Montgomery Simus

California Needs Water and Clean Power. It Might Have a Fix for Both.

Via the New York Times, a look at a pilot program is building solar panels over irrigation canals to generate electricity. As a bonus, the shade prevents water from evaporating.

In California, a sprawling 4,000-mile network of canals winds through citrus orchards and fields of tree nuts, delivering irrigation and drinking water to homes and farms across the state.

The canals are critical in an increasingly arid part of the country. But what if they could help fulfill another urgent need: renewable energy?

To test that idea, researchers, private enterprise and a public utility in the Central Valley are installing solar panels atop the man-made waterways.

The pilot program, called Project Nexus, is testing solar canopies that researchers say could generate gigawatts of power and save billions of gallons of water by providing shade that slows evaporation. It could be transformational if scaled up, researchers say, in helping the state to meet its ambitious climate and biodiversity goals.

California aims to generate 60 percent of its electricity from renewable sources by 2030 and to rely entirely on carbon-free energy by 2045. It also aims to conserve 30 percent of its open land by 2030. Placing solar panels above the canals, rather than on undeveloped land or competing for space with agricultural interests, could be a double win.

The project grew out of a 2021 study by researchers at the University of California, Merced, that aimed to assess the viability of solar installations throughout California’s canal network.

That simulation modeled the effects of covering all 4,000 miles of the state’s major open canals with solar panels. Researchers projected doing so could generate some 13 gigawatts of solar power (roughly half the new solar capacity needed to meet California’s energy targets) and save 63 billion gallons of water annually (enough to serve 2 million people or 50,000 acres of farmland).

The various shapes, conditions and locations of the canals make the study’s maximum projections nearly impossible to achieve in real life, said Brandi McKuin, a project scientist at U.C. Merced and lead author of the study. And the costs would be higher than building solar farms on land. “It’s probably unrealistic to assume that we’re going to cover all 4,000 miles of California’s canals,” she said.

But the modeling showed that building solar on even a portion of the waterways could make a meaningful contribution to energy production and water conservation.

The study received global attention and, soon after it was published, the state of California contacted Dr. McKuin with a question: Could she prove it?

The next year, with $20 million in state funds and a mandate to test their projections, Dr. McKuin and her colleagues found location in the agrarian corridor south of Modesto and a partner in the 139-year-old Turlock Irrigation District, the oldest canal operator in the state. The community-owned utility operates 250 miles of canals and distributes electricity to 240,000 people.

There, Project Nexus partners including the irrigation district, U.C. Merced and a development firm called Solar AquaGrid tested designs, eventually constructing steel scaffolding above two canals, a narrow one no wider than an alleyway and a larger one roughly the width of an eight-lane highway. Together, the installations cover an area about the size of one-and-a-half football fields, generating a combined 1.6 megawatts of power.

They built panels facing south and west, experimenting with a rail system that allows workers to pull the panels aside to access the canal below, all in an effort to prove that the concept might work for the myriad canal types winding across California and beyond.

Having gathered data over a full irrigation season, Dr. McKuin said the initial results from Project Nexus are encouraging.

Preliminary readings from sensors placed in the canal indicate that the shade from the solar panels reduced water evaporation by up to 70 percent. The panels also slowed photosynthesis, reducing aquatic weeds and algae by up to 85 percent.

Solar-covered canals aren’t an entirely novel idea. Two were completed more than a decade ago in Gujarat, in western India. And along I-10 south of Phoenix, the Gila River Indian Community built a project in 2024.

The Gila River project ended up generating 1.5 megawatts of power, 25 percent more than estimates, possibly because the cooling effect of water can make solar panels more efficient. Water temperature dropped a full degree as it traversed 3,500 feet of shaded canal, with no algae growth, said David DeJong, the irrigation project director. He predicted a “paradigm shift” in the West if the technology were to be widely adopted.

Despite the promising initial results, Dr. McKuin said more research would be needed before scaling up the concept. “It’s still really early to say what the economic feasibility of this is,” she said.

A new report from U.C. Merced, expected in the coming months, will be “critical” to determine whether the Turlock Irrigation District will invest in more solar canals, said Josh Weimer, the district’s director of external affairs. Putting solar panels above a canal is more expensive than installing them on an empty patch of desert, and for districts like Turlock to adopt the technology, Weimer said, the construction costs would need to be offset by the combined value of the conserved water, land-acquisition savings and the reduction in aquatic weed maintenance.

The California Department of Water Resources said it is watching closely as it considers solar canals for portions of the State Water Project, a large network of canals and other infrastructure that delivers water to 27 million people. Ultimately, the data from Project Nexus will be “essential” to understanding how solar canals perform in the real world, said Andrew Schwarz, the agency’s climate action manager, in a written statement.

The technology is “absolutely” ready to be developed at scale, said Roger Bales, professor emeritus of Civil and Environmental Engineering at the U.C. Merced who has worked on Project Nexus since its inception. Now, his goal is to build the first 100 miles.

Looking ahead, he thinks solar canals in California could feasibly generate up to a gigawatt of power in the next decade. “We have to get to a hundred miles, and then it might take off,” he said.

The key is finding the “low-hanging fruit,” said Jordan Harris, the chief executive of Solar AquaGrid, for example identifying places across California where canals are adjacent to a localized energy need. That could be a water-pumping station or an electric-vehicle charger along the highway, or anywhere local electrical lines can accept power without the need to build new transmission lines.

Could New Tech Help Save Some Very Rare Whales?

Courtesy of the New York Times, a look at how innovative systems to keep ships from hitting North Atlantic right whales are coming into use. The Trump administration is weighing whether they can replace a bedrock protection.

Trackers that ping satellites every time a whale surfaces for air. Thermal cameras that can detect the animals day and night. Acoustic devices that monitor their calls.

These and other emerging technologies could help protect the world’s last surviving North Atlantic right whales from collisions with vessels, a leading threat to their survival. And the Trump administration, as part of its deregulatory agenda, is considering whether the new systems could replace a simple, core protection almost two decades old: seasonal speed limits for large vessels.

The innovations, some of which are already in use to various extents, collectively show great promise. But they each face limitations and would require enormous resources to roll out comprehensively, meaning they are not yet a viable substitute for broad, mandatory speed restrictions, according to interviews with several of the scientists who are developing the technologies.

“That geographic scale is going to be very, very, difficult,” said Mark Baumgartner, a senior scientist at the Woods Hole Oceanographic Institution who uses underwater acoustic monitoring to detect the whales. “It’s going to be expensive and it’s going to be a major challenge.”

North Atlantic right whales are one of the most endangered species in the United States. Whaling drove them to near extinction. Since 2017 they’ve been experiencing what the government calls “an unusual mortality event,” with an elevated number of whales being found dead or injured. Only about 380 remain on the planet.

Vessel strikes and entanglements in fishing gear are the two primary threats, according to National Oceanic and Atmospheric Administration Fisheries. Global warming also appears to be playing a significant role, as changing ocean conditions cause the whales to follow their prey into new areas where they are less protected from traffic and fishing gear. Females are giving birth to fewer calves, probably because of a combination of stress and difficulty finding food.

For the species to recover, NOAA Fisheries has determined that even one human-caused death per year is too many.

The very nature of North Atlantic right whales makes them especially vulnerable to vessel strikes. They stick close to the coast, often swim at or near the surface and lack a dorsal fin, making them harder to spot.

Since 2008, vessels 65 feet or longer have been required to operate no faster than 10 knots, just under 12 m.p.h., along certain stretches of the coast at certain times of year. During the Biden Administration, NOAA Fisheries sought to expand those areas and include smaller boats, starting at 35 feet. But the boating industry, sport fishermen and harbor pilots fought the measure, saying it would cause economic devastation. The proposal quietly died.

In March, the Trump administration announced that it was reconsidering the existing seasonal speed restrictions and looking at the possibility of replacing them with alternative management areas and technology-based alternatives. The goal, according to the notice, was “to reduce unnecessary regulatory and economic burdens on the regulated community.” The public comment period ends on June 2.

“Technology has the potential to provide transformational tools to prevent species extinction while allowing marine industries, including fisheries, boating, and shipping, to continue to flourish,” Rachel Hager, a spokeswoman for NOAA Fisheries, wrote in response to questions.

The boating industry has welcomed the move. Ten-knot speed limits are prohibitively slow, affecting transit times and fuel efficiency for recreational vessels, said Gettys Brannon, president and chief executive of the South Carolina Boating and Fishing Alliance, a trade group.

“Regulations that were written 20 years ago shouldn’t remain frozen in time despite major advances in technology,” Mr. Brannon said. “In 2008, most people didn’t even have smartphones yet.”

An industry task force is working to identify, develop and implement technologies and monitoring tools to prevent strikes.

But a report assessing the readiness of various technologies, commissioned by NOAA Fisheries and completed in November, found that none had high marks in all three categories evaluated: the ability to detect North Atlantic right whales, the ability to reduce strike risk for individual vessels and the ability to reduce it regionally.

Acoustic monitoring was found to be the closest to ready, according to the report, which was completed by the Mitre Corporation, a nonprofit organization that provides research and development services to the federal government.

But the devices, which use specialized software on buoys and autonomous gliders to detect the calls of right whales, only work if the whales are calling. Mothers with calves, arguably the most important whales to protect, are known to be especially quiet.

Furthermore, expanding acoustic monitoring up and down the East Coast would be a herculean endeavor, said Dr. Baumgartner, who helps develop and deploy the technology. Currently, there are 11 such buoy stations and nine gliders operating in U.S. waters, he said. (In the United States, acoustic detections of right whales currently trigger voluntary speed restrictions.)

Each of the technologies has its strengths and weaknesses. NOAA Fisheries considers thermal imaging to be one of the most promising, Ms. Hager wrote.

Devices sit on individual vessels or on land, detecting whales when they come up for air based on the temperature difference between the water and their exhaled breaths, called blows. Significantly, it works just as well at night.

The technology has scaled up quickly, said Daniel Zitterbart, an associate scientist at Woods Hole who helped develop it and co-founded a company to bring a product to market.

But the devices only work efficiently aboard larger vessels, because they need to sit high above the water, Dr. Zitterbart noted. And critically, the technology will only help if vessel operators act on the detections.

“There is the danger you put this on a boat and people just don’t use it,” Dr. Zitterbart said.

The kind of sonar used to detect submarines is not a good option for finding whales, he said. It introduces more sound into an already noisy ocean and is less effective for objects close to the surface, where whales are vulnerable to strikes and where they spend a lot of time.

Tagging whales with sensors for satellite tracking is enticing, but challenging. To ping satellites, the tag needs an antenna that sits outside of the animal’s body. Tags that don’t penetrate the skin tend to fall off quickly, often within hours.

Alex Zerbini, a senior research scientist at the University of Washington, has developed a tag that anchors below the whale’s blubber. It’s showing success on southern right whales, a related species with a healthier population, he said.

But the approach is controversial. The tag creates an open wound, like a body piercing. As the whale’s body heals around it, the tag is slowly pushed out and dislodged over the course of roughly a year. Regular retagging would be necessary.

Brooke Flammang, a biology professor and director of the Fluid Locomotion Laboratory at the New Jersey Institute of Technology, is working on a new kind of noninvasive tag, modeled on the way remoras, also known as suckerfish, attach themselves to whales and other marine animals. Early results are promising, she said, but she’s still developing the device.

“We need the speed rules to stay in effect while we’re finalizing other technologies for deployment,” she said.

Researchers at Duke University in North Carolina are working on a system that could take in data from all kinds of detection technology, combine it with other metrics and predict the presence of North Atlantic right whales the way meteorologists forecast weather. Measurements like water temperature and chlorophyll concentration, which can signal food, help scientists predict the presence of whales.

“Instead of showing rain potential, it shows whale potential,” said Jason Roberts, a researcher at the university’s Marine Geospatial Ecology Lab.

Currently, based on decades of research, the team can offer monthly predictions. But to effectively protect whales with minimal disruption to vessel traffic, they want to shrink that to days.

“That’s a huge leap, and it’s a very, very different and daunting task,” said Patrick Halpin, a professor of marine geospatial ecology at Duke and the lab’s director. The research and development is still a couple years out, he said, and then implementation and testing would require several more years.

We Can Now Track Animals From Space. Here’s Why It Matters.

Via BBC, a look at how we can now track animal panic from space and why it matters:

After decades of development, wildlife surveillance has finally come of age. The new Icarus satellite is tracking signals hidden in animal behaviour – which could save the lives of cheetahs, rhinos and elephants.

On a blustery morning at the start of Namibia’s winter, a pickup truck idles to a halt on the edge of Okambara Elephant Lodge, a private wildlife reserve 100 miles (161 km) from the capital of Windhoek. Two women and two men – one armed with a rifle – step out onto the red soil.

Throughout June, Okambara is a bone-dry expanse of thorny trees and shrubs. Although the Sun is shining, cool winds keep the park’s animals vigilant, as the wildebeest, zebras and giraffes sniff out scents on the breeze, which could alert them to danger now moving through the bush. Yet the skilled intruders remain hidden downwind.

As the hunters close in on the game, the rifle lefts out a boom. Fear jolts through each species: springbok bounce, skittish zebra break into gallop, and the wildebeest turn and race, some not stopping for hundreds of metres, as they barrel away from danger into Okambara’s wide-open salt plains.

Scientists are now able to study these signals written in animal panic thanks to a new satellite system, named Icarus, which is tracking animal movement and behaviour on an unprecedented scale from space. By monitoring how animals react to the presence of human intruders, conservationists hope to pinpoint and crack down on poachers.

Patterns of panic

Over three days in mid-2024, the intruders in Okambara make around 30 of these salvoes – all captured through the lens of an unmanned drone that hovers overhead. From this sky view, the rapid dispersal plays out, again and again, with animals tracing out signature patterns of panic and withdrawal.

The team of hunters fires dozens of rounds and the game scatters, except for the giraffes, which usually remain impassive and calmly look on from their raised vantage point. Yet by the week’s end, not a single victim has fallen to the hunter’s gun. That’s because, unlike the poachers who have killed hundreds of rhinos in southern Africa, this hunting party is not here for slaughter. Instead, today’s team are scientists doing their best to simulate the arrival of a deadly threat.

The armed interlopers – an ornithologist, an expert game hunter, and two wildlife researchers – are part of an experiment to develop a real-time tracking system that could save the lives of Africa’s most trafficked animals. By recording the distinctive patterns traced by different species as they react to a hunter, the team ultimately aims to train an algorithm that can send out a warning alarm to rangers.

These warning systems are still in development at nature reserves – but the recent launch of a wildlife tracking satellite, dubbed the “Internet of Animals”, aims to link up a truly global system of real-time alerts.   

Okambara, a flat 169 sq km (66 sq miles) reserve, has become the “perfect site to test the system,” says Sierra Jane Mattingly, an ecologist at the Max Planck Institute of Animal Behavior in Germany. Here, 5% of all large animals have been fitted with GPS tags that continually monitor their location. But the real goal is to help wildlife in the most precarious places worldwide.

The lessons learned here are helping in the battle with poachers in national parks in South Africa – home to the world’s largest rhino population – and aims to safeguard the free-roaming wildlife populations in currently unmonitored places like the Congo Basin.

We have the other animals protecting the rhinos because they tell us when the butchers are coming – Martin Wikelski

The project is the realisation of a long-held dream of Martin Wikelski, the ornithologist on hand in Okambara. Wikelski, a world-leading movement ecologist who heads the Max Planck Institute, hopes to tag 100,000 animals across the planet by 2030, with the goal of understanding the signals hidden in animal behaviour. As they beam out their movements to receiver towers or satellites, animals can collectively act as “sentinels” to protect rare giants like rhinos, he explains.

“We have the other animals protecting the rhinos because they tell us when the butchers are coming,” Wikelski says, standing behind a model of a satellite, named Icarus, during a Whitley Award conservation conference in London, shortly before the project launched in November.

With the addition of small ear-mounted tags, rhinos’ competitors and nearby carnivores are turned into their protectors, making allies of cheetahs, zebras and giraffes – arguably the perfect sentinels because rows of the long-necked animals will often remain still and observe threats from a safe distance, with their heads all pointing in the direction of poachers, he says. “So we know where the butcher is.”

Smart watches for wildlife

This grand vision of wildlife connected through network technology is all made possible by tiny pieces of electronics that thousands of animals now carry. Wildlife tags are becoming miniature marvels, says Timm Wild, an electrical engineer at Max Planck. Some can track not only GPS position but their wearer’s activity, heart rate and body heat, as well as operating a mobile sensor taking readings of surrounding temperature and atmospheric pressure.

Today’s tags are tiny enough to be carried by birds or even butterflies, like the rice-grain-sized chip developed by New Jersey-based Cellular Tracking Technologies to track monarchs as they migrate thousands of miles across North America. Wild explains that the cutting-edge sensors are powered by supercapacitors – long-lasting, easily rechargeable alternatives to batteries – meaning they could allow us to see where long-lived species spend each day of their lives. “Lifetime tracking is a challenge that is partly solved now,” Wild says excitedly.

Each of these innovations overcomes a major drawback of a technology that has been around for decades. Animal tags were first used in 1970, when researchers in Wyoming fitted an elk named Monique with a 10kg (22lb) satellite collar, which broadcast an analogue VHF signal. In the decades that followed similar tags beamed new discoveries about ungulate migrations and whale journeys. But these heavy, old tags remain restrictive and have low accuracy. Nearly four decades later, they were still too heavy for 75% of birds and mammals.

Wild, an electrical engineer for carmakers like Daimler and Mitsubishi, began working with wildlife in 2019, and was surprised to see how primitive the tech being used in scientific research was. Digital “Internet of Things” sensors had taken off in the previous decade, fuelled by the rise of consumer technologies. Tracking devices could be precision-measured to centimetre scale and intelligent tracking systems could work out where they were located even when satellite connection was not possible, such as when a car following GPS turns into a tunnel or a blind spot between skyscrapers.

“If you put that technology on a bird or a monkey, you suddenly can see where in the tree they sit and what they actually eat,” Wild says.

Computing in the mud

Wild’s team of about a dozen at Max Planck aimed to close the “huge gaps” between the tech available in consumer electronics and in the field.

Okambara is a good example of the limitations of today’s tech. The current system can transmit 12-byte packages of data about every 10 minutes to a transmitter in the middle of the park. During the poaching simulations, researchers saw isolated GPS points and needed the drone to fill in the rest of the story.

