Last October, Fermi America raised $746 million in an IPO to build what it claimed would be the largest data center in the world. The company successfully pitched investors on a massive campus it tentatively called the President Donald J. Trump Advanced Energy and Intelligence Campus. (The company now generally refers to it as Project Matador.)

Speed and scale were key to Fermi’s pitch. The company told investors it expected to make $1.5 billion in revenue for every 1 GW of capacity built. It said it expected to launch the first one million square feet of space by April 2026 and generate 1 GW of power by the end of 2026. Fermi originally pitched an 11 GW campus; it later increased the targeted capacity to 17 GW.

On its website, the company says that the “initial phase of construction [is] already complete.” But the company has still yet to begin constructing its first buildings, according to Cleanview’s satellite tracker.

Across the country, data center developers are racing to hire construction workers. At the site of OpenAI’s Stargate campus in Abilene 5,000 workers are working through the night to finish the 1.2 GW campus. At Fermi’s site in Amarillo, the workforce has been shrinking, according to local media reports.

In February, The Amarillo Tribune reported rumors of layoffs among construction workers at the site. In response, Fermi’s CEO Toby Neugebauer said that the company was awaiting a ruling on its final Clean Air Permit. “We completed our first phase so fast that we paused construction temporarily,” he said in a statement. A company spokesperson later told KAMR that Fermi still had 100 construction workers on site despite the pause.

Weeks later, on February 25th, Fermi announced that it had secured its air permit and posted a photo to its website with a caption that read, “First phase of construction almost complete.”

Satellite images suggest the company hasn’t begun significant construction activity since then, however.

This is what Fermi’s data center site looked like in August 2025, the month before the company filed its IPO documents.

Two months later, the company went public. That same month, construction workers began clearing land at the site.

By February—the month that Fermi said it had paused construction—this is what the site looked like.

Two months after construction was supposed to resume—and the month that Fermi originally said it would finish its first one million square feet of space—the site looks essentially the same.

Fermi America isn’t the only company attempting to build a mega-campus like this. More than 70 gigawatt-scale data centers have been proposed in the US since the launch of ChatGPT, according to Cleanview’s data center tracker. (1 GW is equivalent to the power demand of an entire city).

Many of these data centers are already under construction, which gives us the ability to compare Fermi’s first 6 months of progress to that of other companies.

Meta announced its 5GW Hyperion campus in Richland Parish, Louisiana in December 2024 and began clearing land at the site two months later. Here’s what the site looked like 6 months later.

In June 2024, OpenAI’s partner Crusoe began building its 1.2 GW Stargate campus in Abilene, Texas. Within six months, Crusoe finished its first building shell and was putting the roof on the second building. The company launched its first 980,000 square foot phase (200 MW) in September 2025, 15 months after it began construction.

50 miles north of Abilene, Vantage Data Centers is building the second Stargate campus for OpenAI. The 1.2 GW project began construction in September 2025. Here’s what the site looked like 6 months later.

The Rise and Stall of the Largest Data Center in the World

Six months ago, Fermi America raised one of the largest data center IPOs in history.

The company was co-founded by Toby Neugebauer, the son of retired Republican Congressman Randy Neugebauer, whose former district includes the Amarillo area where Project Matador sits. Rick Perry, the former governor of Texas and Trump’s first Secretary of Energy, is also a co-founder. When Fermi went public in October 2025, Perry’s 2.5% stake was worth roughly $540 million. Neugebauer’s 28% stake was worth approximately $6 billion.

From the start, Fermi has touted its proximity to the White House.

“If the President of the United States wants to move fast [and] wants to have a shovel-ready project in the very near term, Project Matador is his first, second, and third great choice.”

— Toby Neugebauer, Fermi America CEO

The company tentatively named the campus after the president. On its first earnings call, Neugebauer disclosed that Energy Secretary Chris Wright and Interior Secretary Doug Burgum had personally intervened to help the company secure Siemens turbines during trade negotiations with Germany. Neugebauer said Siemens “wanted confirmation that this would make the United States happy” and that Fermi received “a massive push-push by the United States of America” to get the deal done.

In July 2025, Trump signed an executive order designating AI data centers at DOE facilities as critical defense infrastructure — directly benefiting Fermi’s site adjacent to the Pantex nuclear weapons plant. Neugebauer told investors that “if the President of the United States wants to move fast, wants to have a shovel-ready project in the very near term, that Project Matador is his first, second, and third great choice.”

In November, the company announced it had signed a $150 million funding agreement with an unnamed, investment-grade tenant. Neugebauer told Business Insider that the tenant was Amazon. (After the story published, Fermi issued a statement saying they “categorically deny Business Insider’s claim.”)

Things began to unravel quickly. In Fermi’s first investor presentation in November, the final slide showed that tenant negotiations were behind schedule. When an analyst pressed for details, Neugebauer said they were just three weeks behind and would make up the time. Three weeks later, the exclusivity window on the deal expired. Two days after that, the tenant formally pulled out. The company’s stock immediately plummeted 34%.

In January, a class action lawsuit was filed alleging that Fermi had overstated tenant demand and failed to disclose how dependent the project was on a single tenant’s funding commitment.

In its second earnings call with investors, on March 30th, company executives said they still hadn’t secured a tenant. The CFO said they wouldn’t invest significant capital into construction until they had. He told investors Fermi “could be forced to… surrender collateral to preserve liquidity.” In other words, the company may need to sell some of the Siemens turbines that two cabinet secretaries helped secure.

