You’re out of free articles.
Log in
To continue reading, log in to your account.
Create a Free Account
To unlock more free articles, please create a free account.
Sign In or Create an Account.
By continuing, you agree to the Terms of Service and acknowledge our Privacy Policy
Welcome to Heatmap
Thank you for registering with Heatmap. Climate change is one of the greatest challenges of our lives, a force reshaping our economy, our politics, and our culture. We hope to be your trusted, friendly, and insightful guide to that transformation. Please enjoy your free articles. You can check your profile here .
subscribe to get Unlimited access
Offer for a Heatmap News Unlimited Access subscription; please note that your subscription will renew automatically unless you cancel prior to renewal. Cancellation takes effect at the end of your current billing period. We will let you know in advance of any price changes. Taxes may apply. Offer terms are subject to change.
Subscribe to get unlimited Access
Hey, you are out of free articles but you are only a few clicks away from full access. Subscribe below and take advantage of our introductory offer.
subscribe to get Unlimited access
Offer for a Heatmap News Unlimited Access subscription; please note that your subscription will renew automatically unless you cancel prior to renewal. Cancellation takes effect at the end of your current billing period. We will let you know in advance of any price changes. Taxes may apply. Offer terms are subject to change.
Create Your Account
Please Enter Your Password
Forgot your password?
Please enter the email address you use for your account so we can send you a link to reset your password:
Europe could teach America a thing or two about interconnection.

As the invasion of Ukraine raged last year, all eyes were on Europe’s power grid. Gas prices skyrocketed, Scandinavia’s water levels fell, and France’s trusty nuclear power plants went offline. It was a test of whether the world’s most interconnected energy grid could keep the lights on under extreme stress — and Europe passed. Today, as increasingly volatile weather patterns wreak havoc on infrastructure, the grid is proving to be more important than ever.
“Climate change is going to make us rely on the grid more,” Michael Pollitt, a professor of economics at Cambridge and an expert in energy economics, told me. “It’s not just gas price effects across Europe, it’s low water years and low wind years that will have impacts everywhere.”
This summer’s extreme heat could have been the next greatest threat to the power grid following the invasion of Ukraine. But instead, the stresses posed by recent weather have shown the strength of Europe’s power grid, proving the importance of interconnection in an era of global warming.
Europe’s power grid is made up of a series of interconnected localized grids. The primary one is the continental grid, where about 15% of the continent’s energy is traded across borders every year. This grid serves 400 million consumers across 24 countries, including most of the European Union countries, plus the Balkans and Turkey. These 24 countries are also connected to several other grids: the Nordic grid, the British grid, the Irish grid, and, as of August this year, the Baltic grid. Those additions bring electricity to more than 600 million consumers. In addition, there are discussions about connecting North Africa’s power grid, and especially Morocco, which would provide a rich source of solar energy.
Each country invests in what they do best: Norway champions hydropower, France has nuclear power plants, the U.K. invests in wind turbines, Spain does solar, etc. And each one can sell the excess energy to the grid to assist other countries. When water levels are low in the summer months, Norway relies on countries like Spain, who have ample power from their solar fields. In the cloudier winter months, Norway returns the favor. There is a call for faster progress on interconnection and transmission to make this into an even more reliable “Super Grid.”
This tool provides an interactive map of the grid today, and the expected changes up until 2040.
This single market allows for an energy security not seen in the United States, which has several disconnected state or regional grids with much more limited interconnection. This not only restricts the distribution of renewable energy in the U.S., but it can lead to blackouts, most famously in Texas in 2021.
One reason that Europe’s grid has proved remarkably resilient is that mutual reliance also means mutually assured destruction.
“If a country were to reduce exports, it would reduce costs in their country,” said Pollitt, referring to fears last year that European countries would unplug from the interconnected grid to safeguard their own energy supplies. “But you barely think about that for too long before you realize it’s a nuclear option to keep prices down.”
EU countries came together to agree on a gas price cap to contain the energy crisis in December 2022. But the rise in gas prices was a powerful incentive for countries to increase their reliance on renewables. Wind and solar generated 10% more energy compared with the same period in 2021-2022, saving the region 12 billion euros in gas imports (about the same in dollars), according to Ember, an energy think tank.