Transmitting data back from the far corners of the Earth is a major challenge, Wild says, and heavier tags and bigger batteries are needed to incorporate more memory. His team are working on each of these challenges: “We develop our own prototypes: our own hardware, own software, own 3D printable housings,” he says.

But, for now, newer tags are finding work-arounds that condense raw data into useable insights before transmitting. While some have called this “artificial intelligence on tag”, Wild says that’s a stretch: “It’s like a very, very basic algorithm” that can tell us if a bird has hit a rainstorm or emerged into the sun.

These kinds of real-time notices are especially valuable in conservation, says Mattingly, for example telling us if a particular animal is hunting or resting. “That is great because we can very clearly see if an animal is dead or not,” she says. Projects like Okambara tag both ears to avoid false alarms, as wild animals are continually finding new ways to knock off a tag or crush a tag. But if both ear tags are not moving, they send out a “mortality notice” that alerts a ranger to check on them.

If an animal behaves outside of its usual movement pattern, we can very quickly detect if it’s, for example, sick or if it’s injured – and then hopefully help – Timm Wild

This kind of processing means that tags are now equipped with a level of “situational awareness” about what’s normal and not normal for that animal, says Wild. “For example, we have a lot of data collected about how a zebra normally behaves. Now suddenly it behaves outside of this usual movement pattern – we can very quickly detect if it’s, for example, sick or if it’s injured – and then hopefully help the animal.”

Computing in the mud

This approach has been trialled at Kruger National Park in South Africa, where it has helped rangers to pinpoint wild dogs caught in snares. Out of the 400 wild dogs in the park, some 80 have been freed from snares, showing a major impact on the population, says Louis van Schalkwyk, a wildlife veterinarian based at Kruger, who leads the tricky work of putting tags on wild animals there.

The biggest goal is safeguarding Kruger’s 3,000 rhinos. More than 10,000 rhinos have been poached in South Africa over the last 15 years, according to the International Rhino Foundation. Kruger has long been the species biggest single stronghold but 175 rhinos were killed by poachers in 2025.

“In Kruger, I think the challenge is that it’s a huge place,” says van Schalkwyk. Spanning 19,485 sq km (7,523 sq miles), about the size of Israel, it requires more than 30 of the receiver towers in Okambara, and orders of magnitude more tagged animals to get the same sentinel effect.

Kruger has deployed about 3,000 ear tags on 1,500 rhinos, antelopes, zebras, kudu, oryx and elephants. Van Schalkwyk hopes to build a better “conservation dashboard” by integrating alerts with Earth Ranger, a mapping system that already shows tagged animals and rangers on foot, truck and helicopter.

Currently, tags remain better at providing retrospective data than live updates, he explains. When rhinos have been poached, it can give an indication of the moments leading up to the attack, such as where the attackers had arrived from. “When it works, it’s incredible,” he says.

He’s confident the poacher system will get there but in Kruger it’s not yet a tool that rangers can use every day. “We don’t have an alarm going off here saying there’s 10 zebras telling us there’s someone walking in the bush,” he says.

A lot of things have to work perfectly together for this to happen: “You have to have perfect coverage. You have to have a really good antenna on a very tiny device, which is really tricky”. Then you have to attach it all safely to the ear of “an animal that loves mud and dirt”.

The ‘Internet of Animals’

In late November, Wikelski was nervously preparing for the launch of a satellite that could change the scope of wildlife tracking once again – extending it from a handful of parks to a truly global project. While analogue sensors have communicated basic information with satellites for decades, precise digital tags are usually limited to ground connections. By mid 2027, Max Planck’s Icarus project aims to have six receivers in orbit, making it possible to receive real-time data on animal movements across the planet. Wikelski dubbed the system the “Internet of Animals”.

The first is launching as part of a €70m ($83m) EU-backed fleet of small scientific satellites, serving as a “laboratory in orbit”. Wikelski was stuck waiting all November for the green light from SpaceX to launch the initial Icarus probe into orbit from Vandenberg Air Force Base in Santa Barbara, California.

By the end of the month, he’d endured weeks of false-starts, with four last-minute reschedules, as the US government shutdown resulted in shortages of flight controllers. It felt like being a kid on Christmas Eve, he says, rushing around willing the time to come, “and maybe by 14:00 or 15:00, you’re done,” he laughs. “It’s stupid.”

At 10:44, the condensation-covered pencil-like Falcon 9 rocket erupted from the launch pad, rising on a column of fire and carrying the satellite into orbit. After the years of waiting the whole thing was over in a flash, he says. “Maybe two or three hours afterwards, we already had the confirmation that the satellite is on the right orbit, is communicating, which is really the only thing a satellite has to do.” In May, Icarus launched a second system into orbit – the microsatellite “Raven”. Following a few months of testing, the system will begin receiving data from animals’ tags this summer. 

For animal tracking, the shift from land-based receivers to satellite systems will be like the shift from landlines to mobile phones, Wikelski says. It’s unlikely to solve Okambara’s data bottleneck any time soon.

“I think the biggest impact will be outside of those areas,” Wikelski estimates, in places like the Congo Basin and the Amazon, where animals with large ranges have to navigate wildlife reserves, roads, farms and human settlements. We now have a way to answer questions about the fate of migratory birds and elusive creatures like jaguars, tigers and snow leopards, he says. “What do they need? What else do they need during climate change? Are they going back to certain refugia?”

The core, “absolute key” questions like “where are wild animals living and dying?” have always appeared unanswerable, he says.?”People have somehow accepted ‘Oh, we will never know that,’” Wikelski says. “We can finally do it.”

Case for Using Prices Rather than VPPs to Coordinate Distributed Energy

Via Volts, commentary on how dynamic, time- and location-specific retail prices, communicated directly to consumer devices, would be a better way of coordinating the behavior of thousands of distributed energy resources by cutting out the middleman and leave more value with customers:

David Roberts

Hello everyone, this is Volts for May 8, 2026: “The case for using prices rather than VPPs to coordinate distributed energy.” I’m your host, David Roberts.

The electricity grid has more and more participants. There are increasing numbers of small distributed energy resources (DERs) — solar panels, home batteries, EVs, EV chargers, hot water heaters, commercial and residential HVAC systems, and on and on — that need to be coordinated to work as efficiently as possible for the good of the grid as a whole.

But how should they be coordinated? One answer, which I have discussed many times here on Volts, is what are called virtual power plants, or VPPs. This involves a single aggregator, usually but not always a private party, contracting with dozens, hundreds, or thousands of distributed energy device owners, agreeing to control and coordinate their devices’ behavior, monetize the flexibility, and share the value with them.

My guest today, Bruce Nordman, spent nearly four decades as a research scientist at Lawrence Berkeley National Laboratory working at the intersection of network technology and energy systems, and he thinks VPPs are the wrong answer. Not because they don’t work — they do — but because the gap they’re exploiting, between the real-time locational value of electricity and the flat, time-invariant rates actually charged to customers, shouldn’t exist. If retail electricity prices actually reflected real-time value, there would be nothing to arbitrage — the flexibility value currently being split with aggregators would just stay with customers, captured automatically by their own devices, without any private third party reaching into their home to do it.

The technologies needed to do this — basically, a central price server to send the price signal and devices capable of responding to it — have become cost-effective on a mass scale in recent years, and there are limited experiments underway in some places. In this conversation, Nordman and I dig into it pretty deeply. Okay, very deeply. This is an intensely wonky discussion, more technical than most of what happens on Volts, and a pretty punishing two hours, but if you’re a real head and you want to understand how to actually get the most out of DERs, I think it’s worth it.

With no further ado, Bruce Nordman, welcome to Volts. Thank you so much for coming.

Bruce Nordman

Thank you.

David Roberts

This is a lot, Bruce, and I’ve been thinking about how to ease into it. Where I want to start is at the big abstract level. When you first approached me with this, you said, “Hey, let’s use prices instead of VPPs.” I had in mind, “Oh, you’re going to design some novel tariff for PJM or something.” Something like that. But this is much bigger than that. This is a ground-up new vision for how electricity works. It’s a bit more to get into.

Part of the big vision here is similar to Jonas, who we talked to a couple of weeks ago — Jonas Bergerson — in that this is about transitioning the electricity industry from the old unitary model to a networked system. That is what I talked with Jonas about. This is, I think, an excellent adjunct to that or your sequel, or I do not know exactly what you would call it, but you are involved in, broadly speaking, the same project.

You write that the telephone system and the electricity grid were invented at roughly the same time, roughly the same place. They are interesting analogs to track. As listeners know at this point — we covered it a little bit on the Jonas pod — the telephone system has now entirely been transformed into the Internet. This involved, as I said, going from a unitary system to what’s called a networked system.

The electricity grid notably has not. One of the things I want to talk about starting up front is some of the aspects of what it means to go from a unitary to a networked system. Then we can look at how the Internet did it and how the electricity grid has not done it. You have this chart here of qualities of a unitary system and qualities of a network system. A lot of this I think people understand intuitively.

You’re going from a centralized system to a distributed system. You’re going from a largely analog system to a largely distributed digital system. But here’s one that I want to look at because I think it’s important for understanding the electricity grid and why it’s such a problem. One of the things that the telephone system did is go from tightly coupled systems to loosely coupled.

Briefly, and keep in mind briefly because we have a lot of territory to cover, talk about what it means for systems to be tightly coupled or loosely coupled and maybe tell us how the Internet is loosely coupled and then contrast that to tightly coupled systems in the utility grid.

Bruce Nordman

On the telephone system, when I was a child, if I picked up the phone and called my grandmother, a circuit was set up and there was a continuous flow of data at a constant data rate between the two ends of the system over one specific circuit. Even if no one was talking, it was there.

One thing which enables things to be loosely coupled is storage. On the Internet, you can connect links of different communication technologies that run at different speeds because you can store and forward packets, which accounts for the fact that things are running on different technologies and different speeds. For communications, storage changed everything. Data storage was just as essential as digital communications for being able to move to Internet technology.

David Roberts

One of the themes here — and I might as well just put an exclamation point on it now — is normal people, even people who think about electricity, I don’t think anybody has fully absorbed what a big deal it is to go from a system with no storage to a system with storage. It just fundamentally changes the architectural possibilities. It’s more fundamental than people appreciate. Give me an example of a tightly coupled system.

Bruce Nordman

The old phone system was tightly coupled, where everything was interconnected. The old electricity grid was where every customer site was just part of the grid. It didn’t have any functional identity. It couldn’t operate separately. The grid ended at every end-use device.

David Roberts

For every customer to be part of one big pool, that means if I pull on anything in the customer site, it affects everything, because everything is part of the same system. Everybody’s coupled together. Loosely coupled means that you get some distinction — functional distinction — between these systems such that they can operate somewhat in isolation without affecting one another. Is that fair?

Bruce Nordman

Absolutely. Part of this is the AC frequency of 60 Hz, as we use, is part of what enforces this in that things are all very interconnected and all the load affects the frequency.

David Roberts

One advantage of having loosely coupled systems, as opposed to a tightly coupled system, is if you pull on one string, you affect everything. You can’t fiddle with the machine without fiddling with the whole thing. But if you have loosely coupled systems, you can innovate and change and evolve the systems separately and independently.

Bruce Nordman

Exactly.

David Roberts

That enables much faster evolution and expansion and scope. The one underneath that on the list, I think, is similar and related: entangled technologies versus isolated technologies. Similar sort of concept, correct?

Bruce Nordman

Exactly. This is where the Internet technology example is highly informative, where the architecture of the system does this isolating of complexity so that, as you say, systems can evolve separately. Ideally, the technology inside a customer’s site can evolve separately from the technology on the grid and you can minimize the point of interaction between them.

David Roberts

The reason I think that the telephone system was able to make this leap — one of the reasons you point out — is that we were able to build the new system alongside the old system. For a while we had the two systems going, and then we eased our way from the one to the other.

Bruce Nordman

Originally, Internet communications went over phone lines, as with dial-up modems, and then we flipped it so that with Voice over IP, the phone calls went over the Internet lines. There’s a similar transition from where we are to where we want to be.

David Roberts

What I’m trying to get at here is one of the reasons that the electricity system has not been able to make this similar transition is that you can’t really do what you did with the telephones. You cannot build the new system alongside the old system. It is all one big system and you can’t — what’s the analogy? — rebuilding the plane in flight type of thing — that just makes it trickier.

Bruce Nordman

I would push back on that. This gets back to some of the earlier points you made about the podcast with Jonas, about his approach to networking electricity and mine. On the Internet there is the wide area network and there is the local area network. The local area network is what is inside of each customer site, whether it is your house or an office building or a factory. Then you have the modem and router at that boundary. The wide area network is everything else. There are technologies that only exist inside of LANs, and there is technology that only exists inside of the wide area network. Then of course there are things like the Internet protocol and some other ones that exist in both.

David Roberts

We talk about entangled technologies versus isolated technologies. That’s a great example — the WAN and the LAN. You can fiddle with WAN technology, wide area networks, you can fiddle with local area technology, and you don’t implicate the one and the other. You can work on them separately, which enables rapid innovation. As a final thing, transitioning from the unitary to the networked system — and this is, I think, both the most important one and the most difficult for people to wrap their heads around — is going from a deterministic system to a non-deterministic system.

What’s funny is the reason the telephone people didn’t invent the Internet is that they were running a deterministic system and they were convinced that only a deterministic system could work. They could not imagine how a non-deterministic system could do the communications work that telephones did. Of course, they were wrong. People came along, did it, and they faded into history like dinosaurs.

Now we’re in an electricity — the exact same parallel situation, which is right now the people running it are running a deterministic system and they are convinced that only a deterministic system can work and they cannot imagine how a non-deterministic system could work. Just tell us what is a deterministic system and what is a non-deterministic system.

Bruce Nordman

A deterministic system is where you plan out everything in advance and then you operate according to your plan. Airplanes used to do that. They used to file flight plans with the FAA. They said exactly the route they were going to fly and they had to adhere to that unless they got permission to fly differently so that the FAA always knew what they were going to do. Then maybe a decade ago, I don’t remember exactly when, the FAA said, “Other than these military areas where you’re supposed to stay out of, fly wherever you want.” We just transitioned from a deterministic system to a non-deterministic one. It still works.

David Roberts

You can imagine if you ran the old airplane system, why that notion would freak you out. If you are used to working with deterministic systems, the idea that all the planes are going to run themselves their own direction and you’re just going to coordinate on the fly sounds crazy. Yet it works.

Bruce Nordman

Railroads conventionally are quite deterministic, particularly before — where you plan out where the trains are going to be so that they don’t run into each other. That’s certainly quite important. Car traffic is non-deterministic. When you start driving someplace, you don’t tell somebody in advance exactly the route and timing that you’re going to take. You make it up on the fly by observing the traffic, maybe seeing what Google Maps says and such.

David Roberts

To our point here, the communication system, the Internet, has become non-deterministic. As you said, on the telephone system, if I pick up a line, connect to another person, that is a particular physical circuit that the data is traveling to and from. That’s not how the Internet works. I don’t think people necessarily know this. I think this is worth describing briefly. If you send a packet of information out onto the Internet, it is not the same — there’s not a set path that that packet goes along to its destination. What happens and how do we make it work?

Bruce Nordman

Packets for the same chunk of data, whether it’s an email or some streaming video — different packets in that, and there’s going to be an enormous number of these packets — are going to take different paths and arrive at different times and they are reassembled at the end. If any of them don’t arrive, there are ways to retransmit them with TCP to make the whole system work.

To your point about the phone company being blind to this, multiple people went to Bell Labs in the 1950s and said, “Hey, there’s this new thing called packet switching. We think it’s the future of communications.” The smart engineers said, “Yeah, we know that cannot possibly work.” There are people who say it is impossible for a utility to charge a good retail price. Of course, they are wrong also. It is possible for utilities to charge good and better retail prices. They just choose not to.

To your point about deterministic, the wholesale system for electricity has always been deterministic and my guess is always will be because that is how wholesale markets work. But retail electricity has never been deterministic. People are trying to force it to be through things like VPPs, but that is not necessary in general for balancing supply and demand. Deterministic operation is necessary for addressing distribution system capacity issues, which is its own separate topic that I assume will come up later, but that operates at a totally different scale of a few customers at a time versus millions of customers at a time. The two mechanisms operate in parallel.

David Roberts

The idea here is we’re going to go to a non-deterministic electricity system. People get nervous about that because if you don’t get an email, it’s one thing, but power, home and hearth, your medical machines, whatever — electricity is very important. It’s worth emphasizing that despite rubbing our intuitions the wrong way, I’m not sure our brains are naturally wired to think non-deterministically, but we have found through experience that a massive non-deterministic system can in aggregate be as reliable or more reliable than a deterministic system and can scale much faster.

This is all by way of preface. This is all by way of saying that what you’re working on, what you’re thinking about, similarly to what Jonas is thinking about, similarly to what you and other electricity researchers have been beating your heads on for decades now, which is trying to take this unitary deterministic system, transition it to a massively coordinated networked system and thus reap all the benefits in the electricity system that we got by the similar transition from telephone to Internet, airlines, all these other examples. This is all part of that big transition.

Your basic fundamental point and the reason we’re doing this podcast and the core thing here is that your contention, your hypothesis, is that if you’re going to move from a top-down to a networked system, you can’t have top-down coordination anymore. How do you get coordination? How do you make sure all these loosely coupled systems are operating in concert to produce the desired results? How do you coordinate all this stuff? Your contention is that prices are the only mechanism that can do that at every scale and in every context.

Bruce Nordman

Exactly. I’ve been looking for the last 20 years for other mechanisms. I’ve never found another mechanism that can do that at any scale, in any context.

David Roberts

Before we jump into what that means and how you envision it working, I just want to go over what you see as a few of the highest, high-level benefits of doing this. Why would we want to do this? Why is it worth hashing through how it works and hashing through the trouble? What is the point of the trouble? The main thing, which we just mentioned, is that you think it is context- and scale-agnostic — price can work for a distribution grid, can work for a microgrid, can work for an individual building, can work for a nanogrid inside the building. At every level of organization, the coordination is being done by pricing.

Bruce Nordman

Exactly. One important point here is that I’m a buildings person. I’ve worked on energy use in buildings for the last four decades. I am not a utility grid person, so I try to resist telling people how the utility grid should operate internally. But I also like utility grid people to not tell buildings people how buildings should operate internally because that is not their area of expertise.