When a Cantor Fitzgerald analyst asked for more details, Neugebauer jumped in to walk it back. He said giving up turbines was “not our intention whatsoever,” and added that he’d “auction off my two boys first before I would let one of these gensets go.”

On April 17, two and a half weeks after that earnings call, Fermi announced Neugebauer’s immediate departure. Shares fell as much as 31% in post-market trading.

A Year Behind and Counting

Fermi still has a long way to go before it can generate revenue. In order to begin construction and secure project financing, the company needs to find a tenant. Even after that, building a data center at this scale takes time. Crusoe’s first Stargate campus in Abilene took 13 months from the time it began “vertical construction” to going live, with construction running 24 hours a day. That was for an initial 200 MW phase, which is a fraction of the 1,000 MW Fermi is targeting.

If Fermi secured a tenant this month and replicated Crusoe’s speed, it would bring the first buildings online in May 2027—a full year after the initial timeline projected in its IPO filing.

Share

How we did this analysis

This report is based on Cleanview’s monthly satellite tracker, SEC filings (8-K, 10-K), earnings call transcripts, investor presentations, and news reporting from the Amarillo Tribune and Business Insider.

Cleanview tracks more than 850 planned data centers in the US. For each site, we have satellite imagery dating back to 2022 giving us the ability to track progress over time. We used these satellite images to analyze Fermi’s progress compared to other data center developers.

You can try our new satellite tracking feature for free here:

Try Cleanview

In the eyes of Google’s executives, AI is both a once-in-a-generation opportunity and an existential threat to its search monopoly. The company is spending accordingly: Alphabet invested roughly $90 billion in capital expenditures in 2025 and plans to nearly double that to as much as $185 billion in 2026.

Most of that money will flow to data centers—and the infrastructure needed to power them.

Each new data center requires as much electricity as a small city. Permit documents and site plans we obtained show campus designs that use 800–1,000 MW of power. And Google is building dozens of them. Cleanview now tracks 63 Google data center projects across 20 U.S. states. Finding enough electricity to power all of this development has become arguably the company’s most important operational challenge.

Until recently, Google’s power strategy was straightforward. The company worked with utilities to secure grid connections and then contracted with clean energy developers to match its electricity consumption. Google has spent more than a decade building a reputation as a climate leader on the strength of that model. It was the first major tech company to match 100% of its electricity consumption with renewable energy purchases. It pioneered the concept of 24/7 carbon-free energy. And it has signed more than 22 GW of clean energy power purchase agreements.

But our latest report—and the documents we obtained to produce it—reveal a more complicated story and an evolving power strategy.

In November 2025, Google announced its plans to invest $40 billion in Texas data centers. The company said that investment would fund new data center campuses in Armstrong County and Haskell County. In Armstrong County, Google confirmed to Cleanview that it has partnered with Crusoe Energy to develop a campus known as ‘Goodnight,’ near the town of Claude.

In January 2026, Crusoe filed a permit to build a 933 MW natural gas power plant onsite at the Goodnight campus. Documents obtained by Cleanview show that the gas plant would power two buildings on the campus and would not connect to the power grid. The gas turbines could emit as much as 4.5 million tons of CO2 per year. Cleanview commissioned high-resolution satellite imagery and confirmed the two buildings that would receive power from the gas plant are under construction.

When we reached out for comment, Google confirmed it has partnered with Crusoe to build the Goodnight campus. A Google spokesperson said the company hasn’t signed an offtake agreement with Crusoe for the natural gas plant. Someone familiar with the partnership between the two companies described the negotiations as ongoing and said the amount of power Google would use from the gas plant is still to be determined.

The Goodnight power plant is notable not just for its scale, but for what it lacks. In 2025, Google announced a first-of-its kind agreement to buy power from a 400 MW natural gas plant with carbon capture and storage (CCS) in Illinois. Documents obtained by the Flatwater Free Press suggest Google is exploring a proposed 1,000-3,000 MW gas plant with CCS in Nebraska. (Google told Cleanview they haven’t signed a contract for the project). Unlike these projects, Goodnight’s natural gas plant has no carbon capture technology.

To power its data centers, Google is also investing more in clean energy technology than almost any other entity—public or private—in the world. Since early 2025, the company has signed several gigawatts of new solar and wind PPAs across PJM, ERCOT, SPP, and MISO. It is paying $3 billion to keep aging hydroelectric dams on the grid in Pennsylvania. It is restarting Iowa’s only nuclear plant under a 25-year PPA. And it has signed first-of-their-kind offtake agreements for enhanced geothermal, 100-hour iron-air batteries, advanced nuclear fission, and even fusion energy.

Google is also finding innovative ways to use co-located renewables to connect its data centers in as little as 18 months. In February we published a report on Google’s partnership with AES in Texas, where it will co-locate a 850 MW data center with 600 MW of solar and 945 MW of wind and bypass the crowded ERCOT large-load queue in the process. In this report, we found another example of the strategy that hasn’t been reported on—this time with its soon-to-be subsidiary Intersect Power. Across these projects, Google is bringing nearly 2 GW of data center load online in 2027–2028, powered by co-located wind and solar, without entering the large load queue at all.

Our report shows that Google has pivoted from a grid+renewables strategy to what can be described as an “Everything Everywhere All at Once” strategy when it comes to powering its AI models and compute—one that includes everything from solar and wind to natural gas power.

Get the full report and dataset

The complete ~50-page report includes our analysis of 63 Google data centers across 20 states, detailed case studies, and the company’s evolving power strategy in the AI era. You’ll also get two data files—one covering all 63 data center sites and another with approximately 25 power deals and partnerships we identified.