“I’m very happy to see European solidarity manifest itself and be resilient even though there was some temptation to go it alone,” Kristian Ruby, the secretary general for Eurelectric, the association for the electricity sector in Europe, told me. “By standing together and doubling down on solutions, we’ve seen them keep the lights on during an extremely difficult time.”
Extreme weather is the next big hurdle for the grid to overcome. “There’s no doubt that extreme weather events are becoming a strain on electricity operators,” said Ruby. A recent report from Eurelectric says that all power systems are exposed to the effects of extreme weather, including generation, transmission, and distribution. For hydropower, low water levels are detrimental and extreme cold can cause ice and blockages. Geothermal and nuclear energy become less efficient during heat waves because they require water and cold air for cooling. Many of these plants are also vulnerable to coastal and inland flooding.
This summer in particular, the grid was put to the test. Extreme heat in Spain and Italy pushed the grid to its upper limit. Using power from places like Britain, Norway and Switzerland, Spain was able to provide the power needed. It also benefited from investments in solar panels, which supplied 20 percent more solar power than in the summer of 2022.
The grid’s strength is in its variability. “Different types of weather phenomena call for different coping strategies. Resilience is about diversity. It’s about having a mix of different things. One technology will not solve it alone,” said Ruby.
Renewable energy sources differ based on the conditions in which they are built, which can make the electricity supply more adaptable. If there’s enough interconnection to bring power from, say, where it’s sunny or windy to where it’s needed, countries are much less likely to experience blackouts during severe weather. Whereas with fossil fuel based energy like coal plants, the energy supply is concentrated and more susceptible to shocks.
Despite the success, some experts are concerned that transmission isn’t growing fast enough to handle electrification. People are buying more heat pumps and using electric vehicles, but NGO WindEurope says that the grid itself is not expanding at the same pace. Experts also say that as loads increase, electricity flows will become more complex. Ruby advocates more digitalization in order to handle these complex flows.
The EU Commissioner for Energy Kadri Simson wrote an op-ed piece in the Financial Times this month saying that Europe must sustain a fast pace in rolling out renewables and electrifying the economy. She references the need to integrate intermittent renewable power and adapt more decentralized electricity systems. She says the emphasis needs to be on transmission and distribution grids.
The EU reduced the length of time needed for permitting electricity transmission. It also introduced new emergency legislation last year to accelerate the authorization of renewable projects.
Despite concerns about pace, experts seem generally optimistic about the EU's grid. “EU energy and climate policy are really a success story in European coordination and interdependence,” said Pollitt.
Log in
To continue reading, log in to your account.
Create a Free Account
To unlock more free articles, please create a free account.
On Trump’s electricity insecurity, Rivan’s robots, and the European grid
Current conditions: A series of clipper storms blowing southeastward from Alberta are set to deliver the first measurable amount of snow to the Interstate 95 corridor in the coming days • Planes, trains, and ferries are facing cancellations in Scotland as Storm Bram makes landfall with 70-mile-per-hour winds • In India’s northern Punjab region, a cold snap is creating such a dense fog that travel is being disrupted in some areas.
For the past few days, I have written about alarming forecasts of flooding in the Pacific Northwest as back-to-back atmospheric rivers deluged the region. On Thursday, it became clear just how severe the crisis is becoming, as Washington State issued an urgent order to evacuate more than 100,000 residents, according to The New York Times. Several days of rain have swollen rivers and streams in the Skagit Valley, roughly halfway between Seattle and the Canadian border, putting everyone in the area within a 100-year flood plain. “You can stand downtown here and just see whole Doug firs and cottonwood trees coming down the river, like a freight train,” James Eichner, who fled floodwaters near the Snohomish River farm where he works, told the newspaper. “It’s just a giant steamroller.”
While it’s still difficult to link specific weather events to climate change, federal researchers have connected intensified atmospheric rivers to the planet’s rising temperatures. But the record heat the world has experienced over the past three years isn’t necessarily due entirely to global warming. In a new analysis published in Carbon Brief, climate scientist Zeke Hausfather found that the main culprits turn out to be a combination of El Niño effects, declining sulfur emissions from Chinese coal plants and global shipping no longer masking warming, and stronger than usual solar radiation.