David Roberts

I appreciate that. Most of the Volts guests are very grid-centric, but your take is, “I got this covered here in my building, don’t mess with me, just give me my power, tell me the price.” You’re not asking much of the grid. The grid needs to tell you what the price is and supply you power. All the other work, all the flexibility work, all the coordination work is done on the consumer side. This is a very consumer-centric, distributed, building-centric view.

Bruce Nordman

I use the word customer to make clear that this is applicable to all customer types, whether you are residential, commercial, industrial, or agricultural.

David Roberts

Another merit of pricing — and this to me is the most compelling reason to use prices for this coordination — is that it allows the customer to accrue all the benefits and the value of flexibility. If you get a VPP in there, an aggregator in there, the aggregator takes a cut — around 50%. Yes, but if the devices are simply acting flexibly on their own in response to price signals, all of that flexibility value stays with the customer. The customer gets all of it.

Bruce Nordman

Exactly.

David Roberts

When people hear pricing, they think, “Oh, capitalism,” they think exploitation, they think poor people are going to get screwed. Pricing brings up all that stuff. It should be emphasized this is a very customer-centric view here. This is a view of the grid where customers get the value, reap the most value. Customers are in the driver’s seat in this vision and are getting a higher proportion of the value than they would in any other system.

Bruce Nordman

Exactly. Yes. This is designed around what’s best for customers and customer devices, in devices that exist today and will even more so in the future. It’s designing the relationship between the grid and the customer around what is best for the customer, but also works for the grid. Remember, we invented the grid to serve customers. We didn’t invent customers to serve the grid. I think a lot of grid people don’t quite understand that the customer is supposed to be always right.

David Roberts

The third benefit of pricing that I want to throw out there is — and I think people are not going to be able to really understand this until we talk a little bit about the model and how it works — this vision of pricing maximizes customer privacy and autonomy and cybersecurity because, and we’ll get into the details, under your model, the customer is not sharing any information with the grid above it. Again, we’ll talk through how that works. The point is, in this model, customer privacy is total. What goes on on the customer site is not known by anyone but the customer, and it is not the business of anyone but the customer.

That is as privacy-maximizing as you could get. There is no grid and third parties can do some work — we will talk about that later — but there does not need to be any third party looking into your customer site. Your devices are doing the work themselves and they are doing it based on your preferences and nobody else needs to know or mess with that.

Bruce Nordman

Yes, but the grid does continue to know what happens at the meter and they should have smart meter readings at whatever time interval, whether it is 15 minutes or five minutes that the grid decides to measure at. The grid absolutely needs to and should have that meter data and needs it to function effectively. But that is the limit of what they need to know. They think they need to know more, but they are wrong.

David Roberts

Just to clarify, you said this before, but just to clarify for everybody, so they know, we’re talking about retail prices here. For the moment, we’re leaving the wholesale system, the wholesale energy market, the transmission system, aside. We’re talking about retail systems coordinating distribution grids.

Let’s talk about the model. The core concept here with your model is that the grid ends at the meter, meaning what goes on on the consumer side of the meter is nobody’s business. The grid consults the meter and two bits of information are exchanged. One, what’s the price? Two, how much energy do you need? That is all the grid knows about a particular customer. How much energy do they need? They don’t know what devices are in there. They don’t know what the devices are doing. The customer is a black box to the grid.

Bruce Nordman

Yes. My physics colleagues at Lawrence Berkeley Lab, where I was until recently, assured me that the electrons are all the same.

David Roberts

This is crucial. The grid extends to the meter and the meter tells the grid, “At this customer site we need X amount of power.” That is all that the grid knows about what goes on on the customer site. The utility is not poking its nose in and controlling your water heater. All the grid knows about it is how much energy it needs.

Bruce Nordman

It’s how much energy you consumed in the last time interval. It’s after the fact, not before the fact.

David Roberts

But it’s a quantity. This is the information being exchanged — quantity, nothing beyond that. Let’s talk about what you have to have in place to make this work. To make any of it work, you need what you call a highly dynamic price. There have been a lot of — people who discuss this a lot are familiar with time-varying prices and real-time prices. A lot of other terms for this. You use the term highly dynamic price. What are the characteristics of a highly dynamic price?

Bruce Nordman

Its essence is hourly prices that are different every day. Specifically, it is prices that have intervals between hourly and five minutes that bracket the range of what is reasonable. You set those prices no farther in advance than the day before, and the prices are different every day.

David Roberts

Let me ask a couple of questions about this. You draw the line at hourly or smaller. Do you get more benefits the smaller that increment is? Or do you think hourly is frequent enough to do the job?

Bruce Nordman

The important reason to start with hourly is that that is very comfortable for people to move into this. The barrier here is not technology for this. We have invented all the technology we need, all the communication protocols, the price responsive algorithms. The barrier here is the human beings who are fearful of using prices, even though they use prices everywhere else in their life constantly.

David Roberts

You make a point of saying that the prices are different every day. You mean that they need to be responsive to conditions every day.

Bruce Nordman

Exactly. The conditions on the grid are different every day. If the prices are not different every day, they are the wrong prices. It is a basic function of any business to set the right price for their product. If they are not competent to do that, they need to let someone else who is competent to run a business run the business.

David Roberts

One other thing is that import and export prices are separate. Explain what that means.

Bruce Nordman

I have solar panels and I have net metering, and I pay almost nothing for electricity, even though I derive huge value from the grid because I overgenerate a huge amount in the summer because there is no air conditioning where I live, and I import lots of electricity in the winter. I derive huge benefit from being connected to the grid, but I pay almost nothing. Other people’s bills are higher because I am not paying enough.

David Roberts

This is the cost shift that everybody’s always talking about.

Bruce Nordman

I don’t have an opinion on what the difference should be between import and export prices, but the principle that they can be different and probably should be different is a very reasonable tool for grid operators to have in their set of tools. It’s something that devices can readily accommodate and use, so that’s not a problem. Having a stream of import prices and, if different, a stream of export prices — if you are an exporting customer — that should be 100% of the financial complexity that devices need to contemplate for deciding how to optimize.

We should not have demand charges, which I would describe as evil. They are pathological to good load optimization and to buildings. They were invented in the 19th century. We have far better tools to address the same issue that demand charges are intended to address. We should just get rid of them. They became obsolete a long time ago.

David Roberts

How geographically granular do you envision highly dynamic pricing being? Is this just a price for a utility area or is it going to get as geographically differentiated as it is temporally differentiated?

Bruce Nordman

From the perspective of any individual customer, they don’t care whether it’s millions of people who pay the same price or dozens. It doesn’t affect the customer technology, the communications, or the automation as to how locational it is. That’s really for the grid operators to decide.

David Roberts

But you need to know that to set a good price. The whole point is that the price is a measure of the value of electricity. The value of electricity changes from time to time and place to place and can change place to place, even relatively proximate places. Theoretically, the value of electricity could be slightly different between two even relatively close places. Getting the geographic scale right seems important to accurately capturing value.

Bruce Nordman

It’s similar to the time interval where there are diminishing returns to getting to smaller, smaller areas and there are diminishing issues about how to treat people equitably. In California, we are planning to have locational prices hopefully as soon as next year. In fact, there are some pilot tariffs from the largest, and soon the second largest, utility in the state that have locational prices today that are hourly and the size of the locations is several hundred thousand people.

David Roberts

Interesting.

Bruce Nordman

These are areas where the distribution system load shape is different and the reasons for wanting to shift load are different in those different locations.

David Roberts

Highly dynamic prices: if you want customer sites to optimize based on highly dynamic prices, you have to transmit the highly dynamic prices to the customer sites. Briefly tell us what a price server is, what role it plays, and why it is necessary.

Bruce Nordman

A price server is like a web server. A web server distributes web pages to computers to show to people. A price server distributes prices to machines to use in their optimization. People are free to look at prices, but we don’t want to encourage them to do so because they have much better things to do with their time and they wouldn’t be good at using them. Maybe three days a year they’ll want to look at the prices, and the other 362 they won’t. They don’t need to.

David Roberts

Let me just blow that out because we need to make this point generally. All of this, all of it, is meant to be automated. The customer is not going to know or care what the hourly variations in prices are. They’re not going to be fiddling with their devices on a day-to-day basis telling their devices what to do. All of this is meant to be automated and happening in the background and you as the customer can just chill out and take your hot showers and drink your cold beers. I think that’s implied, but just to underscore it.

Bruce Nordman

Exactly. People express preferences in advance as to how they want their devices to behave, and they can change those preferences anytime they want to.

David Roberts

To return to the original point, the point of a price server is just that you ping it and it tells you the price.

Bruce Nordman

Or really, it pushes out the data to machines on an ongoing basis. It’s technically even simpler than the machine having to ping it when there’s a new price.

David Roberts

This is the very simple point of this: if you’re going to have dynamic prices, you need to tell people. You need to broadcast those prices, which might seem obvious, but it’s hard to get information out of utilities a lot of times. Who has a price server? Is there one of these running somewhere?

Bruce Nordman

Yes, the state of California set one up several years ago called MIDAS and PG&E, my utility, has one. It uses the OpenADR3 protocol. Southern California Edison, the second largest utility, is going to use the same price server. The state will convert theirs to be using that protocol. The concept of a price server dates back to at least 2006, 20 years ago. It is not a new idea because people recognized early on if you are going to have prices to devices, you have to have a way to communicate them.

The price server is the essential part. Then you need communication protocols to move the price from points A to points B, C, and D over both the wide area network and also inside buildings. Then you need the loads, whether they are loads or batteries or EV chargers, whatever they are, that can use the prices. Those are the three elements.

David Roberts

There has relatively recently been developed what’s called OpenADR3, which is a modern, light, replicable, perfectly functional protocol. We know what a price server is. We have a protocol to communicate these things and we know how to create devices that are responsive to them. As you say in your paper, there’s not yet a place in the world, correct me if I’m wrong, where those three pieces are in place and working together.

Bruce Nordman

In California with Pacific Gas and Electric, we do have that. There is a problem with the tariff structure, which is a huge problem that’s not worth our time in this podcast, that makes it problematic. We have the beginnings of it, but we haven’t, for various reasons, had the ability to scale it yet. We have the glimmers of it. California is poised to be the place where this will first happen.

Practically speaking, you’re correct. In some places, we have the price servers but don’t have really good prices yet. Once you have them, manufacturers will spring up and use them because they can deliver more value to their customers and produce a higher quality product.

David Roberts

Here then we get to, in my mind, the really juicy, interesting stuff. This is what I struggled with and this is what I had a little breakthrough about. Let’s talk about the building side. As we say, the only interface in your system between the grid and the building is just this single point of connection, and the only information being exchanged is quantity and price.

Here you are inside your building, your microgrid controller, your central controller of your building devices, which I guess is like a panel or — I don’t know what the central controller looks like. I imagine it will be different for different buildings. Your central controller receives the grid price. This is what’s interesting to me. The value of electricity for that microgrid controller, for that particular building, might be different than the grid controller price. Why is that? What are the considerations at the building level that would mean the value of electricity differs from the grid price? You call this the value of electricity at the building level — you call this the local price, which could be different than the grid price. What are those considerations that would make the value of electricity different for a building than what it says on the tin at the grid level?

Bruce Nordman

New solar customers in California for the last three or four years may be buying electricity at 50 cents, but if they want to sell it to the grid, they’re getting 8 cents. What price should the water heater be getting to decide whether it should be heating water for the next hour or not? If you’re currently importing, presumably you should be getting 50 cents. If you’re currently exporting, if it turned on, it would be consuming 8-cent electricity. It should be getting that signal so that you consume your low-cost electricity instead of higher-cost electricity.

An asymmetric tariff where the import and export are different is the most obvious and widespread reason to have a different local price. I used local as in local area network as in IT, because we want to emulate IT principles and terminology as appropriate as much as possible.

David Roberts

Whether you’re currently importing or exporting will change the local value of electricity. This is where you could import any values that you want. I as a customer might place a very high premium on avoiding greenhouse gases. I can tell my local microgrid controller to adjust the grid price by X amount to reflect that. The value of the electricity inside my building is slightly different than the grid price.

Bruce Nordman

Exactly. That’s the second reason that can create a local price — taking into account the climate pollution from the greenhouse gases. You simply take that GHG signal, multiply by the dollar per ton value that you, as the customer, believe is appropriate, add that to the retail price because climate change is real, and then have your devices optimize to that price. It doesn’t change what you pay on your bill, but it changes how your devices shift their load.

David Roberts

This is key. The point of determining the local value of electricity is not that anyone’s paying anybody anything inside the building, it’s just to put a value on the electricity. The whole point is to channel power to its highest value spot application within the building and work down from there. The microgrid controller, your building controller, gets a price from the grid and then translates that grid price into a local price. It takes the grid price as an input and then adds whatever local considerations there are and comes up with the local price, and then it sends that local price down to these little individual nanogrids that I was talking about before that are in your building.

This is the fun bit to me. That little nanogrid controller takes the local price that it got from the microgrid controller and does the same thing. It says, based on my hyperlocal considerations — you might have a nanogrid that’s an EV charger, whatever, or in a washing machine. That nanogrid controller will, based on its hyperlocal considerations, come up with another local price, an even more local price, a hyperlocal price. That price will be used to coordinate the behavior of the devices on that particular individual circuit.

This is what I want to emphasize — this cascade that’s happening. The price server is sending a price to a building, a building then recalculates to a local price, sends that down to a nanogrid, which then recalculates that to an even more local price. What your grid controllers are doing — your microgrid controllers, your nanogrid controllers — are assessing the value of electricity at that particular time, literally that particular time and place, and coordinating the devices accordingly. This answers the question — when I threw this out on Bluesky, a lot of people were saying, “How can prices — the price is the same everywhere in the building, so how could that possibly coordinate all these different devices?”

The answer is that the price is not the same throughout the building. In fact, the value of electricity might not, but might, differ from nanogrid to nanogrid. The reason it unlocks something for me is that I was getting hung up on the word price. I just think the word value, at least for my purposes, for my brain, value captures this better. You’re trying to compute what is the real-time, geographic-specific value of electricity at this particular spot. That calculation is happening everywhere, at every level.

At every level, from the very bottom, the most specific device itself, right up to the nanogrid to the microgrid to the larger microgrid, the value of power is being calculated at every point in that microgrid and power is being channeled to the highest value uses. I want to get people thinking about value rather than price, because the whole point of this is to make sure that the power itself goes to the highest value uses.

Bruce Nordman

Price is just the unit of measurement by which we measure value. It’s dollars per kilowatt hour. That’s what you use for any of these, so long as you’re in a country that uses dollars. It’s always that same unit that you’re using, no matter what the scale is. My house, where I already have several price-responsive devices, I don’t yet have a local price because I have net metering and I haven’t set it up to take into account the greenhouse gas. Most customers will start out with this using the straight import retail price. That’s fine, that’s the starting point. The communications and automation technologies facilitate this more rich utilization of it trivially. That’s just part of how it works. You get all that for free.

David Roberts

One thing I think is important to point out here is that your building is calculating the value of electricity to you individually, your considerations, your local considerations, and acting accordingly. That is something that a VPP can’t do. A VPP is going to round up a bunch of devices or a bunch of houses and treat them as a class, as all the same. A VPP can’t by its nature calculate what is optimal for each individual customer. Only prices can do that for you.

Bruce Nordman

Correct. The VPP doesn’t have access to the correct information to make the right decision. But also the VPP is acting first for its own interests, second for the interests of their customer — which is the utility grid — and only thirdly for the actual electricity customer. With price responsiveness, 100% of the value, as you mentioned, goes to the customer and 100% of the control is with the customer and the customer devices. Customers may choose to have a third party optimize, say, their thermostat, and that could be the product manufacturer, it could be someone else, it could be a nonprofit, it could be the electricity retailer, and that’s fine.

They might even pay a small subscription fee for that service. But the customer is always in control. The optimization is 100% for the value of the customer because the customer is engaging that third party for the customer’s benefit. The key is that if we only had VPPs as an option, I would be an enormous fan because they deliver important value.

There are some things that VPPs do that price is not the right tool for and should not be used. For example, emergency load reduction. If a power plant or transmission line fails and the grid needs an immediate, within one to two second drop in load. That is a great use of VPPs, and that is not something you would do with pricing. There are other things around the edge that are great uses of VPP. It is not that we should not use them, we should use the right tool for the job. Part of the right tool is the most cost-effective tool.

David Roberts

Here we come to a key question, which is: the grid comes to my meter, tells me a price. My building microgrid controller takes that price as an input and computes the local price and then optimizes my devices on my site, on my consumer site accordingly. What does it pay? How does it negotiate? If the price is different, if the local prices or local value is different than the grid value, what happens between the grid and the meter? There is presumably some negotiation.

Bruce Nordman

The grid only sees the meter readings and the price that the grid announces and then computes the bill. That is the only involvement of the grid for balancing supply and demand. There is the related topic of capacity. Putting that aside, there is no negotiation. All of this local price stuff is completely invisible to the grid.

David Roberts

The customer side is computing based on the local value of electricity. To me at this moment, here is how much I need and here is how much I am willing to buy at this price.

Bruce Nordman

The optimization could be in the central controller or in the individual device, and most buildings will use both of those or in the cloud.

David Roberts

This is something I wanted to ask: do you need the central controller? If the devices are smart, couldn’t they just do the work? Is it necessary that you have a central controller?

Bruce Nordman

It’s not. If the retail import price is all you need, then you can just send that price directly to devices and they can use them. You do not need it, you will want it over time. The transition from having no infrastructure to infrastructure devices is a common one. With cable TV, originally the cable went directly into the back of the TV and then we moved to have set-top boxes — infrastructure devices that decoupled between the two domains. With dial-up modems, your computer was directly connected to the Internet.

Now we connect to our local area network through some switches and routers to the modem and then that connects to the wide area network. Creating infrastructures is natural, but to your point, you don’t have to have one to start using it, as long as you’re just optimizing to the direct retail price, which with better prices is going to be much better than what occurs today. We can do this transition incrementally and organically. It doesn’t have to be an all-at-once thing.

David Roberts

Say I’m in a neighborhood. By nature, my neighbors will be experiencing the same climate as me, the same temperature. They will probably be roughly comparable to me socioeconomically. Our price comes to our neighborhood and then all the devices, all these customer sites take in that price and then act on it. Presumably they’re all going to be making roughly similar calculations and they’re going to all act roughly the same way.