Get the full report

Get the free excerpt

We’re releasing a free excerpt focused on one of the report’s most significant findings: The role of natural gas in Google’s data center strategy. You can read the excerpt here.

Microsoft just announced that it will deploy 1.35 GW of Nvidia’s latest chips at an off-grid data center in West Virginia that will run on 100% natural gas. The massive deployment could boost the company’s data center emissions by 40% based on my estimates.

In Cleanview’s latest report, we identified 46 data center developers that were planning to build their own power plants. One of the strangest projects we came across was Monarch Compute Campus, an 8 GW project in West Virginia.

The company behind the project had never built a data center and had no tenant disclosed. Yet we found a 2 GW order for Caterpillar generators to power the first phase. I didn’t know what to make of it.

Yesterday, Nscale—a neocloud backed by Nvidia—announced it has acquired the Monarch project and signed a deal with Microsoft to deploy 1.35 GW of Nvidia’s Vera Rubin NVL72 GPUs.

Nscale says it plans to bring the first 2 GW of its project online by 2028. It wants to build 8 GW of capacity by 2031.

Before working on our behind-the-meter data center report, I would have been skeptical of this timeline. But in that research, we learned that developers are getting around grid delays and turbine shortages by purchasing equipment like the Caterpillar gensets Monarch will use.

It’s difficult to overstate the scale of this project. 8 GW is a bit less than the entire peak power demand of New York City.

That scale will come with major environmental impacts. Microsoft’s 1.35 GW deployment could result in 3.9 million tons of CO2 emissions per year by my estimate1. Last year, the company’s entire fleet of data centers emitted about 10 million tons. So this deployment would boost emissions by 40%.2

If the project scales to 8 GW, it could emit 23 million tons per year. That would represent about 35% of the entire coal-heavy state’s emissions.3

In our report, we found 45 more data center projects like this that would add a combined 56 GW of capacity. 90% of those projects were announced last year. With Microsoft getting in on the action, I wouldn’t be surprised if that number continues to climb this year.

Read more about behind-the-meter data centers

To read the full behind-the-meter report that featured the Monarch Compute Campus, you can head to Cleanview’s website. We’ve released both a free summary and a ~50-page paid version with two datasets. We also have a discounted option for nonprofits and researchers.

Read the report

Amazon and Meta have said they are building gas plants to power their data centers because it’s the fastest path to power. But this week, Google proved you can do it even faster with co-located renewables. And I found documents showing their strategy.

On Wednesday, Google announced a new data center in Texas that will be powered by renewables built by AES Clean Energy. The press release was light on details, so I used Cleanview’s platform to try to learn more about the project.

In December AES filed a document showing that it plans to connect an 850 MW data center (Google’s) to its massive solar and wind project in West Texas.

The project would use 600 MW of solar and 945 MW of wind power. Using both solar and wind enables near round-the-clock clean energy. And by connecting to the grid, Google gets the reliability it needs when solar and wind output drop.

But the creative part is how this deal enables Google to skip Texas’ large load queue and get online in 18 months instead of 5+ years.

Like the rest of the country, Texas has a massive backlog of data centers trying to connect to its power grid. At the end of 2025, the backlog was 225 GW—equivalent to 20 New York City’s of power demand.

For data center developers like Google that backlog means waiting years to connect to the grid. These delays have led some companies like Meta to start building their own gas power plants as I wrote in Cleanview’s latest report.

But Google found an alternative path—one that relies on a huge amount of onsite renewable energy.

Thanks to a recent rule change, a data center in Texas can piggyback off a power project’s interconnection agreement if its co-located. And that’s what Google appears to be doing here with AES. The documents we found suggest Google is using AES’ grid connection, which took years to secure to get around the ERCOT large load queue.

The wind phase of the project is expected to come online in August 2027. If Google had gone the traditional route, there’s no way they could have achieved that timeline. If they had tried to connect this data center in Virginia, they would have had to wait until the early 2030s.

It’s worth noting that this timeline is similar to the one Amazon and Meta are achieving by using natural gas. They’ve argued that they have to use gas because waiting for renewables delivered through the grid would take too long.

But with this project, Google is proving that it’s possible to build a 850 MW data center in 18 months powered almost entirely by co-located renewables.

Developers and policymakers should take note.

Share

We wrote more about this project in a brief for Cleanview research subscribers. That brief includes a detailed project timeline, the equipment being used, and the broader policy and market context. Visit our website if you’re interested in becoming a subscriber.

Many data centers claim to use clean energy to power their operations. But in a report Cleanview published today, we found that’s increasingly not true. Instead data centers are using natural gas—and doing so in very strange ways.

It can now take as long as 7 years to connect a data center to the power grid. Beginning about a year ago, developers began pursuing new power strategies. Rather than wait, many data centers are now building their own power plants.

In what we believe is the most comprehensive analysis of this trend to date, we identified 46 data centers with a combined capacity of 56 GW that plan to build their own power “behind-the-meter.” That represents roughly 30% of all planned data center capacity in the United States, according to Cleanview’s project tracker.

In the last year, this trend has gone from niche to mainstream. 90% of the projects we identified—representing approximately 50 GW—were announced in 2025 alone. A year ago, behind-the-meter data center power was a curiosity, embodied by xAI’s controversial decision to truck mobile generators into Memphis. Now it’s an increasingly common development strategy.

When we began this research, we were skeptical of many of these projects—as all analysts should be. Data center developers often pursue multiple projects with the intention of only building one (the “phantom project” phenomenon). Turbine manufacturers have said lead times for their equipment now stretch as long as 5-7 years. Local opposition to data centers is rising.