China’s success at building new power plants has earned the country the title of the world’s first “electrostate,” a play on the old petrostate moniker afforded to nations where vast domestic oil and gas resources dominate the economy and politics. In the People’s Republic, blue-gray solar panels glaze entire mountainsides, nuclear reactors come online faster than the U.S. can pick a site for one, and wind turbines spring from the ground. In America, meanwhile, power-thirsty data centers are forced to jury-rig on-site power plants made of old jet engines. But that’s not how President Donald Trump sees it. In response to a Wall Street Journal story outlining how China’s vast grid is boosting Beijing’s artificial intelligence ambitions, Trump posted on his Truth Social website that the newspaper was “WRONG.” He went on: “Every AI plant being built in the United States is building its own Electric Generating Facilities. The approvals are being given carefully, but very quickly, a matter of weeks. Any excess Electricity being produced is going to our Electric Grid, which is being strengthened, and expanded, for other purposes than AI, like never before. In other words, AI has far more Electricity than they will ever need.”
The outburst came as Trump’s approval ratings on the economy sunk to the lowest point ever recorded by an Associated Press-NORC Center for Public Affairs Research poll taken while he’s been in office. Just 31% of U.S. adults now approve of how Trump is handling the economy, down from 40% in March.
Get Heatmap AM directly in your inbox every morning:
Like Tesla, electric vehicle giant Rivian is making a bet on self-driving cars. On Thursday, the company outlined what TechCrunch called “an ambitious effort that includes new hardware, including lidar and custom silicon, and eventually, a potential entry into the self-driving ride-hail market.” The automaker made the announcements at its inaugural “Autonomy & AI Day” event in Palo Alto, where the publication said the company sent “a very public signal to shareholders that it’s keeping pace, or even exceeding, the automated-driving capabilities of industry rivals like Tesla, Ford, General Motors, as well as automakers from Europe and China.”
It’s just the latest extension of Rivian’s empire. In March, Heatmap’s Katie Brigham reported that the company spun off its micromobility division. The startup, named Also, announced a $105 million Series B round on the same day it went independent.
Sign up to receive Heatmap AM in your inbox every morning:
The Republican-led House of Representatives voted largely along partisan lines to pass three bills Thursday designed to speed up permitting for energy projects. While Politico noted the bills are unlikely to be taken up in the Senate, the package managed to pick up six Democrats’ votes in favor of a provision to ease Section 401 Clean Water Act rules and boost pipeline construction. Earlier this week, a bipartisan bill to modernize the federal permitting process and prepare to digitize the experience passed unanimously.
Heatmap’s Jael Holzman broke the news this week that the top three Senate Democrats focused on climate issues oppose another permitting reform bill with bipartisan support, the SPEED Act, on the grounds that it did too little to clear bottlenecks for renewables and transmission lines. For those hoping a long-anticipated bipartisan legislative push may finally come to fruition before tax credits expire for solar and wind projects, at least one thing is clear: A negotiation is underway.
The European Commission proposed what PV Tech called a “two-pronged approach to improving Europe’s energy infrastructure.” The plan aims to speed up and make more transparent the process for permitting projects related to the grid. The proposal also calls for establishing eight new “energy highways” to deliver transmission capacity to areas of “strategic importance” in the 27-nation bloc. The first part, called the “European Grids Package,” calls for creating the Trans-European Network for Energy, or TEN-E, that was first proposed in 2013. The report also noted how far behind Europe was on its previous targets. Despite a goal of 88 gigawatts of new cross-border electricity transmission capacity by 2030, the EU is on pace to build just 41 gigawatts.
The move comes days after European Commission President Ursula von der Leyen notched a victory for her deregulatory agenda when her center-right coalition in the European Parliament broke taboos and teamed up with the far-right to pass legislation easing environmental rules on most companies. As a result of the change, as I wrote in yesterday’s newsletter, 80% of businesses operating in the EU will no longer have to track and report their own environmental impacts.
