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:
It took the market about a week to catch up to the fact that the Chinese artificial intelligence firm DeepSeek had released an open-source AI model that rivaled those from prominent U.S. companies such as OpenAI and Anthropic — and that, most importantly, it had managed to do so much more cheaply and efficiently than its domestic competitors. The news cratered not only tech stocks such as Nvidia, but energy stocks, as well, leading to assumptions that investors thought more-energy efficient AI would reduce energy demand in the sector overall.
But will it really? While some in climate world assumed the same and celebrated the seemingly good news, many venture capitalists, AI proponents, and analysts quickly arrived at essentially the opposite conclusion — that cheaper AI will only lead to greater demand for AI. The resulting unfettered proliferation of the technology across a wide array of industries could thus negate the energy efficiency gains, ultimately leading to a substantial net increase in data center power demand overall.
“With cost destruction comes proliferation,” Susan Su, a climate investor at the venture capital firm Toba Capital, told me. “Plus the fact that it’s open source, I think, is a really, really big deal. It puts the power to expand and to deploy and to proliferate into billions of hands.”
If you’ve seen lots of chitchat about Jevons paradox of late, that’s basically what this line of thinking boils down to. After Microsoft’s CEO Satya Nadella responded to DeepSeek mania by posting the Wikipedia page for this 19th century economic theory on X, many (myself included) got a quick crash course on its origins. The idea is that as technical efficiencies of the Victorian era made burning coal cheaper, demand for — and thus consumption of — coal actually increased.
While this is a distinct possibility in the AI space, it’s by no means a guarantee. “This is very much, I think, an open question,“ energy expert Nat Bullard told me, with regards to whether DeepSeek-type models will spur a reduction or increase in energy demand. “I sort of lean in both directions at once.” Formerly the chief content officer at BloombergNEF and current co-founder of the AI startup Halcyon, a search and information platform for energy professionals, Bullard is personally excited for the greater efficiencies and optionality that new AI models can bring to his business.
But he warns that just because DeepSeek was cheap to train — the company claims it cost about $5.5 million, while domestic models cost hundreds of millions or even billions — doesn’t mean that it’s cheap or energy-efficient to operate. “Training more efficiently does not necessarily mean that you can run it that much more efficiently,” Bullard told me. When a large language model answers a question or provides any type of output, it’s said to be making an “inference.” And as Bullard explains, “That may mean, as we move into an era of more and more inference and not just training, then the [energy] impacts could be rather muted.”
DeepSeek-R1, the name for the model that caused the investor freakout, is also a newer type of LLM that uses more energy in general. Up until literally a few days ago, when OpenAI released o3-mini for free, most casual users were probably interacting with so-called “pretrained” AI models. Fed on gobs of internet text, these LLMs spit out answers based primarily on prediction and pattern recognition. DeepSeek released a model like this, called V3, in September. But last year, more advanced “reasoning” models, which can “think,” in some sense, started blowing up. These models — which include o3-mini, the latest version of Anthropic’s Claude, and the now infamous DeepSeek-R1 — have the ability to try out different strategies to arrive at the correct answer, recognize their mistakes, and improve their outputs, allowing for significant advancements in areas such as math and coding.
But all that artificial reasoning eats up a lot of energy. As Sasha Luccioni, the AI and climate lead at Hugging Face, which makes an open-source platform for AI projects, wrote on LinkedIn, “To set things clear about DeepSeek + sustainability: (it seems that) training is much shorter/cheaper/more efficient than traditional LLMs, *but* inference is longer/more expensive/less efficient because of the chain of thought aspect.” Chain of thought refers to the reasoning process these newer models undertake. Luccioni wrote that she’s currently working to evaluate the energy efficiency of both the DeepSeek V3 and R1 models.
Another factor that could influence energy demand is how fast domestic companies respond to the DeepSeek breakthrough with their own new and improved models. Amy Francetic, co-founder at Buoyant Ventures, doesn’t think we’ll have to wait long. “One effect of DeepSeek is that it will highly motivate all of the large LLMs in the U.S. to go faster,” she told me. And because a lot of the big players are fundamentally constrained by energy availability, she’s crossing her fingers that this means they’ll work smarter, not harder. “Hopefully it causes them to find these similar efficiencies rather than just, you know, pouring more gasoline into a less fuel-efficient vehicle.”
In her recent Substack post, Su described three possible futures when it comes to AI’s role in the clean energy transition. The ideal is that AI demand scales slowly enough that nuclear and renewables scale with it. The least hopeful is that immediate, exponential growth in AI demand leads to a similar expansion of fossil fuels, locking in new dirty infrastructure for decades. “I think that's already been happening,” Su told me. And then there’s the techno-optimist scenario, linked to figures like Sam Altman, which Su doesn’t put much stock in — that AI “drives the energy revolution” by helping to create new energy technologies and efficiencies that more than offset the attendant increase in energy demand.