But we think much of this capacity is likely to come online soon. Many of the projects we identified are already under construction—in some cases with crews working through the night. Many of those that aren’t under construction have been permitted and have placed equipment orders.

We were able to identify specific equipment deals at approximately two-thirds of the projects we tracked—commitments to manufacturers like GE Vernova, Caterpillar, Siemens, and Doosan that signal serious intent.

What makes our report unique is that we didn’t rely on press releases, which show what developers say they are going to build. Instead we tracked down actual equipment deals and permits showing site plans.

This revealed a very different—and surprising—story. Most of the press releases we found mentioned “all of the above” strategies that include renewables. But ~75% of the generation equipment we could identify (23 GW) was natural gas-powered. Virtually none of the developers planned to build renewables in the short term.

But data centers aren’t planning to use your typical gas turbines—hence why many are able to install them this year or next year. Data center developers are instead turning to:

• Mobile gas generators strapped to semitrucks

• Aeroderivative turbines originally designed for aircraft and warships

• Reciprocating engines that ramp fast, but are less efficient

• Refurbished turbines acquired from industrial operations

In our research, we came across a company that typically sells cruise ship engines that struck a deal to power a data center. One developer, unable to secure enough conventional turbines, placed a $1.25 billion order with Boom Supersonic—a company that has never sold a power generation product. Elon Musk’s xAI famously drove in semitrucks with natural gas generators on the back to build what was at one time the world’s largest data center.

On the surface this makes no sense. These are less efficient technologies and the power will cost far more. But an AI data center can earn as much as $10-12 billion per GW. Getting online a few years early can result in a windfall. Power generation efficiency is out. Speed to power is all that developers care about.

As a company, our main job is to track data centers and power projects. Still, much of this data shocked us. The public narrative is that data centers are waiting for grid connections and 5-7 year turbine backlogs. But that narrative is lagging what is actually happening on the ground in rural counties across the country.

To read the full report, you can head to our website. We’ve released both a free summary and a ~50-page paid version with two datasets. We also have a discounted option for nonprofits and researchers.

Read the report

Today, Cleanview is closed for new business in solidarity with those protesting the violence of U.S. Immigration and Customs Enforcement (ICE) in Minnesota. Recently, federal agents and officials have:

• Killed two US citizens exercising their First Amendment rights—Renée Good, a mother of three, and Alex Pretti, an ICU nurse—and then called them terrorists.

• Broken into the homes of US citizens without judicial warrants, breaching the Fourth Amendment, and held them without due process.

• Questioned Americans' right to legally carry firearms—a Second Amendment right—which the NRA has called "dangerous and wrong."

All of these actions are a threat to freedom and democracy in America. That’s why we’re joining Americans from all sides of the political spectrum in condemning these actions.

We’ll open up our “digital doors” for business tomorrow.

If you would like to join me in protesting these actions, you are more than welcome to copy and paste (and edit) my statement and share it with your customers or community.

As I wrote earlier this week, when most Americans learn about what's happening in Minneapolis, they don’t approve of it. They see the cruelty and abuses of power for what they are.

But many people aren't hearing about these injustices. That's why I'm choosing to use our platform as a company and my voice as an individual to speak up.

Every American should know what’s happening right now.

Share

Read more from Distilled

This is a very new trend. A little more than a year ago, virtually all data center developers planned to use the electric grid to power 100% of their projects. There were a lot of advantages to this—mainly the fact that you didn’t need to go into the power plant business if you wanted to stick some computer chips in a warehouse. You could just pay a utility a couple cents per kilowatt hour to run your operation.

In December 2024, there was less than 2 GW of planned behind-the-meter data center capacity, according to our data center tracker at Cleanview. By the end of 2025, developers had announced roughly 40 projects that planned to skip the grid partially or entirely with a combined 48 GW of capacity.

All of these projects are motivated by the same goal: getting their data center online as soon as possible.

It can take as long as 7 years to connect a hyperscale data center to the grid in a place like Virginia now. Building your own behind-the-meter power plant in a red state with lax regulations can get that time down to less than 2 years.

Some of these projects will soon be home to America's largest fossil fuel power plants, like Homer City Energy Campus in PA—a proposed 4.5 GW natural gas plant that will send all of its power to an onsite data center. The project has already been permitted and secured its gas turbines. It expects to start delivering power by 2027.

Other projects will use a combination of technologies—everything from solar, wind, batteries, and even nuclear. Natural gas is by far the most common, though. 72% of projects plan to use it.

But speed comes with a cost. Homer City's 4.5 GW project could soon become one of the largest single sources of carbon emissions in the country, emitting millions of tons of CO2 each year.

At Cleanview we're tracking more than 30 projects that plan to use onsite gas with a combined 48 GW of capacity. Virtually all of those projects were announced in just the last 12 months.

Share

Other trends and stories I’m following

On the potential for an AI power glut

Earlier this month, Tracy Alloway of the excellent Odd Lots podcast wrote in her newsletter:

In a new report, CreditSights analysts led by Andy DeVries crunch the numbers to show that US utilities already appear to be in the process of building twice as much power capacity as is currently forecast to be needed by 2030. In other words, the data center buildout already looks to be outstripping demand for compute.

On the demand side: Using third-party estimates for future power demand from data centers (including estimates from BloombergNEF), the analysts come up with 94 gigawatts (GW) needed by 2030 and 138 GW by 2035. Subtracting the power that’s already in use today, leaves you with demand that’s expected to grow by a net 59 GW and 103 GW, respectively.