Deep in the cloud forests of the Serra do Quiriri mountains in the southern Brazilian Atlantic Forest, scientists discovered an entirely new species of frog. Named Brachycephalus lulai, the tiny amphibian has pumpkin-orange skin with green and brown freckles and obsidian eyes. Males max out at a little over 11 millimeters, while females grow nearly 14 millimeters. They are, according to Phys.org, “among the smallest four-legged animals on Earth.” What makes the discovery so interesting is that the researchers identified the so-called “pumpkin toadlet” by its unique mating song, unlike any known species in its genus, which consists of two short bursts of sound, according to their paper in the journal PLOS One.
How America’s one-time leader in designing small modular nuclear reactors missed out on $800 million.
When Congress earmarked $800 million in the 2021 bipartisan infrastructure law to finance the deployment of the United States’ first small modular reactors, there was one obvious recipient lawmakers and industry alike had in mind: NuScale Power.
The Oregon-based company had honed its reactor to meet the 21st century nuclear industry’s needs. The design, completed in the years after the Fukushima disaster in Japan, rendered a similar meltdown virtually impossible. The output, equal to 50 megawatts of electricity, meant that developers would need to install the reactors in packs, which would hasten the rate of learning and bring down costs in much the same way assembly line repetition made solar, wind, and batteries cheap. In mid-2022, the Nuclear Regulatory Commission certified NuScale’s design, making the company’s reactor the first — and so far only — SMR to win federal approval. Seeing NuScale as its champion, the Department of Energy plowed at least $583 million into what was supposed to be the company’s first deployment. To slap an exclamation point on its preeminence, NuScale picked the ticker “SMR” when it went public on the New York Stock Exchange that year.
That September, I toured the shuttered Oyster Creek nuclear plant in New Jersey, where a very different kind of nuclear company, decommissioning specialist Holtec International, was considering building the first of its own as-yet-unapproved SMRs as part of an effort to get into the energy generation game. Holtec’s trajectory to becoming an active nuclear plant operator seemed all but certain, but a former employee cast serious doubts on whether it would end up producing its own reactors. “NuScale is at the front of the line right now,” the former Holtec employee told me at the time. “It’s more realistic to bet your horses on that.”
But forerunners are not always frontrunners. When the Energy Department finally awarded that $800 million earlier this month to two different reactor companies, neither one was NuScale.
Splitting the funding between two projects, the agency gave $400 million to build GE Vernova Hitachi Nuclear Energy’s 300-megawatt BWRX-300 reactor at the Tennessee Valley Authority’s Clinch River site, just south of Oak Ridge. The other $400 million went to Holtec to fund the expansion of the Palisades nuclear plant in western Michigan using the company’s own 300-megawatt SMR-300 reactor — the same one I saw it prepping for in New Jersey.
“I call it the eff NuScale award,” one industry source, who previously worked at NuScale and requested anonymity to speak frankly about the company, told me, using slightly more colorful language.
NuScale declined my request for an interview.
Spun out of research at Oregon State University and the Idaho National Laboratory in 2007, NuScale appeared at the peak of the last attempt at a nuclear renaissance, when the Bush administration planned to build dozens of new reactors to meet the country’s needs for clean electricity. That just two large reactors conceived at that time — the pair of gigawatt-sized Westinghouse AP1000s completed at Southern Company’s Alvin W. Vogtle Electric Generating Plant over the past two years — seemed to justify NuScale’s smaller approach.
Since America’s first commercial nuclear plant came online at Pennsylvania’s Shippingport plant in December 1957, reactors have been bespoke megaprojects, each designed to particular needs and geological conditions. Atomic energy projects regularly went over budget. In the 1960s and 1970s, when the majority of the nation’s 94 operating reactors were built, that didn’t matter. Utilities were vertically integrated monopolies that controlled the power plants, the distribution lines, and sales to ratepayers. Cost overruns on power stations were offset by profits in other divisions. As appliances such as dishwashers, washing machines, and air conditioners relieved the tedium of managing American households, electricity sales climbed and made billion-dollar nuclear projects manageable.
In the 1990s, however, the Clinton-era drive to end big government brought the market’s efficient logic to the electric grid, which was supposed to bring down rates by making power plants compete against each other. The practical effect was to render a years-long endeavor with steep upfront costs, such as building a nuclear plant, virtually impossible to justify in markets where gas plants, solar farms, and wind turbines could come online faster and cheaper. That those energy sources wouldn’t last as long or provide as much electricity as nuclear reactors did not enter into the calculus.