Which scenario predominates could also depend upon whether greater efficiencies, combined with the adoption of AI by smaller, more shallow-pocketed companies, leads to a change in the scale of data centers. “There’s going to be a lot more people using AI. So maybe that means we don’t need these huge, gigawatt data centers. Maybe we need a lot more smaller, megawatt-size data centers,” Laura Katzman, a principal at Buoyant Ventures, told me. Katzman has conducted research for the firm on data center decarbonization.
Smaller data centers with a subsequently smaller energy footprint could pair well with renewable-powered microgrids, which are less practical and economically feasible for hyperscalers. That could be a big win for solar and wind plus battery storage, Katzman explained, but a boondoggle for companies such as Microsoft, which has famously committed to re-opening Pennsylvania’s Three Mile Island nuclear plant to power its data centers. “Because of DeepSeek, the expected price of compute probably doesn’t justify now turning back on some of these nuclear plants, or these other high-cost energy sources,” Katzman told me.
Lastly, it remains to be seen what nascent applications cheaper models will open up. “If somebody, say, in the Philippines or Vietnam has an interest in applying this to their own decarbonization challenge, what would they come up with?” Bullard pondered. “I don’t yet know what people would do with greater capability and lower costs and a different set of problems to solve for. And that’s really exciting to me.”
But even if the AI pessimists are right, and these newer models don’t make AI ubiquitously useful for applications from new drug discovery to easier regulatory filing, Su told me that in a certain sense, it doesn't matter much. “If there was a possibility that somebody had this type of power, and you could have it too, would you sit on the couch? Or would you arms race them? I think that is going to drive energy demand, irrespective of end utility.”
As Su told me, “I do not think there’s actually a saturation point for this.”
Log in
To continue reading, log in to your account.
Create a Free Account
To unlock more free articles, please create a free account.
“We had enough assurance that the president was going to deal with them.”
A member of the House Freedom Caucus said Wednesday that he voted to advance President Trump’s “big, beautiful bill” after receiving assurances that Trump would “deal” with the Inflation Reduction Act’s clean energy tax credits – raising the specter that Trump could try to go further than the megabill to stop usage of the credits.
Representative Ralph Norman, a Republican of North Carolina, said that while IRA tax credits were once a sticking point for him, after meeting with Trump “we had enough assurance that the president was going to deal with them in his own way,” he told Eric Garcia, the Washington bureau chief of The Independent. Norman specifically cited tax credits for wind and solar energy projects, which the Senate version would phase out more slowly than House Republicans had wanted.
It’s not entirely clear what the president could do to unilaterally “deal with” tax credits already codified into law. Norman declined to answer direct questions from reporters about whether GOP holdouts like himself were seeking an executive order on the matter. But another Republican holdout on the bill, Representative Chip Roy of Texas, told reporters Wednesday that his vote was also conditional on blocking IRA “subsidies.”
“If the subsidies will flow, we’re not gonna be able to get there. If the subsidies are not gonna flow, then there might be a path," he said, according to Jake Sherman of Punchbowl News.
As of publication, Roy has still not voted on the rule that would allow the bill to proceed to the floor — one of only eight Republicans yet to formally weigh in. House Speaker Mike Johnson says he’ll, “keep the vote open for as long as it takes,” as President Trump aims to sign the giant tax package by the July 4th holiday. Norman voted to let the bill proceed to debate, and will reportedly now vote yes on it too.
Earlier Wednesday, Norman said he was “getting a handle on” whether his various misgivings could be handled by Trump via executive orders or through promises of future legislation. According to CNN, the congressman later said, “We got clarification on what’s going to be enforced. We got clarification on how the IRAs were going to be dealt with. We got clarification on the tax cuts — and still we’ll be meeting tomorrow on the specifics of it.”
Neither Norman nor Roy’s press offices responded to a request for comment.
The foreign entities of concern rules in the One Big Beautiful Bill would place gigantic new burdens on developers.
Trump campaigned on cutting red tape for energy development. At the start of his second term, he signed an executive order titled, “Unleashing Prosperity Through Deregulation,” promising to kill 10 regulations for each new one he enacted.
The order deems federal regulations an “ever-expanding morass” that “imposes massive costs on the lives of millions of Americans, creates a substantial restraint on our economic growth and ability to build and innovate, and hampers our global competitiveness.” It goes on to say that these regulations “are often difficult for the average person or business to understand,” that they are so complicated that they ultimately increase the cost of compliance, as well as the risks of non-compliance.