On the supply side: The analysts compile data center grid connection requests as reported by the actual companies doing the hookups, which are the utilities. Utilities generally report both firm commitments and more speculative pipeline projects, but CreditSights is only using actual committed projects here (which would make, if anything, the estimate of supply quite conservative). These connection requests are adjusted by a Power Usage Effectiveness factor to try to isolate the compute portion from other power needs like cooling, lighting etc. Put it altogether and you get 112 GW of firm supply capacity already lined up (even without counting that much larger supply of projects that might be in the pipeline).

It’s worth noting that the supply figures here are in addition to the “behind-the-meter” projects I wrote about above.

As I’ve written often on this newsletter, forecasting is difficult. There’s huge uncertainty around the demand for AI, its future efficiency, and the ability of any of these sluggish utilities to actually build these power plants. But history definitely suggests that some degree of overbuilding is a likely outcome. How much exactly is the trillion dollar question.

Mark Carney’s speech heard around the world

In his Davos speech last week, Canadian Prime Minister, Mark Carney did what few other powerful people have done in the last year: He stood up to Donald Trump and challenged his power. The full speech is worth watching due to its significance. But I’ll also say it was just a well-written, almost poetic, speech.

Carney’s speech, and this excerpt in particular, actually inspired me to write my post about the killings and injustices in Minneapolis:

In 1978, the Czech dissident Václav Havel, later president, wrote an essay called The Power of the Powerless, and in it, he asked a simple question: how did the communist system sustain itself?

And his answer began with a greengrocer.

Every morning, this shopkeeper places a sign in his window: ‘Workers of the world unite’. He doesn’t believe it, no-one does, but he places a sign anyway to avoid trouble, to signal compliance, to get along. And because every shopkeeper on every street does the same, the system persist – not through violence alone, but through the participation of ordinary people in rituals they privately know to be false.

Havel called this “living within a lie”.

The system’s power comes not from its truth, but from everyone’s willingness to perform as if it were true, and its fragility comes from the same source. When even one person stops performing, when the greengrocer removes his sign, the illusion begins to crack.

Friends, it is time for companies and countries to take their signs down.

Trump’s gift to China

Trump’s chaotic and aggressive trade policy was initially seen as a threat to Chinese manufacturing and exports. But now, it appears that it could have been a great gift as US allies strike deals with the country.

Earlier this month, Canada’s government announced a trade deal with China to import its cheap electric vehicles. Per the BBC:

The deal will see Canada ease tariffs on Chinese electric vehicles that it imposed in tandem with the US in 2024. In exchange, China will lower retaliatory tariffs on key Canadian agricultural products.

With the lower EV tariffs, approximately 10% of Canada’s electric vehicle sales are now expected to go to Chinese automakers, said Vivek Astvansh, a business professor at McGill University in Montreal.

Canada isn’t the only country cutting deals with China or absorbing it’s overcapacity. China’s trade surplus rose in 2025, even as exports to the US fell. As the New York Times reports:

China’s surplus, the value of goods and services it sold abroad versus its imports, reached $1.19 trillion, an increase of 20 percent from 2024… For December alone, China’s surplus reached $114.14 billion, propelled by surging exports to the European Union, Africa, Latin America and Southeast Asia.

A few shows that have brought me joy recently

Over the last few years, I’ve come to love Cleo Abram’s YouTube series: Huge If True. Cleo’s optimism and curiosity are infectious. The visuals are beautiful. And the topics are often about scientific breakthroughs or research I had never heard of.

More recently—as in within just the last few days as I’ve been sick with the flu—I stumbled upon the YouTube show Track Star, which had a moment in 2025. On the show, host Jack Coyne gives famous musicians a pair of headphones, plays a song, and asks them to name the artist.

As the New Yorker’s David Remnick said in an interview, “It’s a sophisticated trick.” Coyne does a lot of research before every episode so he often plays songs of significance to the interviewee. “He’s making you tell a little bit about yourself and your life and your obsessions and your pleasures, maybe your heartbreaks. It is, in this ugly world, a relief and a pleasure to see people talking about what makes them feel something very deeply.”

This show with Jon Batiste brought me a lot of joy:

This weekend, federal agents killed Alex Pretti, a nurse who was filming their activities. The Trump administration defended their actions saying that the man was a domestic terrorist who was attacking ICE agents.

This is an insult to every American’s intelligence. The killing was filmed by two witnesses and both videos show that Pretti had one hand up and another on his camera when he was taken to the ground. Federal agents then fired 10 shots into him.

Pretti had a handgun on him at the time he was killed, which the administration is using as evidence that he was a terrorist. But the right to bear arms is a constitutional right. Republicans frequently made this point in 2020 when Kyle Rittenhouse showed up to a protest in Wisconsin with a AR-15 and shot two people.

The killings of Pretti and Renee Good earlier this month aren’t the only injustices that have taken place in Minneapolis.

The Trump administration has detained US citizens, thrown children in jails apart from their parents, and entered homes without warrants. When peaceful protesters have filmed these activities, federal agents have threatened to put them on lists and hurt them.

Earlier this month ICE agents broke down the door of a US citizen, put a gun to his head, and dragged him out in the cold in his underwear—all without a warrant.

Since July, ICE agents have held more than 2,300 people across the US without due process, according to Politico’s Kyle Cheney.

Peacefully protesting is a 1st amendment right; bearing arms, which I wish was less common, is a 2nd amendment right; and requiring a warrant to enter a home is a 4th amendment right.

The Trump administration is challenging all of these constitutional rights. And they are lying to the American people about it, claiming that peaceful protesters are terrorists.