Say, for instance, a cheap electricity price comes down one hour and then every EV charger in the neighborhood is like, “Hey, cheap electricity, let’s charge our EVs.” They all go start charging EVs, you get a massive surge and you overwhelm your neighborhood transformer. In other words, how can a single price that’s going to a neighborhood cause individuals in that neighborhood to act differently? It seems like they’re going to act as a school of fish and that’s going to cause all kinds of capacity problems.

Bruce Nordman

Several things are going on here. One is that if you have a precipitous change in price, you’re likely to get a precipitous change in demand. Many utilities around the US and around the world see this with TOU timer spikes where, to your point, EV chargers are all set to start charging at midnight. You create this problem not only locally, but also for the macro grid.

The answer is don’t have a precipitous change in price. If you have these more fine-grained hourly or half-hourly prices, then you can have the price change slowly, not dramatically. Even within the EV chargers, different cars will need to charge for different amounts of time and need to be complete at different times of the day.

David Roberts

But will they though? They’re probably all going to work in the morning and they’re probably most of them coming home in the evening. Their behavior is going to be roughly similar.

Bruce Nordman

Maybe one needs to charge for two hours, one for three, one for five, another for three, so that if there’s a trough of low price times, then the utility needs to just set the size and shape of that trough to result in the load shape which is best for the grid. Prices should be set based on the result that they produce for the grid, not based on the state of grid infrastructure.

Utility people commonly assume that you must set retail prices based on cost causation, which is wrong. They don’t know that it’s wrong, but it’s wrong. Prices should be set to produce the right result. That’s what airlines do. They don’t charge prices based on the state of the infrastructure, their airplanes and their staff. They set it based on past experience with how customers buy tickets, by geography and calendar, et cetera, and then set prices to maximize the most benefit for them. Utilities should set prices for all of their considerations — for generation costs, for transmission issues, distribution issues, environmental goals, et cetera — produce the best load shape, both in aggregate and also by these locational areas. That will have a loose relation to things like wholesale energy costs, but will not have a tight relation.

David Roberts

I’m not sure everybody’s going to understand the distinction here. The two separate questions here are how to distribute power and then what happens if particular circuits are overloaded. In other words, the capacity of the circuit to handle. A price signal goes to the devices. Devices act a certain way, they overload the circuit. What tools do you have to keep capacity in safe boundaries? If you charge an EV, it is a large power draw rapidly, a big burst of draw in a way that a stove or an air conditioner or whatever is not going to be.

Especially if you have a DC fast charger with multiple stalls. If three EVs hook up all at once, all of a sudden you get a ginormous demand out of nowhere almost instantly.

Bruce Nordman

Or if you have a pole-mounted transformer with 10 houses, six of them have EV chargers and maybe one of them has two, you can easily get the same thing. For my house on an annual basis, ignoring my PV panels, it uses about a kilowatt on average over the course of the year. The highest single hour was a little over 4 kW for everything before I got an EV. An EV charger can be 7 or even 9 kW all by itself.

David Roberts

Wild. Running your EV charger more than doubles the previously high point of your instantaneous load.

Bruce Nordman

For the highest single hour and it could be like eight times the typical. That’s incredibly bursty.

David Roberts

Which makes it a new kind of thing on the grid. This is also worth emphasizing. It’s not that we have a lot of experience dealing with giant bursty loads. There haven’t been a ton of them.

Bruce Nordman

The first place that this capacity issue showed up was in Australia, which — to a recent podcast of yours on rooftop PV — there, a number of years ago, had an increasing number of circuits where on certain days in the afternoon when solar production was at its peak, the transformers and wires were approaching their capacity. They said, “Okay, nobody else on this circuit gets to install PV, forever, unless we spend lots of money increasing the capacity of the system.”

David Roberts

Or they can curtail. Now they are writing that into the contract that they can curtail your individual solar in these circumstances.

Bruce Nordman

Rather than saying no one else can install solar, when in fact most of the time there is extra capacity — it is only a few hours of the year — what they started doing was broadcasting out interval power limits to each customer, not to how much the PV can produce, but to how much the customer site as a whole can export. This is called dynamic operating envelopes. Then they can fairly allocate the capacity to each customer and it is different for each customer, different every day, to make maximum use of the capacity that is there and guarantee that the system will not be overloaded. That is an incredibly powerful mechanism.

David Roberts

An individual customer site is told, “Here’s how much you can export this hour.”

Bruce Nordman

Correct.

David Roberts

If its solar is running, it might take this and say, “Okay, well then instead of exporting all the solar, I’m going to put it in my battery.” That’s the kind of calculation that’s going on on the consumer side in response to this.

Bruce Nordman

Correct. Anything that you might have been in danger of having to curtail, you instead put into a battery, you run the air conditioner, you heat some water, whatever it is, and we want people to shift loads those times anyway. This mechanism just reinforces what is good for the customer and good for the grid as well. There is no financial aspect to this. It is purely a requirement for operating to make sure that we keep the system safe.

Doing this kind of digital capacity management is not new. USB, which has now been around for more than a third of a century, from day one, has had capacity management. You plug a device into the USB port in your computer, it can’t just take power, it has to first request capacity. There’s a negotiation about the capacity. This is fundamentally different from the pricing, where the communication is one direction. For the capacity management, when you get to things like EVs, there does need to be some negotiation so that the grid can allocate through some dynamic mechanism capacity. Devices are able to reserve capacity when they need it and plan for it and guarantee that the system will not exceed it.

The details of that mechanism are still being determined. People have different theories and we need to do more experimentation on that. Pacific Gas and Electric and Southern California Edison in California are doing experiments with some mechanisms. They are more band-aid rather than permanent ones. The protocol you mentioned, OpenADR3, has two mechanisms for it.

This is an area that needs much more attention, but it can be done in coordination with the customer site as a whole, rather than trying to micromanage individual devices. What happens at the meter is the only thing that affects the grid, it simplifies things so much. If that is the point of interaction rather than trying to work around the meter because you do not have the right arrangement of the meter.

David Roberts

The take home here is you’ve got two separate mechanisms — pricing mechanisms and capacity mechanisms. According to you, once you’ve got those two, you’re mostly done. You need to send out price signals and then you need to send out capacity envelopes that people need to stay inside. With those two bits of information — what is the price and what is the hour, what is the capacity envelope for the hour — that’s all the information customer sites need to do their optimization and to do collective optimization.

Bruce Nordman

Correct. Yes.

David Roberts

This capacity will also be a relatively local thing too. Generally these capacity issues come up on individual transformers or —

Bruce Nordman

It could be the first transformer in front of a house, for example, or it could be a substation or the wire in between. It could occur at different points in the distribution system depending on the details of it. That’s all inside the grid where the constraint is — it doesn’t matter to the customer. The customer only needs to know what’s the mechanism for coordinating with the grid so that things work out the best possible for everybody, which is to use all the capacity that’s there and guarantee that it’s not exceeded.

David Roberts

Here you have — and this I think is what comes up with a lot of people when they first hear this — a system that’s based almost entirely on price. Price — how much does it cost, who can pay? When people hear that, their mind goes immediately to, “What about people who can’t pay? How is this not rapacious capitalism, et cetera, et cetera?” There’s a question of if price and capacity are your two mechanisms and your only two mechanisms, how do you ensure other kinds of social or political goals or guidelines? How do you layer other stuff in?

Bruce Nordman

I care as much about this as anyone else does. When I started thinking about this, I started thinking about it in the context of off-grid systems in developing countries and how to make that work. That’s how I originally came to be using price. Simply changing the price does not increase system costs. In fact, changing the price is a way to reduce system costs and therefore reduce bills. That has to be clear in people’s minds — dynamic pricing is about achieving lower bills. It’s not about achieving higher bills.

In California, what we are doing for the near term and what we should do forever is that you charge an hourly price and you calculate the bill according to that and according to some TOU price. If you would have paid less under the TOU, you pay that to make sure that the people who today are more expensive to serve than others continue to pay the TOU and the people who are less expensive to serve and then also can do additional load shifting, pay the hourly. Over time more and more people will just be doing the hourly and fewer and fewer will be on this TOU backstop. That is easy to understand and easy to implement.

David Roberts

How do you protect low-income ratepayers?

Bruce Nordman

They continue to pay the TOU if they are more expensive to serve. Many low-income people today are less expensive to serve and are paying more than they should because they’re consuming more at the low-price times. Others are the reverse. It’s not that low-income people are some monolith. The current system is not fair. We have to recognize that. Current tariffs are not fair. Absolutely. If we want to make things fair, we can move in that direction and move to it in a way which protects anybody who has non-optimal load shapes. We can help them change their load shapes through acquiring devices, controls, or new appliances as we’ve done with energy efficiency for decades.

David Roberts

Is there any functional distinction between an individual customer site, a building, say, that is its own small microgrid, and a larger microgrid in which that building is a part? To the larger grid, the microgrid is the black box, correct?

Bruce Nordman

That’s absolutely true. This community is essentially a single large customer from the perspective of the utility.

David Roberts

That’s what I’m trying to get at. Theoretically, a whole community could be, from the grid’s point of view, a single customer if they were all on the same microgrid.

Bruce Nordman

Correct.

David Roberts

They could calculate their local price based on their local values.

Bruce Nordman

But that’s the retail price for that community. That would be at best a locational retail price. It’s still not a local price inside of a single customer site. They’re similar, I can see that, but you still have a utility-customer relationship between the community and the final customer. That’s not a problem. That’s just not the same as inside of a customer site.

David Roberts

The whole question is: how do you allocate your power to your highest value? How do you allocate power? The way you want to allocate power is you want to put it where it’s most valuable first and work down from there. To do that, you need to calculate what is the value. The key insight here is that the value from the grid perspective is going to be slightly different than the value from the building perspective. Value is being calculated all the way down, such that when you get to the device level, the actual device optimization is working on extremely local, very specific value information. You are optimizing for value at every level.

Bruce Nordman

Correct.

David Roberts

That, I think, is the point here.

Bruce Nordman

The water heater doesn’t have to know where the price came from or why. It only cares about what the price is. The grid could go down. Your house continues to operate as a microgrid because you have a battery and the price may change. The water heater re-optimizes. You may have a higher price or a lower price, depending on what is going on. The water heater doesn’t care about why, it just cares what the price is. All of this topological complexity in no way affects how devices operate.

David Roberts

Let’s return briefly to the point about non-deterministic systems, because I think now that we have described how this works, how price is being used as a coordination and optimization mechanism down to the device level. In what sense is this system non-deterministic?

Bruce Nordman

Retail customers have always been non-deterministic in their consumption of electricity, with a few niche exceptions. In general, you don’t promise in advance to your utility grid how much electricity you’re going to consume each hour for the next month or the next year. You just consume whenever you want to. That’s always been the case about how electricity works. I’m not changing that at all. We can either charge bad prices or we can charge good prices.

What’s a good price? A good price is a price which results in a better overall situation for the grid, the entity setting the prices — that’s their job, to set the price which produces the best result for them. In the past, the price was just a revenue generation mechanism. That’s all it was. In the present and future, the price is also a control mechanism to shift load. What you need to do to set retail prices in this context is go through two separate calculations.

One calculation occurs today, which is the revenue requirement of the utility. How much revenue do they need to collect at the end of the day, month, or year. This doesn’t change that one bit, except that we will lead to needing less revenue because the system can be more efficient. Secondly, you go through a calculation of what’s the shape of the prices — 24 hourly prices or 48 half-hour prices. The grid, based on past experience, determines what shape for this locality will produce the best result for the grid. Then shift that shape up or down so they collect the right amount of revenue. Those two things happen in parallel so that you meet both objectives: collect the right amount of revenue and charge the right shape so that the right load shifting happens.

In general, the water heater, because you are only shifting load, you are never shedding load with your water heater, only cares about the difference in prices, it does not care about the absolute price. These two things work seamlessly together to shape load and to collect revenue.

David Roberts

I come back again to how this is going to produce a more efficient system. I think one of the advantages of this is that it makes flexibility not something that you try to add on top of the system or that you try to tweak the system afterwards. Flexibility will just be part of normal electricity consumption. In this, the flexibility is part of the operation of the system, not some separate thing.

Bruce Nordman

Precisely. Historically what we do is we sold electricity at the wrong price. Utilities got results that they didn’t like for whatever reasons — too high a generation cost, transmission or distribution issues. Then we created these complicated, expensive mechanisms like VPPs to account for the fact that we charge at the wrong price. Those are expensive bandages on the system or like digital paint on the old analog system. What this is, if you charge the right price to begin with, then you get the result you wanted. You don’t need the band-aid, you don’t have a wound that you need to address because you’re not creating the wound by charging the wrong price.

In other contexts we just charge the right price. Airlines charge the right price. They don’t sell their tickets at a flat price and then have third parties go in and try to convince people to change their flight times to even out the supply and demand of airplane seats. They just charge the right price to begin with.

David Roberts

This is the analogy that finally occurred to me. I was thinking about how to explain this to normal people. This is the analogy that I think works: imagine if airlines sold all seats on all flights for the same flat price. They would get terrible results from that. Some flights would be massively oversubscribed, some undersubscribed. Imagine a third party came along, rounded up all the ticket holders and said, “Hey, the airline’s getting terrible results. There’s some value to the airline of you being flexible. Can you move your flight, individual customers? Can you move your flight?” Then that third party rounded up all the customers, took their flexibility, and sold it to the airlines. That is the VPP model.

Bruce Nordman

Exactly. It’s complicated and it’s expensive.

David Roberts

What you are saying is just don’t charge that dumb flat price in the first place. If you charge prices that reflect temporal and geographic value, you will get the results you want and then you won’t have to do this ad hoc flexibility stuff afterward. Is that fair? The airline analogy — you think that works?

Bruce Nordman

Yeah. The key difference in the last few years is that we now have all of the technologies at very low cost, often free, to make the automatic control highly practical. We have the communication protocol — we had communication protocols for pricing for the last 20 years, but they were old and cumbersome and difficult to deal with. Now we have modern, simple ones. We have pervasive Internet connectivity so that devices can get the prices. We have extremely low-cost computation on integrated circuits. The integrated circuits that go into appliances cost four or five dollars.

David Roberts

This is worth emphasizing. This is all cheap, often free. The benefit of all these electronic developments over the last 30 years is this is all just basic circuits and communication stuff that is trivially cheap. It’s not some big, fancy, complicated stuff we’re talking about. Your water heater just needs to ping the Internet, and that is a solved problem.

Bruce Nordman

And ping your local price server in your central building device to get the price.

David Roberts

Who is then pinging the central price server.

Bruce Nordman

The reason many appliances have Wi-Fi on them today — when you think, “Does my dishwasher need to have Wi-Fi?” — no, it doesn’t, but the microprocessors have Wi-Fi and Bluetooth on the chip whether they are used or not. It is literally free for them to include it. You can’t get much less expensive than free. I will soon have three price-responsive devices today covering my space heat, water heat, and EV charging. Also, my phone optimized it to a GHG, all of it through free software upgrades from the manufacturer for devices that were not grid responsive when I bought them.

David Roberts

No kidding. You bought non-grid responsive appliances and they just had the chips in them that were capable of Wi-Fi communication. Because they included them trivially regardless, a software update from the manufacturer then turned on its ability to communicate with the web.

Bruce Nordman

I should be more specific here. My space heat and hot water come from a heat pump which the heat pump can’t communicate with. But I have a clever controller commercial product. When I bought it, it optimized to time-of-use prices in the manufacturer’s cloud. I then got the free software update where it optimizes to hourly prices in the controller in my house and it receives those prices that were OpenADR3. That was a free software update. The EV charger, when I got it, optimizes to TOU prices in the cloud and they are working on integrating with the PG&E price server to be able to use the hourly prices and optimize in their cloud to it. That PG&E price server uses OpenADR3.

David Roberts

Really most of what you are talking about in this system is software tweaks. There is just not a lot of hardware, certainly not expensive hardware, involved in any of this. Most of this is software, correct?

Bruce Nordman

For any new product or any product that you can get an over-the-air software update, an increasing number, this is pretty trivial. There is this long transition time where lots of people have old water heaters. Maybe it’s an electric resistance water heater. There are commercial products today that will control them and are Internet connected and could readily operate to an hourly price just as well as to the TOU price they’re operating to now.

David Roberts

There’s no barrier to making literally any appliance price responsive. Even if it’s dumb and old, you can get a controller to attach to it to make it —

Bruce Nordman

Well, a dumb device generally only works if you can depower it. For a resistance water heater you can just shut off the power to shift its load. I’m sure there are some devices where that’s not a workable option. Some devices will need replacement. Many devices can have these external controls which do cost something. Doing anything costs something. To have something be part of a VPP, it has to be able to communicate and that’s all you need for the price response. You don’t need anything more for price response than you need for VPPs, you just avoid a whole lot of cost and complexity.

David Roberts

Let’s talk about how to get from here to there, which is part of what’s broken my brain about all this. This is a very wholesale change in the way the electricity system works. Just to review, what we need to make this work is a central price server, a protocol that helps communicate, and price-responsive devices. We have price-responsive devices. That’s a solved technical problem. We have the protocol. That’s a solved technical problem. We know how to make a price server. That’s technologically trivial. Technologically speaking, this is all good to go.

It’s not technology development that is the barrier here. Is the road in this direction — does it start with utilities changing how they charge prices? Just with utilities implementing more dynamic prices? Is that step one here?

Bruce Nordman

That’s step one. It could be either a unitary utility acting on its own, or it could be a utility forced by a regulator, such as in Illinois where the legislature passed a law saying they must offer hourly prices different every day, and they’ve had it for 10 years.

David Roberts

Oh really? Illinois has highly dynamic prices of the type you are talking about.

Bruce Nordman

Yes, but the variation over the course of the day isn’t very much. Your opportunity to save is limited. It’s there, but it’s not that great. They have only very limited automation. They don’t have a price server in the way that we’re talking about. Having the prices is a prerequisite for everything else, the prices and the price server, because it’s not that difficult for manufacturers to put in the ability to be price responsive. In general, they are reluctant to spend time doing so until they have a critical mass of customers who could actually use them.

I’ve had this conversation with many manufacturers over many years and they say, “I’m not going to innovate into a vacuum. I need to have a business case. I need to have the customers and the customers need to be able to pay the prices.” Everything comes down to the prices. It’s not a chicken-and-egg thing. The prices have to come first.