SMRs were supposed to solve that dilemma. The most common metaphor harkened to aerospace: Traditional nuclear plants were built to local specs, like airports, whereas SMRs would be built like airplanes rolling off the factory floor. A utility looking to generate a gigawatt of electricity could build one AP1000, or it could buy 20 of NuScale’s 50-megawatt units. Vogtle Unit 4, which came online last year, ended up costing 30% less than Vogtle Unit 3, the debut AP1000 that started up in 2023, since it could rely on the previous unit’s design and supply chain. If NuScale’s reactors followed the same trajectory, the cost savings by the time the 20th reactor came online would be stupendous.
But what works on paper doesn’t always pan out in concrete. In November 2023, less than three months after Vogtle Unit 3 entered into service, NuScale’s first project — a half-dozen of reactors near the Idaho National Laboratory, meant to sell electricity to a network of municipal power companies in Utah — collapsed as inflation ballooned costs.
The company seemingly hasn’t been able to catch a break since then. Last year, the U.S. Export-Import Bank approved a loan to fund construction of a NuScale project in Romania; in August, the company announced that a final investment decision on the plant near Bucharest could be delayed until 2027. Over the summer, a project developer in Idaho floated the idea of building NuScale reactors at the site of a giant wind farm the Trump administration canceled. But NuScale denied the effort in an email to me at the time, and nothing has yet come of it.
The company has lately shown some green shoots, however. The NRC approved an upgrade to NuScale’s design in July, raising the output to 77 megawatts to make the reactor roughly 50% more powerful. In September, NuScale’s exclusive development partner, Entra1, inked a deal with the TVA to build up to six of its reactors at one of the federal utility’s sites in southeastern Tennessee.
“It’s too early to discount NuScale,” Chris Gadomski, the lead nuclear analyst at the consultancy BloombergNEF, told me.
But the TVA project was also too early-stage for the Energy Department to make a bet, experts told me.
“This isn’t necessarily the government picking winners here as much as the market is supporting projects at these two sites, at least pending government approval,” Adam Stein, the director of nuclear energy innovation at the think tank Breakthrough Institute, said. “The government is supporting projects the market has already considered.”
By contrast, GE-Hitachi’s Clinch River project has been in the works for nearly four years. The BWRX-300 has other advantages. GE-Hitachi — a joint venture between the American energy-equipment giant GE Vernova and the Japanese industrial behemoth Hitachi — has decades of experience in the nuclear space. Indeed, a third of the reactors in the U.S. fleet are boiling water reactors, the design GE pioneered in the mid-20th century and updated as an SMR with the BWRX-300. Making the technology more appealing is the fact that Ontario Power Generation is building the first BWRX-300, meaning that the state-owned utility in Canada’s most populous province can work out the kinks and allow for the TVA’s project to piggyback off the lessons learned.
While Holtec may be a newcomer to nuclear generation, the company has manufactured specialized containers to store spent reactor fuel for more than three decades, giving it experience in nuclear projects. Holtec is also close to bringing the single reactor at the Palisades plant back online, which will be the first time a nuclear plant returns to regular operation in the U.S. Like NuScale’s, Holtec’s SMR is based on the pressurized water reactor design that makes up nearly 70% of the U.S. fleet.
The point is, both companies have existing nuclear businesses that lay the groundwork for becoming SMR vendors. “GE is a nuclear fuel and services business and Holtec is a nuclear waste services and decommissioning business. That’s what they live on,” the former NuScale employee told me. “NuScale lives on the thoughts, prayers, and good graces of investors.”
Shares of NuScale today trade at roughly double the price of its initial public offering, which is at least in part a reflection of the feverish stock surges for SMR companies over the past year. The artificial intelligence boom has spurred intense excitement on Wall Street for nuclear power, but many of the established companies in the industry are not publicly traded — Westinghouse, GE-Hitachi, and Holtec are all privately held. That could be an advantage. Last month, the prices of most major SMR companies plunged in what the journalist Robert Bryce said indicates the “hype over SMRs is colliding with the realities of the marketplace.” NuScale saw the steepest drop.