None of this is about the issue of immigration. Democrats and Republicans can disagree on border enforcement or deportation policy. That’s a sign of a healthy democracy. But what’s happening in Minneapolis is something else entirely: it’s the stuff of a Founding Father’s nightmare.

Most Americans don’t approve of what’s happening in Minneapolis either. Since the federal government sent thousands of agents to the city to sow chaos, Trump’s poll numbers on the issue of immigration have fallen significantly.

When people see the videos—when they watch what actually happened to Alex Pretti and Renee Good—they recognize that something has gone deeply wrong. That’s why I think it’s important for all of us to talk about what’s happening and simply describe what we’re seeing with our eyes.

Every American should know what’s happening right now.

But there was one notable exception: In California, carbon emissions fell by 8%.

The decarbonization of California’s power grid has been driven by huge growth in solar and battery capacity in recent years. The state has added more than 300 large-scale solar projects and more than 200 large-scale battery projects in the last 5 years, according to Cleanview’s project tracker. Small-scale solar and battery adoption has also been significant.

The growth of solar and batteries is now cutting into the state’s fossil fuel power plant generation. Since 2023, natural gas generation has fallen by 10%.

Much of this fossil fuel displacement is happening in the afternoon and evening hours when electricity demand is highest. Batteries that charge up on abundant solar power during the day now discharge during these evening peaks when natural gas peaker plants would otherwise fire up.

The state still has a long way to go in its efforts to reach 100% clean power, but the trends are going in the right direction. That wasn’t true as recently as just a few years ago. The state’s electricity emissions were stubbornly flat for much of the last decade.

The growth of solar and batteries wasn’t the only factor behind California’s decline in emissions last year though. Falling electricity demand also helped. Homes and businesses consumed 2.4% less electricity in 2025 than they did the year before.

Here the state was an outlier, too. On every other major US electric grid—in Texas, in the Midwest, in the Northeast—electricity demand rose. In Arizona, it rose by a staggering 8%. That demand growth, which was driven in many places by data centers, meant that coal and gas plants had to run more frequently.

In California there’s been very little growth in data centers or other industrial sources of electricity demand. The state’s high electricity prices and slow permitting process have made it unattractive to hyperscale data center developers. (Slow permitting has also plagued renewable projects for decades, slowing the energy transition).

Here’s a map showing all the planned data centers in the US from our project tracker at Cleanview. Note how few are in California relative to places like Texas, the Midwest, and the Mid-Atlantic.

California’s real test will come when its electricity demand begins to rise again—which is likely in the near future.

The state has ambitious plans to electrify its transportation and buildings, putting millions of electric vehicles on the road and heat pumps in homes. That electrification will be essential to cutting emissions economy-wide, but it will also mean more demand on the grid.

Once demand starts climbing again, California will face the same challenge every other US grid struggled with last year. To decarbonize in that environment will require deploying clean energy even faster.

Share

More from Distilled

At all hours of the night, in windowless buildings scattered across Phoenix’s sprawl, machines that never sleep now hum away, drawing power around the clock. Hundreds of miles of yellow fiber optic cables connect thousands of cabinets filled with these machines—each one serving up a video, ChatGPT query, or newsletter like this.

Few places in the world are seeing more demand from data centers than Phoenix. In the last few years, developers and tech companies have built more than a dozen large-scale data centers. The combined capacity of operating data centers in Arizona is now more than 2,000 MW, according to Cleanview’s data center tracker.

But that’s nothing compared to the growth that is likely to come over the next decade.

Arizona's three major utilities are grappling with data center interconnection requests that would dwarf their existing grids. Arizona Public Service (APS)—which currently peaks around 8,200 MW—has 30,000 MW of data center requests sitting in their queue.

"We've never sat in a position before where somebody's asking you to triple the size of your company," Jacob Tetlow, APS's Executive Vice President, recently told KTAR News. APS's average data center request now stands at 500 megawatts—enough to power roughly 375,000 homes.

Already data centers are causing electricity demand to rise faster in Arizona than almost any other place in the country. In 2025, electricity demand in Arizona grew by 8%—a growth rate 4x larger than that of the United States as a whole.

Utilities have received so many requests from data center developers that they’ve had to start turning some away for fear of breaking the grid.

“The utility cannot commit to serving them because it would put existing customers at risk of having poor reliability,” Jose Esparza, APS Senior Vice President of Public Policy, told a NARUC conference in November 2024.

Utilities have found a way to accommodate many of the requests, however. Nighttime electricity demand—a good proxy for data center demand in Arizona—grew by 45% between 2019 and 2025.

Data centers’ round-the-clock demand for power has made decarbonization in Arizona more difficult.

In the last decade, solar power generation has grown by more than 400% in the state. The state has added 4,748 MW of battery storage capacity—equivalent to the peak output of 5 nuclear power plants. That’s helped the state cut its emissions significantly since 2010.

But solar and short-duration batteries provide hardly any power to these machines when it’s 3am. That’s one reason why, after falling for more than a decade, Arizona’s electricity emissions have risen by 3% since ChatGPT launched in November 2022.

The data center surge has driven Arizona utilities to walk back decarbonization commitments. In August 2025, APS abandoned its 2050 "100% clean energy" commitment, replacing it with an aspirational "carbon neutral" goal that allows offsets. The company also eliminated interim targets and extended operations at the Four Corners coal plant to 2038.

Utilities are also planning large investments in natural gas power plants. When it announced that it would renege on its climate targets, APS announced a new planned 2,000 MW gas plant.