David Roberts

Yeah, but skate to where the puck is going. Everything is going to talk to the Internet eventually. Just go ahead and do it.

Bruce Nordman

Then we also need to get global consensus on the maximum complexity of the tariffs as it affects device optimization. Just to be clear, you can have fixed charges which do not affect device optimization. But the parts of the tariff that affect how devices optimize to reduce bills the most need to be limited to the stream of import prices and stream of export prices. People need to recognize that needs to become the global standard so that a manufacturer in China, shipping products all over the world, can ship products that will work all over the world because everything works the same way.

David Roberts

I see. You want to standardize price format to make this replicable and interoperable across different utilities, different countries —

Bruce Nordman

Everywhere! We’ve done this with IT technology. It’s not that we haven’t done this globally before. We just need to get a critical mass of organizations that operate according to these principles and that other people will see the merit of it and then be consistent with it. The question is how do we get to that initial critical mass? If you take the principles that things should be simple and universal, you’re 80% of the way towards having the right answer. Some utilities will think they’re special and different. You and I know they are special and they aren’t different in this respect.

How you calculate the price can be different everywhere because that doesn’t change the communication. This is like web browsing where you have the web server, HTML, and web browsers, and HTML defines the maximum complexity of the web page. This price structure defines the maximum complexity of how prices should be done on a retail basis.

David Roberts

Got it. It’s a high priority for you to keep this simple, to keep the pricing differentiation as simple as possible.

Bruce Nordman

Exactly.

David Roberts

Import price, export price, but it is going to change hour to hour. It’s not simple in that respect. It’s changing temporally all the time.

Bruce Nordman

It can be a day. The vast majority of people who pay such prices, there are day-ahead prices and that is absolutely the place to start. Once people have the automation, they will realize that you could shift from day-ahead to day-of prices. The communication, the automation do not change and that is going to result in a slightly lower bill. This will happen later on, but we just have to know that the devices do not care if it is day-ahead or day-of.

The key thing here is that we don’t want to put digital paint onto our 19th-century electricity technology. We need electricity technology initially just inside of buildings and then later into the utility grid — which is digitally native and networked — instead of using network technology to manage our 19th-century systems.

David Roberts

This is a very big set of ideas and topics, a very big idea. Do you have a next thing that you are pushing for? You said we can step toward these incrementally. What’s next on your personal list that you would like to see happen moving in this direction?

Bruce Nordman

Aside from every electricity retailer offering hourly prices that are different every day or something that is beyond flexibility?

David Roberts

That’s a big step, Bruce. Much more practically, earthbound — is there a particular utility you’re trying to get a pilot thing started with? Very practically, what needs to happen?

Bruce Nordman

PG&E offers hourly prices different every day to any customer type today and has for more than a year. Southern California Edison is supposed to soon and the state of California’s Energy Commission policy is that every Californian should have access to such prices in less than a year. I don’t think we will make that deadline, but we are trying to get there.

David Roberts

That’s the price server that you’re talking about. Making them available to everyone is just about setting up the price.

Bruce Nordman

It’s about offering the prices and having the price server to communicate them. In some places they have the prices but they don’t have the price server.

David Roberts

Let’s stipulate in our fantasies, two years from now, California has the dynamic prices, has the server. Then what is left to do for Californians, individual consumers, is just to buy the price-responsive devices. Everything will be in place except for the consumers.

Bruce Nordman

In many cases people will get devices that can get software updates to do it or they can acquire a central controller that can do it. When they replace a product, they will replace it with one that does this. Any cloud-connected device can certainly do this and just optimize in the cloud. There are lots of paths for many existing devices to become price responsive very quickly and then it will take time for most of the devices to turn over and get replaced.

David Roberts

Long story short, California is moving in this direction and there is reason to believe that in some amount of time, two years, five years, this is how the California electricity system is going to work.

Bruce Nordman

Yes. We also plan to still have VPPs and use them both. The balance of it should be determined by the market and the need. There’s always going to be some — we will make good use of VPPs for the foreseeable future because we won’t turn everything over instantly, there’ll be just a shift in the balance over time. There are some limited services that VPPs are better at. It’s not that they don’t have a role, it’ll just be changing.

I was thinking the other day that it’s like hybrid cars. Hybrid cars are sort of like VPPs in that they’re kind of a band-aid solution. Hybrid cars are actually growing, as I understand it, even as electric vehicles are also growing. Over time, eventually the EVs will grow and start taking up and the hybrids will decline over time. That’s, I think, a very good way to think about it.

David Roberts

The point I’m trying to make here, and what I want listeners to take away, is that this is not just Bruce’s elaborate fantasy that he’s concocted. This is a thing that’s happening in the world. There are systems moving in this direction, and this is something that could happen within our lifetimes. We could see a system working.

Bruce Nordman

Millions of people in at least 20 countries pay such prices today, unfortunately most without the right automation. The automation is easy to set up. The barriers are all in people’s heads and institutional. They are not technological.

David Roberts

Maybe this is a hard question to answer, but among the nerds who think about stuff like this, the class of nerds in the world who are beavering away investigating electrical systems and thinking about how to reform them, is there a broad consensus behind this basic vision or is it a battle? Is it a debate? Are you an outlier? Are you a minority? What’s the state of opinion in Electricity World about this? Are a lot of people captured by this and want to move this way, or do you feel like you’re fighting the tide?

Bruce Nordman

I feel like I’m fighting the tide, but there are lots of reasons to be optimistic, at least in the long term. More and more places are offering more dynamic prices. The automation of their use — people haven’t figured out that that’s an essential part of this. The desire for privacy and autonomy certainly is only growing. The desire for lower bills is growing. Both of those are facilitated by this.

One problem here is that there’s no group of companies that are making lots of money off of this, that are a trade association lobbying for this, because the customer saves the money. There’s not these companies pulling money out. That’s part of why this is so good. With VPPs you’ve got tons of venture capital, so they can have big PR budgets and get lots of attention to this, whereas there’s no industry behind this for pricing because it’s all about saving money rather than about extracting money from people.

David Roberts

Yes, it is a value shift in the direction of customers and as a cynic, you would expect that not to be hugely popular among the entities who were previously receiving that value. What you really need is consumer groups to get behind this. This gets back to what I was saying earlier — consumer advocates, when they hear “let’s just throw everybody out on the mercy of the market and use nothing but prices,” have negative connotations around that. There is a lot of communications work to be done with the consumer field about this, about why it is consumer friendly.

Bruce Nordman

We’re already at the mercy of the market. We’re already paying all of the revenue requirement for utilities. All this does is reduce how much revenue they need. In a way it’s a path towards lower bills — it’s not a path towards higher bills. That’s the key.

David Roberts

I know that, you know that. But I don’t think ordinary people when they hear pricing necessarily think consumer first. Which is just to say that there’s a lot of translation and communication work to be done here to help people understand why this is a consumer-friendly reform and not rapacious capitalism or whatever.

Bruce Nordman

Internet technology developed from researchers and academics and government labs. It wasn’t in the market where it all developed because it had business backing until it reached scale and then people figured out that they could have businesses around it, but the development wasn’t driven by those business interests. That’s the problem here — we don’t have business interests driving the development of this, even though once it exists people will realize there is lots of opportunity to change existing products or introduce new products.

David Roberts

Obviously, we’ve barely scratched the surface. People can go to your website, got all kinds of papers digging into this further, all the ins and outs. One thing I just leave listeners with is, and this I think is a good provocative question you ask, which is: if not prices, then what? As you say, VPPs are good, but customer sites are going to get very complicated and trying to micromanage 10,000, 50,000 customer sites from a central location is just going to get impractical quickly. We need a coordination and optimization mechanism that can operate locally, that can operate at different scales, that can operate off grid or on grid, et cetera. If not pricing, then what is that going to be?

Bruce Nordman

Exactly. I’ve looked for 20 years and I’ve never found a second mechanism. Let’s suppose that the utility sees their transformers getting overloaded and wants to turn down my EV charger, but I really want to charge my EV and I’ve got a stationary battery I could discharge. They shouldn’t turn down my EV. They should just say, “You can’t import as much power,” in which case I’ll just discharge it from the battery and keep charging my EV. All the decisions need to be made locally because the information to make the right decision is local. The customer’s preferences and economic impacts all occur locally.

David Roberts

The customer is in control here, in the driver’s seat here, getting all the value of flexibility, in control of the preferences, in control of the operation of their devices. One thing I like about this vision is that you have solved the privacy problems at a stroke here and the enshittification problem too. I think this is going to avoid that too, because the utilities are not capturing you, the VPPs are not capturing you. You are running your own stuff.

Bruce Nordman

Because it’s platforms that enshittify, it’s not technologies. You’re not required to be part of any platform. With this, it’s only about technology.

David Roberts

I’ll leave it there. Good Lord. Thank you to any listener who’s still with us. Thank you, Bruce, for walking us through this again. People can go to brucenordman.com and learn a lot more about all this. Thanks, Bruce.

Bruce Nordman

Thank you so much. I would love to hear from other people who have questions or want to help out with this.

David Roberts

Thank you for listening to Volts. It takes a village to make this podcast work. Shout out especially to my super producer, Kyle McDonald, who makes me and my guests sound smart every week. It is all supported entirely by listeners like you. If you value conversations like this, please consider joining our community of paid subscribers at volts.wtf, leaving a nice review, telling a friend about Volts, or all three.

Thanks so much, and I’ll see you next time.

Fin-Tech: How Sharks Could Sharpen Ocean Forecasts

Via Anthropocene, an interesting look at how sensors strapped to 19 sharks off America’s east coast cut errors in a leading climate model by as much as 43%:

The vast ocean dwarfs our efforts to understand it. Sensor-laden buoys, high-flying satellites and sophisticated computer models can only do so much to plumb the depths of the waters covering more than two-thirds of the planet.

But a creature with intimate knowledge of the ocean might help humans get a more accurate picture of what lies beneath. Sharks could serve as mobile, wide-ranging sensor systems, collecting data that improves our understanding of ocean conditions in ways that might inform fisheries management and other critical activities, according to new research in the journal npj Climate and Atmospheric Science.

“Sharks are already moving through parts of the ocean that are challenging for us to observe,” said Laura McDonnell, the lead author and a postdoctoral scientist at the Woods Hole Oceanographic Institution (WHOI) in Massachusetts. “This research shows that data they collect can help fill important gaps.”

Scientists have attached sensors to sharks for years, but usually with the intent of understanding what’s going on with the animals. In 2022 I spent five days on the Atlantic Ocean near Africa with a team of scientists catching sharks and drilling holes in dorsal fins to attach light-bulb-sized sensors. They wanted to know how the sharks’ behavior changed as they swam through patches of low-oxygen water.

Neil Hammerschlag, a co-author of the new paper, was using sensors in much the same way as a marine ecologist at the University of Miami (UM) when, in 2018, he spoke with UM atmospheric scientist Ben Kirtman about the possibility of using data from the sensors to study the ocean, rather than the fish.

“Marine predators like sharks naturally seek out dynamic ocean features such as fronts and eddies,” explained Kirtman. “These are areas where models often lack sufficient observations.”

As a Ph.D. student at UM, McDonnell took on the question of whether this might work.  In waters off the Northeast U.S. coast, McDonnell and colleagues attached sensors to the dorsal fins of 18 blue sharks and one shortfin mako shark in October 2021. Then they set them loose, like so many fast-moving drones.

In the following months, the equipment decorating the sharks’ fins collected moment-by-moment measurements of temperature and depth, two pieces of data critical to understanding the state of the ocean in a particular place. When the animals surfaced, the tags transmitted the information to satellite and on to the scientists. All told, they collected more than 8,200 snapshots of ocean conditions from the sharks. While the data was concentrated off the coast of the eastern U.S. north of Virginia, the sharks roamed as far south as Florida and out into the middle of the Atlantic. They also gave scientists glimpses of conditions at depths of almost 2,000 meters as they dove.

The researchers took this trove of information and used it to fine tune a computer program commonly used to model current ocean conditions based on data from ocean-going buoys and other sources. In certain parts of the ocean, the shark-enhanced approach was significantly closer to reality than the standard model, when scientists tested to see how well the models simulated ocean conditions during the time when the sharks were collecting the data. (The ability for computer models to accurately reconstruct past conditions is a standard test for ocean and atmospheric models.)

The improved performance was particularly notable along the shallow continental shelf, where it reduced the model’s error by 43% for November and 33% in December. That added up to the model being around 1.5°C closer to the mark when it came to sea surface temperatures, a significant improvement in an environment where subtle temperature shifts can drive major ecological changes.

 “For fisheries and coastal communities, small improvements in ocean forecasts can make a big difference,” said Camrin Braun, an oceanographer at WHOI who worked on the study. “Reducing uncertainty helps people plan, whether that’s where to fish, how to manage resources, or how to respond to changing conditions.”

That doesn’t mean sharks will be replacing other data-gathering, cautioned McDonnell. This was only a short-term experiment, and there is no mention of a more comprehensive effort to enlist sharks to the front lines of ocean forecasting. But it does show that tags formerly used to just understand the sharks could do double duty by shining more light onto broader mysteries of how the ocean is changing.

McDonnell, et. al. “Improved seasonal climate forecasting using shark-borne sensor data in a dynamic ocean.” npj Climate and Atmospheric Science. April 28, 2026.

Internet of Animals: How Tracking Animal Movement May Save the Planet

Via MIT’s Technology Review, a look at how researchers have been dreaming of an Internet of Animals. They’re getting closer to monitoring 100,000 creatures—and revealing hidden facets of our shared world.

There was something strange about the way the sharks were moving between the islands of the Bahamas.

Tiger sharks tend to hug the shoreline, explains marine biologist Austin Gallagher, but when he began tagging the 1,000-pound animals with satellite transmitters in 2016, he discovered that these predators turned away from it, toward two ancient underwater hills made of sand and coral fragments that stretch out 300 miles toward Cuba. They were spending a lot of time “crisscrossing, making highly tortuous, convoluted movements” to be near them, Gallagher says. 

It wasn’t immediately clear what attracted sharks to the area: while satellite images clearly showed the subsea terrain, they didn’t pick up anything out of the ordinary. It was only when Gallagher and his colleagues attached 360-degree cameras to the animals that they were able to confirm what they were so drawn to: vast, previously unseen seagrass meadows—a biodiverse habitat that offered a smorgasbord of prey.   

The discovery did more than solve a minor mystery of animal behavior. Using the data they gathered from the sharks, the researchers were able to map an expanse of seagrass stretching across 93,000 square kilometers of Caribbean seabed—extending the total known global seagrass coverage by more than 40%, according to a study Gallagher’s team published in 2022. This revelation could have huge implications for efforts to protect threatened marine ecosystems—seagrass meadows are a nursery for one-fifth of key fish stocks and habitats for endangered marine species—and also for all of us above the waves, as seagrasses can capture carbon up to 35 times faster than tropical rainforests. 

Animals have long been able to offer unique insights about the natural world around us, acting as organic sensors picking up phenomena that remain invisible to humans. More than 100 years ago, leeches signaled storms ahead by slithering out of the water; canaries warned of looming catastrophe in coal mines until the 1980s; and mollusks that close when exposed to toxic substances are still used to trigger alarms in municipal water systems in Minneapolis and Poland. 

a tiger shark seen underwater with a camera on its flank
Attaching 360-degree cameras to tiger sharks helped demystify the animals’ strange movements around the Bahamas.
COURTESY OF BENEATH THE WAVES

These days, we have more insight into animal behavior than ever before thanks to sensor tags, which have helped researchers answer key questions about globe-spanning migrations and the sometimes hard-to-reach places animals visit along the way. In turn, tagged animals have increasingly become partners in scientific discovery and planetary monitoring.

But the data we gather from these animals still adds up to only a relatively narrow slice of the whole picture. Results are often confined to silos, and for many years tags were big and expensive, suitable only for a handful of animal species—like tiger sharks—that are powerful (or large) enough to transport them. 

This is beginning to change. Researchers are asking: What will we find if we follow even the smallest animals? What if we could monitor a sample of all the world’s wildlife to see how different species’ lives intersect? What could we learn from a big-data system of animal movement, continuously monitoring how creatures big and small adapt to the world around us? It may be, some researchers believe, a vital tool in the effort to save our increasingly crisis-plagued planet. 

Wearables for the wild

Just a few years ago, a project called ICARUS seemed ready to start answering the big questions about animal movement. 

A team led by Martin Wikelski, a director at the Max Planck Institute of Animal Behavior in southern Germany and a pioneer in the field, launched a new generation of affordable and lightweight GPS sensors that could be worn by animals as small as songbirds, fish, and rodents. 

These Fitbits for wild creatures, to use Wikelski’s analogy, could produce live location data accurate to a few meters and simultaneously allow scientists to monitor animals’ heart rates, body heat, and sudden movements, plus the temperature, humidity, and air pressure in their surroundings. The signals they transmitted would be received by a three-meter antenna affixed to the International Space Station—the result of a €50 million investment from the German Aerospace Centre and the Russian Space Agency—and beamed down to a data bank on Earth, producing a map of the animals’ paths in close to real time as they crisscrossed the globe.

Wikelski and his peers hoped the project, formally the International Cooperation for Animal Research Using Space, would provide insights about a much wider variety of animals than they’d previously been able to track. It also aimed to show proof of concept for Wikelski’s dream of the past several decades: the Internet of Animals—a big-data system that monitors and analyzes animal behavior to help us understand the planet and predict the future of the environment.

Researchers have been laying the groundwork for years, connecting disparate data sets on animal movement, the environment, and weather and analyzing them with the help of AI and automated analytics. But Wikelski had his sights on something even grander and more comprehensive: a dashboard in which 100,000 sensor-tagged animals could be simultaneously monitored as near-real-time data flowed in from Earth-imaging satellites and ground-based sources. 

By bringing together each of these snapshots of animals’ lives, we might begin to understand the forces that are shaping life across the planet. The project had the potential to help us better understand and conserve the world’s most vulnerable species, showing how animals are responding to the challenges of climate change and ecosystem loss. It also promised another way to monitor the Earth itself during a period of increasing instability, transforming our animal co-inhabitants into sentinels of a changing world. 