But Brett Rampal, a nuclear analyst at the consultancy Veriten, said NuScale’s “current focus around its relationship with Entra1” could make the company more nimble than its rivals because it can “pursue potential projects absent a direct utility customer, like GE, or owning the asset themselves, like Holtec.”
One factor the market isn’t apparently considering yet: whether the type of SMR NuScale, GE-Hitachi, and Holtec are designing actually pencil out.
The Energy Department’s funding was designed for third-generation SMRs, meaning shrunk-down, less powerful versions of light water reactors, an umbrella category that includes both boiling and pressurized water reactors. The option to go smaller existed in the heyday of nuclear construction in the 1970s, but developers at that time found that larger reactors delivered economies of scale that made more financial sense. Neither Russia, the world’s top nuclear exporter and the only country to deploy an SMR so far, nor China, the nation building the most new atomic power plants by far, including an SMR, has filled its order books with smaller reactors. Instead, the leading Chinese design is actually a bigger, more powerful version of the AP1000.
Calculations from the Massachusetts Institute of Technology estimate that the first BWRX-300 will cost significantly more than another AP1000, given that the GE-Hitachi model has yet to be built and the Westinghouse reactor has an established design and supply chain. That reality has propelled growing interest in building large-scale reactors again in the U.S. In October, the Department of Commerce brokered a landmark deal to spend $80 billion on 10 new AP1000s. This week, Westinghouse’s majority owner Brookfield inked a deal to complete construction on the aborted VC Summer AP1000 project in South Carolina.
At the same time, the Energy Department has kicked off a pilot program designed to hasten deployment of fourth-generation reactors, the type of technology that uses coolants other than water. Bill Gates’ molten salt-cooled reactor company, TerraPower, just cleared its final safety hurdle at the NRC for its so-called Natrium reactor, setting the stage to potentially build the nation’s first commercial fourth-generation nuclear plant in Wyoming.
“From a marketing point of view, everyone has consistently said that light water reactor SMRs will be the fastest to market,” Stein said. But the way things are going, both NuScale and its peers could get lapped yet again.
Citrine Informatics has been applying machine learning to materials discovery for years. Now more advanced models are giving the tech a big boost.
When ChatGPT launched three years ago, it became abundantly clear that the power of generative artificial intelligence had the capacity to extend far beyond clever chatbots. Companies raised huge amounts of funding based on the idea that this new, more powerful AI could solve fundamental problems in science and medicine — design new proteins, discover breakthrough drugs, or invent new battery chemistries.
Citrine Informatics, however, has largely kept its head down. The startup was founded long before the AI boom, back in 2013, with the intention of using simple old machine learning to speed up the development of more advanced, sustainable materials. These days Citrine is doing the same thing, but with neural networks and transformers, the architecture that undergirds the generative AI revolution.
“The technology transition we’re going through right now is pretty massive,” Greg Mulholland, Citrine’s founder and CEO, told me. “But the core underlying goal of the company is still the same: help scientists identify the experiments that will get them to their material outcome as fast as possible.”
Rather than developing its own novel materials, Citrine operates on a software-as-a-service model, selling its platform to companies including Rolls-Royce, EMD Electronics, and chemicals giant LyondellBassell. While a SaaS product may be less glamorous than independently discovering a breakthrough compound that enables something like a room-temperature superconductor or an ultra-high-density battery, Citrine’s approach has already surfaced commercially relevant materials across a variety of sectors, while the boldest promises of generative AI for science remain distant dreams.
“You can think of it as science versus engineering,” Mulholland told me. “A lot of science is being done. Citrine is definitely the best in kind of taking it to the engineering level and coming to a product outcome rather than a scientific discovery.” Citrine has helped to develop everything from bio-based lotion ingredients to replace petrochemical-derived ones, to plastic-free detergents, to more sustainable fire-resistant home insulation, to PFAS-free food packaging, to UV-resistant paints.