Reading this now, the passage echoes the comments I’ve heard from industry groups and tax law experts describing the incredibly complex foreign entities of concern rules that Congress — with the full-throated backing of the Trump administration — is about to impose on clean energy projects and manufacturers. Under the One Big Beautiful Bill Act, wind and solar, as well as utility-scale energy storage, geothermal, nuclear, and all kinds of manufacturing projects will have to abide by restrictions on their Chinese material inputs and contractual or financial ties with Chinese entities in order to qualify for tax credits.
“Foreign entity of concern” is a U.S. government term referring to entities that are “owned by, controlled by, or subject to the jurisdiction or direction of” any of four countries — Russia, Iran, North Korea, and most importantly for clean energy technology, China.
Trump’s tax bill requires companies to meet increasingly strict limits on the amount of material from China they use in their projects and products. A battery factory starting production next year, for example, would have to ensure that 60% of the value of the materials that make up its products have no connection to China. By 2030, the threshold would rise to 85%. The bill lays out similar benchmarks and timelines for clean electricity projects, as well as other kinds of manufacturing.
But how companies should calculate these percentages is not self-evident. The bill also forbids companies from collecting the tax credits if they have business relationships with “specified foreign entities” or “foreign-influenced entities,” terms with complicated definitions that will likely require guidance from the Treasury for companies to be sure they pass the test.
Regulatory uncertainty could stifle development until further guidance is released, but how long that takes will depend on if and when the Trump administration prioritizes getting it done. The One Big Beautiful Bill Act contains a lot of other new tax-related provisions that were central to the Trump campaign, including a tax exemption for tips, which are likely much higher on the department’s to-do list.
Tax credit implementation was a top priority for the Biden administration, and even with much higher staffing levels than the department currently has, it took the Treasury 18 months to publish initial guidance on foreign entities of concern rules for the Inflation Reduction Act’s electric vehicle tax credit. “These things are so unbelievably complicated,” Rachel McCleery, a former senior advisor at the Treasury under Biden, told me.
McCleery questioned whether larger, publicly-owned companies would be able to proceed with major investments in things like battery manufacturing plants until that guidance is out. She gave the example of a company planning to pump out 100,000 batteries per year and claim the per-kilowatt-hour advanced manufacturing tax credit. “That’s going to look like a pretty big number in claims, so you have to be able to confidently and assuredly tell your shareholder, Yep, we’re good, we qualify, and that requires a certification” by a tax counsel, she said. To McCleery, there’s an open question as to whether any tax counsel “would even provide a tax opinion for publicly-traded companies to claim credits of this size without guidance.”
John Cornwell, the director of policy at the Good Energy Collective, which conducts research and advocacy for nuclear power, echoed McCleery’s concerns. “Without very clear guidelines from the Treasury and IRS, until those guidelines are in place, that is going to restrict financing and investment,” Cornwell told me.
Understanding what the law requires will be the first challenge. But following it will involve tracking down supply chain data that may not exist, finding alternative suppliers that may not be able to fill the demand, and establishing extensive documentation of the origins of components sourced through webs of suppliers, sub-suppliers, and materials processors.
The Good Energy Collective put out an issue brief this week describing the myriad hurdles nuclear developers will face in trying to adhere to the tax credit rules. Nuclear plants contain thousands of components, and documenting the origin of everything from “steam generators to smaller items like specialized fasteners, gaskets, and electronic components will introduce substantial and costly administrative burdens,” it says. Additionally the critical minerals used in nuclear projects “often pass through multiple processing stages across different countries before final assembly,” and there are no established industry standards for supply chain documentation.
Beyond the documentation headache, even just finding the materials could be an issue. China dominates the market for specialized nuclear-grade materials manufacturing and precision component fabrication, the report says, and alternative suppliers are likely to charge premiums. Establishing new supply chains will take years, but Trump’s bill will begin enforcing the sourcing rules in 2026. The rules will prove even more difficult for companies trying to build first-of-a-kind advanced nuclear projects, as those rely on more highly specialized supply chains dominated by China.
These challenges may be surmountable, but that will depend, again, on what the Treasury decides, and when. The Department’s guidance could limit the types of components companies have to account for and simplify the documentation process, or it could not. But while companies wait for certainty, they may also be racking up interest. “The longer there are delays, that can have a substantial risk of project success,” Cornwell said.
And companies don’t have forever. Each of the credits comes with a phase-out schedule. Wind manufacturers can only claim the credits until 2028. Other manufacturers have until 2030. Credits for clean power projects will start to phase down in 2034. “Given the fact that a lot of these credits start lapsing in the next few years, there’s a very good chance that, because guidance has not yet come out, you’re actually looking at a much smaller time frame than than what is listed in the bill,” Skip Estes, the government affairs director for Securing America’s Energy Future, or SAFE, told me.