When ICARUS first went into space in 2018, it was widely celebrated in the press. Yet what should have been a moment of glory for Wikelski and the field of animal ecology instead became a test of his will. The ICARUS antenna first went down for a year because of a technical issue; it went back up but was only just out of testing in February 2022 when the Russian invasion of Ukraine halted the project altogether.

Wikelski and his peers, though, have used the time since to innovate and evangelize. They now envision a more complete and technologically advanced version of the Internet of Animals than the one they hoped to build even just a few years ago, thanks to innovations in tracking technologies and AI and satellite systems. They have made even smaller and cheaper sensors and found a new, more affordable way to work in space with microsatellites called CubeSats. Their efforts have even gotten NASA to invest its time and resources into the possibility of building the Internet of Animals.

Now Wikelski and his collaborators are again on the verge, with an experimental CubeSat successfully transmitting data as part of a testing phase that started last June. If all goes as planned, another fully operational ICARUS CubeSat will begin collecting data next year, with more launches to follow. 

The potential benefits of this system are extraordinary and still not yet fully understood, says Scott Yanco, a researcher in movement ecology at the University of Michigan. Perhaps it could help prevent mountain lion attacks or warn about a zoonotic disease about to make a jump to humans. It could alert researchers of behavioral changes that seem to happen in some animals before earthquakes, a phenomenon Wikelski has studied, and determine what conditions tell boobies in the Indo-Pacific to lay fewer eggs in years before strong El Niños or signal to weaver birds in the Niger Delta to build their nests higher up before floods. 

“You can talk to 100 scientists about this,” Yanco says, “and they’re all going to give you a different answer of what they’re interested in.”

But first, a lot still needs to go right. 

Animals as sentinels

When I first spoke with Wikelski, in early 2022, ICARUS was live, tracking 46 species from the ISS 400 kilometers overhead. Wearing a pair of square-rimmed glasses and speaking in a German accent with a tone of unfailing urgency, he was excited to tell me about a tagged blackbird who made a 1,000-or-so-kilometer crossing from Belarus to Albania. 

That was actually pretty routine, Wikelski said, but almost everything else he had been seeing over the past year of road-testing had been stranger than expected. White storks were crossing back and forth over the Sahara five times a season, without apparent reason. Cuckoos, which are tree-dwelling birds ill suited to long periods at sea, were making uninterrupted journeys from India to the Horn of Africa. “Now, any time you look, totally novel aspects appear, and novel connections appear across continents,” he told me.

This could have been a mystifying mess. But for Wikelski, it was “beautiful data.” 

The practice of tagging animals to monitor their movements has been used for more than 100 years, though it began with a stroke of luck. In the 1820s, a hunter in a village in central Africa threw a 30-inch spear that lodged itself nonfatally in the neck of a white stork. This became what might have been the world’s first tag on a wild animal, says Yanco: the bird somehow flew back to Germany in the spring, helping settle the mystery of where storks disappeared to in the winter. 

By the 1890s, scientists had started tracking wild birds with bands fitted around their legs—but 49 out of every 50 ring-tagged birds were never seen again. Starting in the 1960s, thousands of birds received very-high-frequency radio tags known as “pingers,” but these were only powerful enough to broadcast a few kilometers. To capture the data, researchers had to embark on cartoonish chase scenes, in which tagged birds were pursued by an oversize homing antenna pointed out the roof of a car, plane, or hang-glider. 

Wikelski tried all three. During a stint at the University of Illinois in Urbana-Champaign in the mid-’90s, he was studying thrushes and would gun an Oldsmobile around the Midwest at over 70 miles per hour. He’d set off as the songbirds got going at around 2 a.m., which tended to draw the attention of local police. Wikelski found that contrary to the conventional wisdom, thrushes used just 29% of their energy on their overnight migrations, less than they expended hunting and sheltering during stopovers. But the hassle of his process, which also entailed capturing and recapturing birds to weigh them, convinced Wikelski that, among other things, he needed better tools.

Thinking bigger (and higher) 

It was not immediately clear that the solution to Wikelski’s problems would be in space, though the idea of tracking animals via satellite had been explored decades before his Oldsmobile experiments. 

In fact, NASA invented space-based animal tracking back in 1970 when it strapped a transmitter collar the weight of two bowling balls around the neck of Monique the Space Elk, a local news celebrity at the time. (Monique was actually two elks: the anointed Monique, who wore a dummy collar for testing and press photos, and another, who accidentally caught a misfired tranquilizer dart and subsequently got the satellite transmitter collar.) After the Moniques met untimely deaths—one from starvation, the other at the hands of a hunter—the project went dormant too. 

But its research lived on in Argos, a weather monitoring system established in 1978 by the National Oceanic and Atmospheric Administration (NOAA) and the French space agency. It pioneered a way to track a tagged animal’s location by beaming up a short stream of analog data and measuring wave compression—the so-called Doppler shift—as a polar-­orbiting satellite zoomed overhead at thousands of miles an hour. But this captured locations to only a few hundred meters, at best, and typically required a clear line of sight between tag and satellite—a challenge when working with animals below the canopy of rainforests, for instance. 

Wikelski worked extensively with Argos but found that the technology didn’t enable him to capture the highly detailed whole-life data he craved. By the late ’90s, he was on an island in Panama, exploring an alternative approach that followed hundreds of animals from 38 species, including small mammals and insects. 

Using six long-distance radio towers, Wikelski and Roland Kays, now the director of the Biodiversity Laboratory at the North Carolina Museum of Natural Sciences, started to develop the Automated Radio Telemetry System (ARTS), a radio collar tracking system that could penetrate thick canopy. Crucially, ARTS revealed interactions between species—for example, how predatory ocelots support the island’s palm trees by eating large quantities of rabbit-like agoutis, after the rodents bury palm seeds underground as a snack for later. The researchers also found that despite what everyone believed, many of the animal inhabitants don’t remain on the island year-round, but frequently travel to the mainland. Kays and Wikelski had demonstrated in microcosm the kinds of insights that fine-grained multispecies tracking could provide even in challenging environments.

But Wikelski was frustrated that he couldn’t follow animals off the map. “If we don’t know the fate of an animal, we will never be able to really do good biology,” he says. The only solution would be to have a map with no edge. 

This was around the time that GPS trackers became small enough to be used in animal tags. While radio tags like those used by Argos estimated location by transmitting signals to receivers, GPS systems like those in cars download data from three or more satellites to triangulate location precisely. 

Wikelski became a man possessed by the idea of using this technology to create a truly global animal monitoring system. He envisioned digital tags that could capture GPS data throughout the day and upload packets of data to satellites that would periodically pass overhead. This idea would generate both excitement and a lot of skepticism. Peers told Wikelski that his dream system was unrealistic and unworkable.

At the turn of the millennium, he took a position at Princeton with the notion that the institutional pedigree might earn an audience for his “crazy” idea. Not long after he arrived, the chief of NASA’s Jet Propulsion Laboratory came for a talk, and Wikelski asked whether the agency would benefit from a satellite system that could track birds. “He looked at me as if I came from a different planet,” Wikelski remembers. Still, he got a meeting with NASA—though he says he was laughed out of the building. By this time, the agency had apparently forgotten all about Monique. 

Undeterred, in 2002 Wikelski launched ICARUS, a half-joke (for fans of Greek mythology) at his own immodest ambitions. It aimed to use digital GPS tags and satellites that would relay the information to a data center on Earth nearly as instantly as the ARTS system had.

Wikelski’s big ideas continued to run into big doubts. “At the time, people told us technology-wise, it will never work,” he says. Even 10 years ago, when Wikelski was making proposals to space agencies, he was told to avoid digital tech altogether in favor of tried-and-tested Argos-style communication. “Don’t go digital!” he recalls people telling him. “This is completely impossible! You have to do it analog.” 

Moving away from the fringe

In the two decades since ICARUS was established, the scientific community has caught up, thanks to developments in consumer tech. The Internet of Things made two-way digital communications with small devices viable, while lithium batteries have shrunk to sizes that more animals can carry and smartphones have made low-cost GPS and accelerometers increasingly available.

“We’re going from where we couldn’t really track most vertebrate species on the planet to flipping it. We’re now able to track most things,” says Yanco, emphasizing that this is possible “to varying degrees of accuracy and resolution.” 

The other key advance has been in data systems, and in particular the growth of Movebank, a central repository of animal tracking data that was developed from Wikelski’s ARTS system. Movebank brings together terrestrial-animal tracking data from various streams, including location data from the Argos system and from new high-res digital satellites, like ICARUS’s antenna on the ISS. (There are also plans to incorporate CubeSat data.) To date, it has collected 6 billion data points from more than 1,400 species, tracking animals’ full life cycles in ways that Wikelski once could only dream about. It is now a key part of the plumbing of the animal internet. 

The field also had some practical successes, which in turn allowed it to marshal additional resources. In 2016 in London, for instance, where air pollution was responsible for nearly 10,000 human deaths a year, researchers from Imperial College and the tech startup Plume Labs released 10 racing pigeons equipped with sensors for nitrogen dioxide and ozone emissions from traffic. Daily updates (tweeted out by the Pigeon Air Patrol account) showed how taking a pigeon’s path through the neighborhoods revealed pollution hot spots that weather stations missed.

Diego Ellis Soto, a NASA research fellow and a Yale PhD candidate studying animal ecology, highlights an experiment from 2018: flocks of storks were outfitted with high-resolution GPS collars to monitor the air movements they encountered over the open ocean. Tagged storks were able to capture live data on turbulence, which can be notoriously hard for airlines to predict.

Among the critical roles for these animal sensors was one that was once considered eccentric: predicting weather and the world’s fast-changing climate patterns. Animals equipped with temperature and pressure sensors essentially act as free-roaming weather buoys that can beam out readings from areas underserved by weather stations, including polar regions, small islands, and much of the Global South. Satellites struggle to measure many environmental variables, including ocean temperatures, which can also be prohibitively expensive for drones to collect. “Eighty percent of all measurements in Antarctica of sea surface temperature are collected by elephant seals, and not by robots or icebreakers,” Ellis Soto says. “These seals can just swim underneath the ice and [do] stuff that robots can’t do.” The seals are now tagged yearly, and the data they collect helps refine weather models that predict El Niño and sea-level rise.

When the ICARUS antenna was installed on the ISS in August 2018, it seemed poised to unlock even more capabilities and discoveries. In the antenna’s short life, the project recorded the movements of bats, birds, and antelope in near-real time, from Alaska to the islands of Papua New Guinea, and transferred the data to Movebank. But when the experiment ground to a premature halt, Wikelski knew he’d have to do something different, and he concocted a plan by which ICARUS could continue—whether it could rely on a major space agency or not.

Another shot

Rather than a system of major satellites, the new incarnation of ICARUS will run on CubeSats: low-cost, off-the-shelf microsatellites launched into low Earth orbit (around the same height as the ISS) for around $800,000, meaning even developing nations that harbor space ambitions can be part of the project. CubeSats also offer the benefit of truly global coverage; the ISS’s orbital path means it can’t pick up signals from polar regions further north than southern Sweden or further south than the tip of Chile.

There’s currently one ICARUS CubeSat in testing, having launched into orbit last summer. If all goes well, a CubeSat funded by the Max Planck Society, in collaboration with the University of the Bundeswehr Munich, will launch next April, followed by another in winter 2025, and—they’re hoping—another in 2026. Each further addition allows the tags to upload once more per day, increasing the temporal resolution and bringing the system closer to truly real-time tracking. 

map of the earth with flight pattern of tracked birds shown in red

MAXCINE/MAX PLANCK INSTITUTE OF ANIMAL BEHAVIOR

Outfitting even small animals with lightweight, inexpensive GPS sensors, like the one on this blackbird, and monitoring how they move around the world could provide insights into the global effects of climate change.

Wikelski and his partners have also rededicated themselves to making even smaller tags. They’re close to the goal of getting them down to three grams, which would in theory make it possible to track more than half of mammal species and around two-fifths of birds, plus hundreds of species of crocodiles, turtles, and lizards. ICARUS’s tags are also now cheaper (costing just $150) and smarter. ICARUS developed AI-on-chip systems that can reduce the energy use by orders of magnitude to cut down on the size of batteries, Wikelski explains. There are also new tags being tested by scientists from the University of Copenhagen and Wikelski’s institute at Max Planck that harvest energy from animal movements, like a self-winding wristwatch. Finally, these new ICARUS sensors can also be reprogrammed remotely, thanks to their two-way Internet of Things–style communications. A new ecosystem of tag makers—professional and DIY—is further driving down prices, open-sourcing innovation, and allowing experimentation. 

Still, not everyone has bought into ICARUS. Critics question the costs compared with those of existing terrestrial monitoring initiatives like MOTUS, a national Canadian bird conservation program that uses a network of 750 receiving towers. Others argue that researchers can make better use of the thousands of animals already tracked by Argos, which is upgrading to more accurate tags and is also set to launch a series of CubeSats. The total cost of a fully realized ICARUS system—100,000 animals at any one time, some of which die or disappear as new ones are tagged—is around $10 million to $15 million a year. “If you’re thinking about how to tag a moose or bighorn sheep, you might need to hire a helicopter and the whole team and the vet,” says Ellis Soto, who has long collaborated with Wikelski. “So the costs can be extremely, extremely limiting.” 

But, proponents argue, the initiative would beget a lot more information than other Earth-imaging space missions and be significantly cheaper than sending humans or drones to collect data from remote locations like polar ice sheets. Wikelski also emphasizes that no one entity will bear the cost. He is working with local communities in Bhutan, South Africa, Thailand, China, Russia, and Nigeria and gets requests from people across the world who want to connect tags to ICARUS. With cheap satellites and cheap tags, he sees a route to scale. 

Even as ICARUS explores a grassroots future, one of the biggest changes since the initial launch is the backing Internet of Animals technology has received from the biggest giant in the field: NASA. The agency is now two years into a five-year project to explore how it might get more involved in building out such a system. “We’re very much focused on developing future mission concepts that will come after the current set of ICARUS missions,” says Ryan Pavlick, a researcher in remote sensing of biodiversity at NASA’s Jet Propulsion Laboratory. In 2024, this will mean “architecture studies” that aim to understand what technical systems might meet the animal-tracking needs of stakeholders including NOAA, the US Fish and Wildlife Service, and the United States Geological Survey. 

While NASA’s project aims to deliver benefits for the American people, a fully realized Internet of Animals would necessarily be global and interspecies. When we spoke in November 2023, Wikelski had just got off the phone discussing how ICARUS can help monitor the global “deal for nature” established by the UN’s COP15 biodiversity conference, whose targets include reducing extinction rates by a factor of 10. 

Jill Deppe, who leads the National Audubon Society’s Migratory Bird Initiative, has boundless enthusiasm for how an Internet of Animals could affect organizations like hers. For a century, Audubon has watched migratory birds disappear on journeys to Chile or Colombia. A system that could tell us where birds are dying across the entire Western Hemisphere would allow Audubon to precisely target investments in habitat protection and efforts to address threats, she says.  

“Our on-the-ground conservation work is all done on a local scale,” says Deppe. For migratory birds, ICARUS can link these isolated moments into a storyline that spans continents: “How do all of those factors and processes interact? And what does that mean for the birds’ survival?”

Movebank’s live-updating dashboard also makes more dynamic conservation action possible. Beaches can be closed as exhausted shorebirds land, wind farms can halt turbines as bats migrate through, and conservation-conscious farmers—who already aim to flood fields or drain them at times that suit migrating flocks—can do so with real knowledge. 

In return, will animals really help us see the future of the planet’s climate? 

No one is suggesting that animals take over from the system of satellites, weather stations, balloons, and ocean buoys that currently feed into meteorologists’ complex models. Yet technology that complements these dependable data streams, that captures the ever-changing biological signals of seals, storks, sharks, and other species, is already starting to fill in gaps in our knowledge. Once considered cryptic signs from the fates, or harbingers of doom, their behaviors are messages that have only just begun to show us ways to live on a changing planet. 

Uzbekistan Wants Nuclear Energy, But Can It Afford the Water Cost?

Courtesy of The Diplomat, a look at how Uzbekistan is betting on nuclear power to secure its energy future – but it is doing so in one of the most water-stressed regions in the world

The Uzbek government’s plan to build a nuclear power plant (NPP) in the Jizzakh region, alongside the creation of a centralized radioactive waste system, marks a decisive shift toward long-term nuclear infrastructure. Officials have presented the project, implemented in partnership with Russia and under the oversight of the International Atomic Energy Agency, as a rational response to rising electricity demand and necessary to further economic growth.

Yet beneath this explanation lies a difficult question: can a water-intensive energy system, which nuclear power is, be sustained in a region where water demand already exceeds supply?

According to a United Nations report, water demand in Uzbekistan exceeds supply by 23 percent and overall water stress has reached 123 percent – a sign that scarcity is no longer a future risk, but a present reality.

A Strategic Response – or a Strategic Bet?

Uzbekistan’s turn to nuclear energy reflects mounting structural pressures. Electricity demand across Central Asia is projected to grow by around 40 percent by 2030, requiring at least 62.8 gigawatts of new capacity.

Political analyst Elyor Usmanov frames the NPP project as a structural response to a widening energy gap, driven by industrial growth, population expansion, and the declining flexibility of gas-based generation.

The project also promises measurable gains. Official estimates suggest the NPP could supply up to 15 percent of Uzbekistan’s electricity while saving roughly 3.6 billion cubic meters of natural gas annually.

But nuclear energy is not a short-term solution. It is a long-term commitment. Infrastructure of this scale operates for decades, embedding financial and technological dependence. As Usmanov notes, such projects act as a “long-term anchor” – economic, technological, and political.

This dependence is already taking shape. Agreements linked to Uzbekistan’s nuclear program are valued at up to $24.7 billion, underscoring both the scale of investment and the depth of long-term commitments. 

The government has approved the creation of a national radioactive waste management center, responsible for transporting, processing, and storing radioactive materials over the long term. The project also includes infrastructure upgrades and the development of specialized laboratories between 2026 and 2027.

As Andrey Ozharovsky, a nuclear physicist and co-founder of the public program Radioactive Waste Safety notes, radioactive materials remain hazardous for decades – and in some cases centuries – requiring continuous monitoring and stable institutional capacity.

In this sense, nuclear energy extends far beyond electricity generation, creating a long-term obligation to manage risk across generations.

A Region Under Pressure

These developments are unfolding in a region already under severe water stress.