“Additional natural gas generation is essential to support our existing customers and to begin addressing unprecedented requests from extra-large energy users, such as data centers,” Tetlow, the company’s CEO, said.

Data center water use has emerged as another flashpoint in Arizona’s arid landscape. “This has been the driest 12 months in 126 years. We are on red alert, and I think data centers are an irresponsible use of our water,” Mesa Vice Mayor Jenn Duff told NBC News.

Some tech giants are responding by eliminating water cooling. Joe Kava, Google’s VP of Global Data Centers, announced the company’s Mesa facility would use no water for cooling. Microsoft has said its data centers “will use zero water for cooling for more than half the year, leveraging a method called adiabatic cooling.”

This will put less pressure on scarce water resources, but it will push electricity demand up even higher. Companies use water cooling more in order to use electric cooling systems less. Take away the water cooling systems and each data center will use more electricity. In a place like Arizona that might be the right choice, but it’s a tradeoff nonetheless.

Some communities have chosen to block data centers altogether.

In December 2025, Chandler’s City Council voted 7-0 to reject a $2.5 billion AI data center despite lobbying from former Senator Kyrsten Sinema. A few months earlier, Tucson’s City Council unanimously halted Project Blue, a proposed $3.6 billion, 600 MW Amazon-linked data center that would have consumed 15% of Tucson Electric Power’s current generation capacity.

“In addition to the water use, this was a problem because of the energy use,” Councilmember Kevin Dahl said.

Tech companies and data center developers continue to push ahead, despite the opposition. Google, Microsoft, Meta, and Apple are all expanding their data center operations in Phoenix.

A proposed 1.5 gigawatt Hassayampa Ranch development backed by investor Chamath Palihapitiya received zoning approval in December 2025. The developer of the project told Fortune, “We have probably six to eight large hyperscalers that are interested in looking at it.”

Share

Read more from Distilled

In the first 6 months of this year alone, China installed a similar amount of solar capacity.

In addition to being a world leader in solar, China also has the world’s largest coal fleet, which continues to grow in size. But the country’s investments in solar and wind have allowed Chinese utilities to run those coal plants less frequently.

For more than a decade, clean energy’s share of total electricity generation has been rising. Here’s a chart from Kyle Chan’s excellent vibe-coded dashboard using Ember data showing that trend:

China is also doubling down on electric vehicles. The country has been making investments in the sector for more than a decade now. Those investments have helped China become not just the largest EV producer in the world, but the largest producer of all vehicles.

In 2025, while Elon Musk was gutting investments in US foreign aid through DOGE that would have saved millions of lives, Wang Chuanfu—the CEO of China’s top EV maker, BYD—was quietly opening factories and showrooms in countries around the world. Musk’s embrace of far-right politics and Nazi parties in Europe alienated millions of customers and destroyed what was once one of the world’s best brands.

Tesla’s sales fell 9% in 2025, while BYD’s grew 28%. By year’s end, BYD had outsold Tesla by over 600,000 electric vehicles, becoming the world’s top EV maker.

China, and the companies that operate in the country, are far from perfect. Even if the country were a model global actor, it would be difficult to recreate the country’s engineering-led culture in America.

Still, there are lessons that America and other countries should pay attention to. For me, one of the main lessons in the Chinese clean energy story is that industrial policy can work, but only if committed to over decades.

Share

Read more

I wrote a much longer story about the history of China’s EV industry and industrial policy here. It focuses on the battery manufacturer CATL, but it’s a good history of the entire EV sector.

To learn more about China’s clean energy industrial policy, I recommend Henry Sanderson’s Volt Rush. My favorite book on China, in general, is Evan Osnos’ Age of Ambition.

When I think of whose jobs are most threatened by AI, I tend to think of low-income earners. AI, and automation more broadly, are already displacing call center workers, cashiers, and ride-share drivers.

That’s why this chart from a recent Wharton Budget Model study surprised me. It shows that some of the highest income earners are at most risk, while the lowest income earners are safest.

As the authors of the study write, “Occupations at the bottom of the wage distribution are the least exposed to AI, since many of these jobs are predominantly manual labor or personal services. Exposure generally rises with earnings until the 80th-90th percentiles, which include programmers, engineers, and other professionals.”

The study argues that half of the work of these high income earners can be done by AI today. The researchers found that in more than a quarter of US employment, AI can perform between 90 and 99 percent of the work required with minimal oversight.

I know many readers will find these numbers absurd. Google’s AI overviews struggle to answer basic questions like “What year is it?” ChatGPT only recently figured out how to count the number of r’s in the word strawberry.

But I think the estimates in this study actually understate the potential labor disruption.

If you look at the research this study used to estimate AI’s capabilities, you’ll notice many used old models like ChatGPT 3.5 and 4 and Github’s Copilot (circa 2023). Most studies on the same topic use the same underlying research.

These tools were primitive compared to today’s models, though. They didn’t have web search access, reasoning capabilities, or access to every file on your computer like today’s paid models do.

I’ve been thinking about all of this as I’ve been using Claude Code in my own work at my company Cleanview. It is consistently doing tasks in hours that used to take me days. It built most of our free project explorer that I shared on this newsletter last month. That project took Claude a couple hours. A similar project took me two weeks a year ago.

I always hesitate to write about my own experiences with these tools. I find that I have an almost allergic reaction to stories like the one I just shared given that so many come from AI hype bros that don’t share many of my values or concerns for the impact of these tools on society and our planet.

But I’ve noticed that many people who don’t use these tools—often out of moral disgust with the way they were illegally trained on creatives’ work, or their environmental impacts, or their potential to enrich billionaires—are overlooking or underestimating their true capabilities.