On Wednesday, the company unveiled two new platform capabilities that it says will take its approach to the next level. The first is essentially an advanced LLM-powered filing system that organizes and structures unwieldy materials and chemicals datasets from across a company. The second is an AI framework informed by an extensive repository of chemistry, physics, and materials knowledge. It can ingest a company’s existing data, and, even if the overall volume is small, use it to create a list of hundreds of potential new materials optimized for factors such as sustainability, durability, weight, manufacturability, or whatever other outcomes the company is targeting.
The platform is neither purely generative nor purely predictive. Instead, Mulholland explained, companies can choose to use Citrine’s tools “in a more generative mode” if they want to explore broadly and open up the field of possible materials discoveries, or in a more “optimized” mode that stays narrowly focused on the parameters they set. “What we find is you need a healthy blend of the two,” he told me.
The novel compounds the model spits out still need to be synthesized and tested by humans. “What I tell people is, any plane made of materials designed exclusively by Citrine and never tested is not a plane I’m getting on,” Mulholland told me. The goal isn’t to achieve perfection right out of the lab, but rather to optimize the experiments companies end up having to do. “We still need to prove materials in the real world, because the real world will complicate it.”
Indeed it will. For one thing, while AI is capable of churning out millions of hypothetical materials — as a tool developed by Google DeepMind did in 2023 — materials scientists have since shown that many are just variants of known compounds, while others are unstable, unable to be synthesized, or otherwise irrelevant under real world conditions.
Such failures likely stem, in part, from another common limitation of AI models trained solely on publicly available materials and chemicals data: Academic research tends to report only successful outcomes, omitting data on what didn’t work and which compounds weren’t viable. That can lead models to be overly optimistic about the magnitude and potential of possible materials solutions and generate unrealistic “discoveries” that may have already been tested and rejected.
Because Citrine’s platform is deployed within customer organizations, it can largely sidestep this problem by tuning its model on niche, proprietary datasets. These datasets are small when compared with the vast public repositories used to train Citrine’s base model, but the granular information they contain about prior experiments — both successes and failures — has proven critical to bringing new discoveries to market.
While the holy grail for materials science may be a model trained on all the world’s relevant data — public and private, positive and negative — at this point that’s just a fantasy, one of Citrine’s investors, Mark Cupta of Prelude Ventures, told me over email. “It’s hard to get buy-in from the entire material development world to make an open-source model that pulls in data from across the field.”
Citrine’s last raise, which Prelude co-led, came at the very beginning of 2023, as the AI wave was still gathering momentum. But Mulholland said there’s no rush to raise additional capital — in fact, he expects Citrine to turn a profit in the next year or so.
That milestone would strongly validate the company’s strategy, which banks on steady revenue from its subscription-based model to compensate for the fact that it doesn’t own the intellectual property for the materials it helps develop. While Mulholland told me that many players in this space are trying to “invent new materials and patent them and try to sell them like drugs,” Citrine is able to “invent things much more quickly, in a more realistic way than the pie in the sky, hoping for a Nobel Prize [approach].”
Citrine is also careful to assure that its model accounts for real world constraints such as regulations and production bottlenecks. Say a materials company is creating an aluminum alloy for an automaker, Mulholland explained — it might be critical to stay within certain elemental bounds. If the company were to add in novel elements, the automaker would likely want to put its new compound through a rigorous testing process, which would be annoying if it’s looking to get to market as quickly as possible. Better, perhaps, to tinker around the edges of what’s well understood.
In fact, Mulholland told me it’s often these marginal improvements that initially bring customers into the fold, convincing them that this whole AI-for-materials thing is more than just hype. “The first project is almost always like, make the adhesive a little bit stickier — because that’s a good way to prove to these skeptical scientists that AI is real and here to stay,” he said. “And then they use that as justification to invest further and further back in their product development pipeline, such that their whole product portfolio can be optimized by AI.”
Overall, the company says that its new framework can speed up materials development by 80%. So while Mulholland and Citrine overall may not be going for the Nobel in Chemistry, don’t doubt for a second that they’re trying to lead a fundamental shift in the way consumer products are designed.
“I’m as bullish as I can possibly be on AI in science,” Mulholland told me. “It is the most exciting time to be a scientist since Newton. But I think that the gap between scientific discovery and realized business is much larger than a lot of AI folks think.”