Across Central Asia, water availability per capita has declined from around 8,400 cubic meters to approximately 2,500, and could fall to 1,700 by 2030 – a threshold associated with chronic scarcity.

At the same time, inefficiencies compound the problem: up to 40 percent of water is lost through outdated irrigation systems, while agriculture consumes roughly 80 percent of available resources.

Climate change is intensifying these pressures. Under high warming scenarios, up to 80 percent of the region’s glaciers could disappear, further reducing already limited water supplies.

At the same time, experts increasingly note that the region’s water crisis is not just by scarcity, but also by inefficiencies in management, outdated infrastructure, and the lack of coordinated regional governance. 

Nuclear energy introduces a structural tension into this already fragile system. Nuclear power plants require substantial volumes of water for cooling. Globally, they rank among the most water-intensive forms of energy generation.

This raises a broader question: are large-scale infrastructure decisions are being fully aligned with long-term resource constraints? 

Location is critical. Uzbekistan’s planned nuclear site lies near the Aydar-Arnasay lake system, including Lake Tuzkan, a hydrologically sensitive area.

Water expert Bulat Yessekin notes that such a facility could consume more than 70 million cubic meters of water annually, an amount comparable to a large city.

He warns that this demand, combined with thermal pollution, could accelerate ecosystem degradation, particularly in the Aral Sea basin, which has already lost up to 92 percent of its water volume.

Technological solutions such as dry cooling can reduce water use – in some cases by up to 90 percent – but they do not eliminate the underlying constraint.

The implications of Uzbekistan’s nuclear power ambitions extend beyond environmental concerns. In Central Asia, water is central to both economic stability and regional security. 

According to Elyor Usmanov, a core challenge lies in the absence of integrated water-energy planning. Without it, infrastructure designed to solve one problem may intensify another. This is particularly relevant in a region defined by transboundary rivers. Projections suggest that new infrastructure projects, including Uzbekistan’s NPP and Afghanistan’s Qosh-Tepa Canal, could reduce the flow of the Amu Darya by 8 to 20 percent.

Afghanistan, which historically withdrew only limited volumes of water from the Amu Darya basin, is increasingly asserting its claim to a share of these resources. The country is, at present, in a water crisis.  Officials in Kabul emphasize that their position is grounded in the legitimate right to equitable use of transboundary waters, while also expressing readiness to cooperate with neighboring states.

As Afghan representatives have stated, the country’s objective is “to receive only what it is entitled to,” framing its water policy not as a challenge to regional stability, but as part of a broader push for balanced and lawful resource distribution.

An Array of Risks 

Beyond water consumption, nuclear infrastructure introduces long-term environmental and health risks. 

Ozharovsky warns that the most serious dangers arise when radioactive materials enter water systems, where they can lead to prolonged internal exposure through drinking water and food chains.

A large-scale study in the United States (2000–2018) identified approximately 115,000 cancer deaths – or around 6,400 annually – statistically associated with proximity to nuclear power plants.

While such findings remain debated, they underscore the importance of long-term monitoring, transparency, and risk governance.

Even where probabilities are low, the consequences of inadvertent exposure or accidents can be severe.

In a worst-case scenario, damage to reactor systems could release radioactive elements such as iodine, cesium, and strontium, with contamination spreading across hundreds of kilometers. In today’s geopolitical environment, nuclear facilities may also be viewed as strategic vulnerabilities.

The economic dimension of nuclear energy is equally uncertain, as are the full geopolitical repercussions of Uzbekistan’s NPP.

Globally, large-scale nuclear projects often face delays and cost overruns. In one case, projected costs rose from $9.8 billion to $25 billion during implementation.

In some instances, projects are abandoned altogether, leaving financial burdens that ultimately fall on taxpayers. Russia’s role extends beyond construction to include fuel supply, maintenance, and long-term operation. This creates a structural dependence, one that must be actively managed through diversification and the development of domestic expertise. These kinds of arrangements arrangements, while common in nuclear energy projects, may also limit flexibility over time, particularly in areas such as fuel supply, maintenance, and technological upgrades.

Uzbekistan’s nuclear ambitions reflect a rational response to real challenges: rising demand, aging infrastructure, and economic transformation. But nuclear energy is not a neutral solution.

In a region where water availability has already fallen from 8,400 to 2,500 cubic meters per capita – and continues to decline – the introduction of a water-intensive energy system creates a structural tension.

The question is not whether Uzbekistan needs more energy – it does – but whether Tashkent’s seeking of long-term energy stability via nuclear power can be achieved without intensifying the region’s unfolding water crisis.

Soy Republics

Via Phenomenal World, a look at corporate concentration and the far right in South America:

Early this year, in the Brazilian Amazon, a coalition of 14 indigenous groups rose upagainst a government decree that planned to privatize the waterway between Itaituba and Santarém—two towns in the state of Pará—as well as a public project to dredge the River Tapajós. The giant commodities trader Cargill controls ports in both towns that act as key nodes of the logistics infrastructure through which soy grown in the rainforest and savanna is sent to global markets, mainly in Europe and Northern Africa. In 2022, around 13 percent of all soy exported by Cargill from Brazil was shipped from Santarém, including almost two-thirds of the beans grown in the rainforest.

The groups launched the protest by camping in front of a Cargill terminal in Santarém and demanding that the government cancel both the privatization of the river and its dredging. The communities claimed that the plans, on which they hadn’t been consulted, posed a direct threat both to their livelihoods and to the rich biodiversity of the region, risking the release of harmful mercury from the riverbed. At its height, the camp had more than a thousand people. After just over a month, the government agreed to meet their demands. The reaction from the mainstream press was one of shock and horror. 

The episode brings to the fore the critical role played by multinational corporations in conflicts around the Amazon. By looking at the South American soy boom in the biome and beyond, we can begin to grasp the ways in which corporate concentration in one of the region’s main export activities has contributed to both the rise of a far-right agribusiness bloc and popular mobilizations in defense of the environment. While one of my previous columns discussed how policies oriented towards “land-sparing” helped to create the conditions for bolsonarismo, here I will discuss another crucial factor in South America’s political evolution: the power relations within supply chains.

Capital in the Amazon

Global capital’s interest in the rainforest is nothing new. In the 1920s, for instance, Henry Ford bought a massive piece of land in the biome which he named Fordlândia and hoped to transform into a plantation of rubber trees from which to source the inputs for the tyres of his mass-produced cars. Yet the relative significance of foreign investment has fluctuated over time. Critical scholars studying the region have generally claimed that it would be misleading to make it a central focus, since the Amazon’s main problems have flowed from domestic political and economic factors. “The role of international capital in producing deforestation in the Amazon has been relatively minor,” wrote Susanna Hecht and Alexander Cockburn in The Fate of the Forest in 1990. Hecht later took aim at attempts to connect deforestation to the international demand for beef. ‘‘The so-called hamburger connection,’’ she argued, ‘‘simply does not operate in the current Amazonian context.’’ (The literature on this question, from that period, was summarized by Andrew Hurrell.)

In the last three decades, however, the situation has changed beyond recognition. The rearticulation of the world economy around China made South America a key provider of primary commodities to global production, and transformed the role of transnational corporations in the region—including in the rainforest. There have been a number of sophisticated efforts to map this phenomenon empirically. A recent estimate indicates that 56 percent of Amazon deforestation driven by soy croplands between 2020 and 2022 can be attributed to international, rather than domestic, consumption. In an article from 2018, Victor Galaz and his co-authors argued that the main drivers of land-use change in the rainforest are soy and beef production, and that these two economic activities are dominated by only 8 corporations: 4 giant grain traders (ADM, Bunge, Cargill, and Louis Dreyfus, usually referred to as ABCD), the largest private soybean producer in the world (a Brazilian firm, Amaggi), and 3 Brazilian meat-processing firms (JBS, Marfrig, and Minerva). Galaz et al. also showed that many of these companies are tightly linked to the 16 financial firms most involved in activities threatening biomes critical for the Earth’s climate, and especially to the ‘‘big three’’ asset managers (BlackRock, Vanguard, and State Street, all headquartered in the US).

The connection of ADM, Bunge and Minerva to these 16 financial firms is mainly through stockownership, whereas in the case of JBS and Marfrig, ‘‘investors’ latent influence’’ operates predominantly through debt. (The three remaining firms—Amaggi, Cargill, Louis Dreyfus—are privately owned, which means that the data on their intertwinement with financial capital is limited.) The authors therefore arrive at a blunt conclusion: “the ‘Financial Giants,’ through their common blockholding power, have a previously ignored, yet considerable potential influence in companies shaping biomes critical for the stability of the climate system.’’ Last year, researchers from University College London and University of Exeter went further still, tracing the financial flows—including lending and equity and debt issuance—towards 24 corporations, including the 8 firms mentioned above, which have been ‘‘linked to significant land use change and degradation’’ in the Brazilian Amazon between 2014 and 2023.

Academics have also made occasional efforts to get these transnational corporations to change their ways. In 2019, a group of 21 prominent researchers proposed ‘‘expanding the focus from ‘corporate social responsibility’ to ‘corporate biosphere stewardship,’’’ advocating ‘‘a new business logic with the purpose of shepherding and safeguarding the resilience of the biosphere for human well-being.’’ Others appealed to central banks and financial regulators, arguing that it is their mandate to deal with ecosystem degradation that ‘‘poses escalating systemic risks to economic and financial systems.’’ To restrict a ‘‘substantial portion’’ of financial flows to companies that threaten the tipping points of the Brazilian Amazon and the Indonesian peatlands, they asserted, ‘‘would only require coordination across relatively few financial centres.’’ 

The hope implicit in this literature echoes Rudolf Hilferding’s remark that the concentration brought about by finance capital could be seized upon to overcome capitalism itself—‘‘taking possession of six large Berlin banks,’’ he wrote in 1910 in Finance Capital, ‘‘would mean taking possession of the most important spheres of large-scale industry, and would greatly facilitate the initial phases of socialist policy”—except the goal here is no longer to bring about socialism, but simply to salvage a liveable planet by preventing ecosystem tipping points from being reached. 

Can’t beat them? Join them

The expectation that central banks might come to the rescue in a deepening climate crisis was evidently grounded in an ephemeral conjuncture, which came to an end with Donald Trump’s return to the White House last year. Yet even if this push had succeeded, it may have had difficult side effects. As Yannis Dafermos argued, ‘‘attempts of private finance to protect itself from climate risks’’—and attempts by monetary authorities to push them to do so—may exacerbate climate finance injustice by increasing borrowing costs faced by Global South countries and deepening their financial vulnerability.

Research on the role of global capital in the climate crisis must therefore consider the impacts of corporate concentration on country-level politics, investigating the alliances built by transnational corporations with domestic elites and the potential resistance they may encounter. Seen through this lens, the case of the South American soy boom offers useful insights. Since the late 2010s South America has been responsible for more than half of global soy production and almost two-thirds of global exports of the commodity. In 2024, Argentina, Brazil and Paraguay accounted for 61 percent of all exports.

The data made available by Trase.earth is clear about the dominance of the giant grain traders in Brazilian and Argentine soy exports, as shown in the figures below. (Data includes soybeans and “the raw equivalent of the traded sub-products,” that is, ‘‘soy cake and soy oil are converted to soybean equivalent tonnes”.) Statistics for Brazil start in 2004, when the ABCD firms responded for 48.2 percent of total soy exports, in tonnes. By 2009, their share had peaked at 59.8 percent, before gradually falling over the following years, to 42.2 percent in 2022—the most recent information available. Despite the decline, the four companies remained in control of a sizable share of a booming business: in 2022, exports totalled 93.7 tonnes of soy, up from 36.2 in 2004. The data for Argentina is limited to the period between 2015 and 2019 and shows that the share of the ABCD firms oscillated between a quarter and a third of the total. Adding a fifth giant commodity trader, Glencore, brings this figure above 40 percent.

Founded either in the 19th or in the early 20th century, the ABCD firms saw their global influence increase at the turn of the millennium in parallel with the expansion of future trading in commodities. Jennifer Clapp argued that this financialization of the commodity trade allowed them to “wield enormous power in shaping food systems.” Profiting from volatility that caused hunger and popular rebellions throughout the world, these corporations were known, according to Gustavo de Oliveira and Mindi Schneider, for “rerouting cargo ships mid-ocean to gain marginal profits on large volumes, and speculating on futures markets with the privileged information that results from controlling significant shares of non-transparent markets.” 

In South America, their dominance was established in the 1990s through the surge in acquisitions that formed part of the process of neoliberalization. By extending the subordination of soy farmers, they set in motion a cumulative process that further tightened the companies’ grip. They squeezed prices paid to producers and so pushed them to combine intensive use of herbicides with genetically-modified seeds, “in order not to incur price deductions at the point of delivery, deepening the technological treadmill and the farmers’ need for finance.” By the early 2000s, they had come to control“the entire transportation logistics…as well as the port terminals, cargo ships and processing facilities that ultimately crushed the soybeans into meal and vegetable oil.”

South American big capital did not put up much resistance to this trend. In Brazil, Amaggi maintained its powerful position, controlling more than 8 percent of soy exports from Brazil in 2022 and around 1.6 percent of the soy croplands in the mid-2010s, partly by coordinating with the global traders. It established joint ventures with Bunge and Dreyfus and obtained funding from Mitsui (a large Japanese trader). It also had significant political influence. Its owner, Blairo Maggi, was governor of the main soy growing state in Brazil (Mato Grosso) between 2003 and 2010, senator between 2011 and 2016, and Minister of Agriculture between 2016 and 2019.

In Argentina, taxes on exports of unprocessed soybeans, strengthened large domestic agro-industrial capitalists who produced soy oil, allowing them to keep a share of total exports similar to the one of the ABCD firms (see figure above). The three largest Argentine corporations (Vicentin, AGD, and Perez Companc) controlled from a quarter to a third of total exports, between 2015 and 2019. (Embroiled in a series of scandals after 2019, Vicentin would eventually be rescued by an Argentine businessman with the support of Cargill.)

Chinese state capital posed a more significant challenge. Following the so-called “2004 soybean crisis” in China, when a sudden change in the global price of the commodity pushed a number of Chinese processing firms into bankruptcy, the ABCD firms saw their control of the Chinese market surge. In response, the government prioritized the growth of Chinese traders, especially COFCO, in a dispute dubbed the “battle of the beans.” (See Tomaz Fares’ detailed account.) The ripples were eventually felt in South America: in Brazil, COFCO started exporting soy in 2005 and, since 2014, became responsible for a share between 5 and 8 percent of total exports; in Argentina, its share was even higher, around 10 percent, between 2015 and 2019. A substantial part of the post-2009 decline in the share controlled by ABCD in Brazil is explained by the rise of COFCO. Yet the shift in market shares may exaggerate the actual challenge. Being treated initially with “particular hostility” from the ABCD, who aimed “to maintain their oligopolistic control over soybean exports,” COFCO was forced to build alliances with the incumbents, reaching a preferential agreement with ADM. If you can’t beat them, join them.

Corporate concentration and South American politics

The rise of COFCO is unlikely to bring effective competition to the industry, simply expanding from four to five the firms at the helm of the oligopsony of South American soy. Given their focus on processing, marketing, and logistics, the real material process of growing soy is still done by a multitude of farmers, large and small. A tiny minority, usually backed by foreign and domestic financial capital, owns vast tracts of cropland and manages to retain some autonomy relative to the traders. Estimates with data from the mid-2010s indicated that 7 giant farm management companies—pools de siembra, as they are called in Argentina—controlled almost 7 per cent of all hectares planted with soy across Argentina, Brazil, and Paraguay. In Brazil alone, 5 of them (including Amaggi) accounted for about 5 per cent of the soy cropland area.

Yet these are the exceptions, giant as they are. The vast majority of the soy farmers are significantly smaller, including large numbers of medium- and smallholders, who due to their size are deeply subordinated to the trading oligopsony. They are not only price takers but have lost most of the control of their own operations. “The nature and ‘autonomy of farming’ are increasingly disciplined and structured by external forms of management based on the application of technological packages and farm-service logistics,” Oliveira and Hecht wrote. In Argentina, this transformation of farming is referred to as sojización, characterized “by higher levels of capitalization, mechanization, foreign investment and economies of scale.”

The political constituency represented by this multitude of farmers, connected to the most dynamic sector of the region’s economies, could in principle have been mobilized against the oligopsony, demanding a different organization of the supply chain. But instead, subordinated rural groups, suspicious of the center-left governments that presided over the consolidation of the trading oligopsony, turned sharply to the right. In Brazil, these radicalized soy farmers rose through the ranks of state-level soy producer associations to hegemonize rural politics for the country as a whole, providing Bolsonaro with a key electoral base. This marked the birth of what Caio Pompeia calls agri-bolsonarismo. 

A similar development can be identified in Argentina. In 2008, when the government of Cristina Kirchner moved to increase taxes on the exports of soy, it was resoundingly defeated by a large coalition in a struggle that is now referred to as the crisis del campo, which Diana Córdoba and her co-authors described as a “rural protest of unprecedented scale in which disparate and historically fractured rural interests united in a common protest that lasted for months.” They went on to explain how

Landowners, contractors, workers, hauliers and other rural actors—representing the rural elite, smallholders and the rural working class—participated together, using trucks and farm equipment to install blockades on rural routes and cutting off the movement of agricultural (and other) products throughout the core agricultural region.

Some have argued that the origins of Argentina’s rightward movement—first with the election of Mauricio Macri in 2015, then with the victory of Javier Milei in 2023—can best be found in this crisis of 2008. In both Argentina and Brazil, then, the economic transformation led by the giant grain traders increased the countries’ foreign vulnerability at the same time as it sowed the seeds of anti-democratic politics. These examples from recent South American history reveal that corporate concentration in the global food industrial system not only threatens food security and fuels climate change, but also weakens democratic institutions. When the indigenous groups faced down Cargill earlier this year, it was not only their livelihoods that were at stake. 

PlantNet: A “Shazam” for Plants

Via The Conversation, a report on a new citizen science nature app that’s geared towards the scientific community:

The Conversation: What can you tell us about PlantNet users?

Pierre Bonnet and Alexis Joly: An impact study carried out a few years ago identified that 12% of users used the app for work, either for research, land management, farming or teaching purposes. The large majority of users use PlantNet out of curiosity or personal interest.