Another issue SAFE has raised is that the way these rules are set up, the foreign sourcing requirements will get more expensive and difficult to comply with as the value of the tax credits goes down. “Our concern is that that’s going to encourage companies to forego the credit altogether and just continue buying from the lowest common denominator, which is typically a Chinese state-owned or -influenced monopoly,” Estes said.
McCleery had another prediction — the regulations will be so burdensome that companies will simply set up shop elsewhere. “I think every industry will certainly be rethinking their future U.S. investments, right? They’ll go overseas, they’ll go to Canada, which dumped a ton of carrots and sticks into industry after we passed the IRA,” she said.
“The irony is that Republicans have historically been the party of deregulation, creating business friendly environments. This is completely opposite, right?”
On the budget debate, MethaneSAT’s untimely demise, and Nvidia
Current conditions: The northwestern U.S. faces “above average significant wildfire potential” for July • A month’s worth of rain fell over just 12 hours in China’s Hubei province, forcing evacuations • The top floor of the Eiffel Tower is closed today due to extreme heat.
The Senate finally passed its version of Trump’s One Big Beautiful Bill Act Tuesday morning, sending the tax package back to the House in hopes of delivering it to Trump by the July 4 holiday. The excise tax on renewables that had been stuffed into the bill over the weekend was removed after Senator Lisa Murkowski of Alaska struck a deal with the Senate leadership designed to secure her vote. In her piece examining exactly what’s in the bill, Heatmap’s Emily Pontecorvo explains that even without the excise tax, the bill would “gum up the works for clean energy projects across the spectrum due to new phase-out schedules for tax credits and fast-approaching deadlines to meet complex foreign sourcing rules.” Debate on the legislation begins on the House floor today. House Speaker Mike Johnson has said he doesn’t like the legislation, and a handful of other Republicans have already signaled they won’t vote for it.
The Environmental Protection Agency this week sent the White House a proposal that is expected to severely weaken the federal government’s ability to rein in planet-warming pollution. Details of the proposal, titled “Greenhouse Gas Endangerment Finding and Motor Vehicle Reconsideration,” aren’t clear yet, but EPA Administrator Lee Zeldin has reportedly been urging the Trump administration to repeal the 2009 “endangerment finding,” which explicitly identified greenhouse gases as a public health threat and gave the EPA the authority to regulate them. Striking down that finding would “free EPA from the legal obligation to regulate climate pollution from most sources, including power plants, cars and trucks, and virtually any other source,” wrote Alex Guillén at Politico. The title of the proposal suggests it aims to roll back EPA tailpipe emissions standards, as well.
Get Heatmap AM directly in your inbox every morning:
So long, MethaneSAT, we hardly knew ye. The Environmental Defense Fund said Tuesday that it had lost contact with its $88 million methane-detecting satellite, and that the spacecraft was “likely not recoverable.” The team is still trying to figure out exactly what happened. MethaneSAT launched into orbit last March and was collecting data about methane pollution from global fossil fuel infrastructure. “Thanks to MethaneSAT, we have gained critical insight about the distribution and volume of methane being released from oil and gas production areas,” EDF said. “We have also developed an unprecedented capability to interpret the measurements from space and translate them into volumes of methane released. This capacity will be valuable to other missions.“ The good news is that MethaneSAT was far from the only methane-tracking satellite in orbit.
Nvidia is backing a D.C.-based startup called Emerald AI that “enables AI data centers to flexibly adjust their power consumption from the electricity grid on demand.” Its goal is to make the grid more reliable while still meeting the growing energy demands of AI computing. The startup emerged from stealth this week with a $24.5 million seed round led by Radical Ventures and including funding from Nvidia. Emerald AI’s platform “acts as a smart mediator between the grid and a data center,” Nvidia explains. A field test of the software during a grid stress event in Phoenix, Arizona, demonstrated a 25% reduction in the energy consumption of AI workloads over three hours. “Renewable energy, which is intermittent and variable, is easier to add to a grid if that grid has lots of shock absorbers that can shift with changes in power supply,” said Ayse Coskun, Emerald AI’s chief scientist and a professor at Boston University. “Data centers can become some of those shock absorbers.”
In case you missed it: California Governor Gavin Newsom on Monday rolled back the state’s landmark Environmental Quality Act. The law, which had been in place since 1970, required environmental reviews for construction projects and had become a target for those looking to alleviate the state’s housing crisis. The change “means most urban developers will no longer have to study, predict, and mitigate the ways that new housing might affect local traffic, air pollution, flora and fauna, noise levels, groundwater quality, and objects of historic or archeological significance,” explainedCal Matters. On the other hand, it could also mean that much-needed housing projects get approved more quickly.
Tesla is expected to report its Q2 deliveries today, and analysts are projecting a year-over-year drop somewhere from 11% to 13%.