Looking at who and where PlantNet users are based reflects the way technology is used in different parts of the world. In Asia, we have quite a few young users, because they are the most connected. We can also observe that the Chinese flora, which is nevertheless very rich, is poorly represented by the users of the application, and this is explained in particular by the fact that non-state or non-Chinese platforms, which are not carried by Chinese companies or partners, are much less present on this market.

Tropical parts of the world are very biodiversity-rich, is PlantNet particularly big there?

P.B. and A.J.: Brazil is in eighth place with just over 600,000 annual users. Indonesia and India are in the top 20. Currently, the bulk of PlantNet’s activity is still in Europe and North America. Several things explain this. PlantNet has already been launched in France and in Europe, and has therefore been further promoted and attracted media attention in this region of the world. Backed by user demand, the application was also initially adapted to the French and Mediterranean flora before it was gradually extended to include other European flora, then North American, and tropical.

It should also be taken into account that in tropical areas, species richness is certainly far greater, but access capacities are often more limited. The road network is less developed; 3G connectivity can be poor, particularly in forested mountain areas.

In the field, plant biodiversity can also be more complex to photograph, take for example, the many tropical plants that are epiphytic that is to say, that grow on another plant, especially at canopy level. When we talk about trees that are several tens of metres high, it immediately becomes more complex to photograph.

All this makes tropical plants and flowers considerably less known. The application covers almost 100% of European flora, compared with coverage of a few tens of percent for tropical countries. This is something that is not unique to PlantNet, and generally applies to all institutional databases, especially since covering tropical areas costs more.

But despite this, we are working with partners in Costa Rica, Guyana, Brazil, Cameroon, Madagascar and Malaysia, in particular to extend our coverage of the number of species.

In the tropics or elsewhere, what can be noted about the plants that users are looking to identify? Are the most common plants the most in demand or not necessarily?

P.B. and A.J.: There is necessarily a correlation, because very rare plants are necessarily rarely observed. But we also note that certain plants are very common, but of little interest, because they are “discreet”. These can be crop weeds, plants found by the roadside but which do not have noticeable flowers, which are pollinated by the wind with slender, greenish flowers, such as species of fescue (Festuca spp.), or bromes (Bromus spp.). They are less observed because they are actually less attractive. We can also note that tree searches are popular, whereas certain herbaceous or epiphytic plants tend to be extremely rare.

People actually often observe plants that they like, or plants that seem useful to them. Incidentally, we get a lot of requests about fruits, berries, and plums, probably because people want to know if they are edible or not.

But the goal of tracking useful plants for humans was not the main objective of the PlantNet project from the get-go. We had to adapt it, however, to meet expectations in terms of this type of use while remaining fairly moderate regarding the volume of information we provide.

At the same time, we are working more and more with people who study animal health, either in human health and who use the PlantNet service or data for their own work. Like for instance, people who worked for ToxiPlant which identifies plants that are toxic for horses. We also regularly consult doctors on different uses of PlantNet for identifying allergenic plants, especially those that cause skin allergies. We also liaise with the French regional agency for monitoring air quality ATMO Occitanie, which uses shared data on flowering plants listed through PlantNet that it integrates into its air quality estimation models indicating pollen counts.

Are there certain plants that would benefit from being photographed more?

P.B. and A.J.: Yes. There are plants that interest us, or our partners, but for which we have very little data. These include allergenic plants for respiratory allergies, such as male junipers, which release pollen when their cones open.

These cones are very discreet. Junipers are therefore photographed, but rarely with illustrations that show the development stage of the cones. Seeing as they cause allergies, some of our partners would like to collect more data on the subject. We hope to remedy this through animation features, either directly through the platform or through social media, to generate interest in collecting data on this type of plant or on rare, endangered plants or plants that have conservation issues, such as Marsilea strigosa Willd. and Arenaria provincialis Chater & P. Halliday (native to the South-eastern France).

We also have the case of certain plants that interest us for agriculture, weeds, for example, which we have barely identified at the stage of young shoots, such as ‘Imperatacylindrica_ (L.) Raeusch. or that Galium aparine L.

Your application encourages users to take several different pictures. This can be photos of flowers, leaves, fruits, bark, or of the entire plant, for example, to help them identify plants. What type of data do you have the most of?

P.B. and A.J.: Data on flowers, especially in their blooming seasons – during spring and early summer. Flowers attract interest and trigger observation. They are also the most effective visually, and have typically been used a lot by botanists in the past.

On PlantNet, after flowers, we notice a decreasing degree of performance for fruit, leaves, and then twigs and bark, which are sometimes more difficult to take pictures of, even if users are always encouraged to combine several criteria (flower and fruit, flower and leaf…) each time to maximise their chance of correctly identifying a plant.

What happens when PlantNet can’t match a photograph with an existing species?

P.B. and A.J.: Troubleshooting “no results found” is still a tricky business for all forms of AI, and PlantNet is no exception. Some species are very poorly represented, and it is very difficult to teach the model behind the app to differentiate between something that is very rare in the learning database and something that we do not have at all. We therefore seek to quantify the uncertainty, rather than decide when the model is uncertain or not. We are working with our team on creating confidence intervals. That is why, in some cases, the application will provide several species.

One thing that makes this work difficult is that new species often resemble existing species. A new species is thought to be very picturesque, but it isn’t always the case, hence the confusion.

There is also the issue of hybrid, ornamental plants. You will always come across new plant creations. We cover them, but not as well as other plants, even though we are trying to overcome this difficulty.

We also set out to simply identify more and more plants. Today PlantNet covers 85,000 species out of an estimated 400,000, with 2,000 new species being discovered each year. These discoveries are made by taxonomists worldwide and via the World Flora Online network, which brings together several dozen universities, herbaria and botanical gardens, which PlantNet joined in 2025.

Working closely with these networks will enable us to improve cases where PlantNet performs poorly, particularly when new species emerge thanks to the research carried out by the taxonomists, who actually further knowledge by dividing what was previously thought as a single species, but which represents, in fact, several of them or, conversely, by grouping together what were thought to be different species, but which turns out in reality to be only one plant.

Let’s come back to the 10% of users who use PlantNet for work. Who are they? What line of work are they in?

P.B. and A.J.: PlantNet’s data is very useful for building species distribution models predicting where a given species is commonly found. There are certainly biases in our data, depending on where people do and don’t go, but they can help us to gain a better understanding of the factors that influence these distributions, including climate change. This provides answers to questions like, “will species change habitats or stay put?” or “what is the human impact on species distribution?” Many ecologists download the PlantNet data and couple it with other data for species distribution modelling.

PlantNet data can also help in the detection of invasive species that are often notable for their size, density, visual appearance, or novelty, such as Carpobrotus edulis (L.) N.E.Br or Mirabilis jalapa L. We are working on this subject with the French Biodiversity Office, which is interested in using plant monitoring cameras to detect the presence of invasive plants, in order to contain their expansion within the areas where they are found.

We have an ongoing project that focuses on improving the classification of species that are present in farming environments and recognition of plant pathogens whether they are viruses, bacteria, fungi…

Have you discovered any unusual uses of the app that have surprised you?

P.B. and A.J.: Perhaps the most surprising was a Dutch museum which used PlantNet to identify plants featured in the paintings it had on display.

Other surprises have included people using PlantNet to identify a tattoo of a plant, or a plant-inspired pattern of an oilcloth… along with other more fun uses of the app like on the Explorama or Geo Quest apps which use our automated identification service.

PlantNet supports the diversity of possible uses, not by trying to integrate them, but rather by sharing its plant-identifying services. We have more than 20,000 accounts that have incorporated PlantNet’s service into their own mobile or web application, or in their data processing workflow.

And what other uses for PlantNet could be developed?

P.B. and A.J.: We have processed just over 1.3 billion plant ID requests. Among this data, there is a lot of material for characterising environments and species… However, the data is still difficult to use due to its sheer volume. These photos most probably contain new species and data on species that do not exist elsewhere. Photos users have posted provide potentially interesting information about the associated communities (not just in the foreground, i.e.; the plant photographed, but in the background) and about plants’ surrounding environments. While it’s not what the photo primarily sets out to do, it does offer potentially very interesting data on plant associations, that for the moment, remain untapped.

Little use has been made of our data to study the impact of current, fairly fast-moving climate change. PlantNet has, in this respect, collected extensive data over the last five to ten years that could allow us to gain greater understanding on how plants react to such swift environmental changes. From one year to the next, there may be a lot of variability, but for now, it is difficult to measure this impact.

What can PlantNet users do to help further research going forward?

P.B. and A.J.: Creating a user profile significantly increases the benefits for research. By creating an account, people agree to terms of use and facilitate re-exploitation for research. The more qualitative the data is, the more relevant it is to research activities. Geolocation is, for example, very valuable to us. It is also beneficial for users with a potentially higher level of determination.

We encourage researchers around the world to feel free to explore the full potential of PlantNet, whether it be via our shared data or the services we offer.

The Limits of the UAE’s Push for Food Security

Via New Lines Magazine, a look at how – as Iran blocks the Strait of Hormuz – the Gulf state’s strategic investments across Africa face a reality check:

Deep in the heart of the Sahara Desert, an industrial complex of greenhouses rises above the dunes. Running underneath the sand is a 35-mile-long pipeline transporting water to the property in this remote part of Mauritania.

The owners? A firm from the United Arab Emirates. The goal? Farming blueberries.

In recent years, companies from the UAE have invested heavily in farming projects across Africa. There are currently 56 projects spanning millions of acres from Sudan to the project in Mauritania, which I reported from last year as part of a year-long investigation into the UAE’s agriculture acquisitions in Africa.

Despite their country’s staggering wealth, UAE officials have long known that food is the country’s Achilles’ heel. High temperatures, lack of water and infertile soil make growing most crops exceedingly challenging, meaning the UAE has to import about 90% of its food. Increasingly, Emirati agribusiness has looked to Africa to reduce its anxiety about food shortages, quietly preparing for distant climate shocks and conflicts by investing $11.9 billion in East African agriculture alone since 2009, according to the Africa Center for Strategic Studies.

Now, as the war with Iran enters its third week, the entire strategy is being tested and cast into doubt. Iran has blocked the vast majority of ships from transiting through the Strait of Hormuz, the narrow strip of water between the UAE and Iran, through which 90% of the UAE’s food supply has historically flowed. The country holds six months of grain and staple stockpiles, but the blockage threatens fruit and vegetable supplies that depend on continuous imports, and experts warn that food shortages will occur if the war continues.

But the otherworldly scene at the farm in Mauritania revealed the extent of the UAE’s food challenges — even outside of war. The state-backed Emirati firm had said it was going to grow blueberries. But there were no blueberries in sight.

Why was one desert nation trying to grow fruit in another desert nation, over 6,000 miles away? Over the course of the past year, I’ve been investigating Emirati projects in Africa, seeking to understand why they are investing in remote sites where other companies would not dare, pushing through vast losses and, sometimes, allegations of land and water grabs.

Some farming investors say it is simply naivete, but others say the effort to transform arid land into arable is part of a long-standing Emirati tradition of projecting power and ingenuity, which in recent years has been a central part of its foreign policy effort to expand influence in Africa and beyond.

“The UAE is all about soft power,” said Naser Alsayed, a Middle East environmental expert at the think tank Chatham House. “When one state-owned entity invests, it doesn’t come alone.”

Before oil, Emiratis mainly depended on pearling — with divers plunging to depths of 6o feet to retrieve the shellfish and their valuable gems — and small-scale, oasis-fed agriculture. But as oil revenue filled government coffers from the 1960s, the first president of the newly formed UAE, Sheikh Zayed bin Sultan Al Nahyan, plowed vast funds into transforming the desert for modern, large-scale agriculture. “Give me agriculture, and I will give you civilization,” the founding father is often quoted as saying. Cynthia Gharios, a scholar at the University of Munster, says these projects were less about food security and more about projecting an image to the world of sovereignty and modernity — selling a country where oil wealth and ingenuity could transcend ecological limits.

But desert farming soon exhausted precious water supplies, and the state opted to abandon it. The idea of projecting Emirati sophistication through extreme farming was etched into the minds of elites, however, and the strategy shifted to purchasing land abroad, first experimenting in Sudan. This further crystallized during the 2008 price crisis, when food shortages rocked the UAE. In 2018, the UAE’s national food strategy announced it would become number one in the global food security index, an ambitious feat for a nation importing 90% of its food. As part of this, its footprint in Africa expanded into new areas — Egypt, Ethiopia, Mauritania, Morocco, Angola, Uganda, Zimbabwe.

But Emirati companies, eager to avoid accusations of neocolonial land grabbing and resource exploitation, promised to leverage their experience in hostile environment farming.

“We go to remote areas away from farmers, so they don’t say that we have taken their food or land,” Ahmed al-Falasi, a board member of Jenaan, an Emirati company with 250,000 acres across Egypt, Sudan and Ethiopia, told Reuters.

When it came to blueberries, the claims made by Elite Agro, a firm from Abu Dhabi with investments in Morocco, Egypt and Ethiopia, were big. On a government contract, the group said it would grow over 10,000 acres. If true, that would have made it the biggest single farm in Mauritania.

I wanted to see it. So, armed only with the name of a village in southern Mauritania, I set off through the arid landscape, occasionally interrupted by sparse, wispy acacia trees.

Many of these investments are shrouded in mystery. The estimate of 56 operations was put together by Land Matrix, a registry for large-scale land acquisitions, but it acknowledges that remote farms with a lack of company transparency make a true accounting of Emirati deals impossible. Access for journalists ranges from limited to nonexistent, adding a further challenge for anyone wanting to understand the reality behind the glitzy pictures of Emirati and African officials signing deals in boardrooms.

Under closer inspection, these estimates flatter the Emiratis. I’ve read countless documents and articles promising vast investments in Sierra Leone, Angola, Kenya and Tanzania, and nothing follows from them.

Even when deals progress into real-life production, it’s no guarantee that large-scale projects live up to the promises. In northern Senegal, I visited a company growing livestock fodder to export to the Gulf. Unpaid employees and rusting irrigation pipes above barren land were all that was left when the company went bust after one year. Herders and farmers still aren’t able to use the land.

“It’s the waste that hurts me,” said Doudou Ndiaye Mboup, a former electrician for African Agriculture. This is what people in the business of land prospecting call a zombie deal. It’s the worst possible outcome — no food for the Gulf, no food or employment for locals, while depriving them of land to grow food or raise livestock.

“Businesspeople walk away, but communities suffer,” said Rene Velvee, co-founder of the nonprofit Grain, which campaigns for land rights.

In Sudan, Land Matrix estimates that companies from across the Gulf have acquired 1.9 million acres, but many are failing to produce crops while blocking locals from using the land, according to analyses. One satellite review of large-scale land acquisitions estimates that only 4.2% of large-scale investments are producing food. Another study says 24% fail outright.

When I arrived in the village in Mauritania, no one — farmers, police, the mayor’s office — knew anything about this farm. The deputy mayor, dressed in flowing white robes, took pity on me and made some calls. Thirty minutes later, I was on my way to the alleged location, two hours north.

We turned off Mauritania’s only highway, where the deputy mayor said it was only a few more miles till our destination. After one hour down this gravel road, increasingly despondent about my chances, a tractor emerged out of a cloud of dust.

A rugged man stepped out of the cab, his head wrapped in a headscarf, curious about an unexpected visitor. As I asked him about the company, he looked at me vacantly. Only when I mentioned the Emiratis did a toothy grin appear, and he ordered me to follow his pickup. He zigzagged over sand dunes, navigated through herders and their cattle, and out of nowhere an industrial complex of greenhouses emerged from the desert.

The farm existed, but no blueberries were growing. The fine dust from the Sahara spoiled the delicate fruit, and the company had to grow hardier watermelons instead. Of the 10,000 acres leased, the greenhouses occupy barely 1%, the majority standing empty.

Experts and food investors greet the idea of growing blueberries and watermelons in the Mauritanian desert with astonishment, even ridicule. It doesn’t even contribute to the UAE’s food security, because the watermelons are exported to the U.K. and Europe.

An employee from Al Dahra, a sovereign wealth fund-backed agribusiness with 22,000 acres in the military-controlled Egyptian desert, says crops are sold on the global market, not transported back to the UAE.

It poses a question: What motivations lie behind such risky investments by companies claiming to pursue Emirati food security? The real harvest is influence, not food, said Martin Keurlitz, assistant professor at the American University of Beirut.

“Agriculture is a wonderful diplomacy tool,” Keurlitz explained. Farming projects, ostensibly promising benefits to the poor, are part of a wider strategy to forge stronger relations with elites of stronger nations. Projects don’t even need to materialize to be useful. Even an announcement of a deal is a foot in the door to more lucrative and influential industries — mining, logistics, energy.

Husam Mahjoub, the founder of Sudan Bukra, told me that just because farming projects in Sudan sit empty, it doesn’t mean the ventures were a failure. By investing in farmland, Emirati officials and business executives could see the inner workings of Sudanese politics. Gold exports from Sudan, worth more than $1 billion in 2023, now dwarf the food trade.

But this influence has a dark side, especially in Sudan. It’s widely understood that the Emiratis funded the insurgent Rapid Support Forces in Sudan, who have driven instability for decades, displacing 9.5 million people and committing atrocities in El Fasher late last year.

Presumed to be a bastion of safety in the Middle East, the Iran war will fundamentally reshape the UAE’s priorities at home and abroad. Gulf officials are already considering invoking force majeure on overseas investment contracts to ease financial pressures caused by the war.

The impact of increased food security concerns on risky agribusiness investments in Africa remains to be seen. Alsayed from Chatham House said the UAE should focus on expanding its already extensive logistics and port networks across the continent.

“The smartest strategy of their plan is the focus on ports,” Zayed said. Control of ports is control of food supplies, he added.

Meanwhile, Al Dahra, the sovereign wealth fund-owned agribusiness, shows no sign of slowing down. In January, the firm announced that it sought to become the world’s largest farming company, aiming to reach a footprint of 1.2 million acres worldwide. In February, it announced a deal of 45,000 acres in Tanzania.

As the Strait of Hormuz remains closed for the foreseeable future and food shipments are blocked, the UAE’s African breadbasket strategy is of little use — and would be even if it were more successful and less experimental. Despite the greenhouses in Mauritania and land in Egypt, the UAE is still a prisoner of geography, relying on fragile supply chains. One day, UAE food security and the metamorphosis of the African desert into lush green fields may be a reality. For now, it’s a mirage.