I see many people online cite the viral MIT study that tanked the stock market temporarily last year as evidence that these tools aren’t capable. That study found that “95% of AI pilots at companies fail.” But that was one study with a small sample size. Many other studies have found AI offers huge productivity gains and is increasingly able to replace jobs.

The way some people dismiss AI’s current capabilities worries me because many of these critics have influence and power that could shape the trajectory of all this. If you don’t believe these tools can displace tens of millions of workers, why think about policies that could protect or support them?

For much of the last three years, I’ve oscillated between thinking that AI is transformative to believing that it’s mostly overhyped. At times, I’ve felt frustrated that I can’t land on a single position. But more recently, I’ve come to embrace that ambiguity and uncertainty.

In a world that is complex and fast-changing, the alternative feels absurd.

What I’ve been reading

I’ve been reading a lot about AI’s potential impact on labor markets recently. I’ve really enjoyed these two books:

• A World Without Work: Technology, Automation, and How We Should Respond

• The Technology Trap: Capital, Labor, and Power in the Age of Automation

If you’re looking for an entry point into the topic, as I was, I recommend starting with the first one. If you want a wonky, long-sweeping history of automation, start with the second.

If you’ve read any good books or articles on this topic, I’d love to hear about them. Please send them my way.

More from Distilled

Electricity demand grew by 2.6% in 2025 due to the growth of data centers, electrification, and extreme heat among other factors. It was the second consecutive year of growth after decades of flat demand.

Renewable energy’s contribution to the power grid rose again in 2025, but not enough to offset the country’s rising electricity demand.

Collectively, the country’s solar and wind projects produced 73.5 million MWh more electricity in 2025 than they did the year before. That’s as much electricity as about 7 million US homes use in a year. Solar’s output grew by 31%, while wind had another sluggish year, only growing by 2%.

Total electricity demand grew by 107 million MWh—or about 10 million homes worth. Electricity generation from fossil fuels filled the remaining gap, rising by about 32 million MWh.

Strangely, most of the growth in fossil fuel power generation came from coal-fired power plants. Coal generation rose by 80 million MWh—more than solar and wind combined.

Trump can take some credit for bringing coal back. Throughout the year, the administration forced utilities to keep operating old coal plants that were supposed to retire, in most cases against the wishes of the utility and at significant cost to consumers.

But the more important factors behind the resurrection of coal, which had long been in decline, were high natural gas prices and growing electricity demand. “The power sector’s consumption of coal has increased this year as the average cost of natural gas in the electric power sector increased over 40% compared with 2024. More overall electricity demand has also supported coal consumption this year,” analysts at the Energy Information Administration wrote at the end of last year.

Coal and natural gas power plants are always competing by bidding into electricity markets across the US. When gas prices are higher, coal plants win those bids more often, generating more electricity and emissions.

Carbon emissions from the power grid rose by 4.4% in 2025, due almost entirely to the growth of coal power generation. That’s more emissions growth than any year since 2021, which was a weird COVID recovery year.

For much of the last decade, the power grid has been a rare example of a huge sector getting cleaner by the year. The story of America’s decarbonization since about 2010 has really been a story of grid decarbonization. Basically no other part of the economy of significance has seen much progress.

But that decarbonization happened during a period of falling electricity demand. When solar and wind produced more electricity, they cut into a shrinking pie, and cut emissions in the process.

2025 proved that we’re no longer in that era.

More from Distilled

• This newsletter recently turned 3 years old. To everyone who has read my work—and especially to paid subscribers who have supported it financially—thank you. Your support and kind words over the years have meant the world to me.

• We’ve made a big impact together. In 3 years, these stories have reached tens of millions of people. According to a literal academic study on Distilled’s work, these stories change people’s mind and makes them more open to sustainable solutions. This summer, a Distilled analysis was used on the floor of the US Senate.

• Going forward, I’m going to be writing about more topics than climate change and clean energy. This newsletter has always touched on politics, economics, technology and much else. Beginning in 2026, I plan to do so more directly by expanding the breadth of coverage.

• I’m going to publish more short posts in between my longer stories. These stories will be more visual and quicker to read. You can read the first one here.

The longer version

A personal highlight from 2025

When I started Distilled, I set a goal for myself that felt ambitious at the time: Produce work that influences how Congress writes clean energy policy. This year, that goal became a reality.

This summer, Trump and Congress set out to dismantle many of the clean energy policies originally passed in the Inflation Reduction Act. The first version of the bill passed by the House in May was bad. It aimed to phase out clean energy tax credits within months. But all the pundits believed that the “cooler heads” of the Senate would pass a better bill. They were wrong.

One Saturday morning in June, I woke up and saw that the Senate released their version of the big bill. I woke up and quickly read through the energy provisions. It was immediately clear that the bill would be even more disastrous than the House version.

The bill threatened to cut off America’s fastest-growing sources of electricity—solar and wind—at a time when the country was desperately in need of new power. In addition to removing tax credits, it actually created a new tax on solar and wind projects.

Over the last two years, I’ve been building a dataset of ~10,000 planned clean energy projects through my company Cleanview. That work enabled me to quickly run an analysis to see how many of these projects would be at risk if the bill passed.

The number was staggering. Republicans’ poorly designed policy threatened a wave of energy project cancellations, hundreds of billions in lost investment, and electric grid blackouts by the end of the decade.

I quickly wrote all of this up and by 10am that morning I published the results to this newsletter and on social media. Then I started texting everyone I knew who had any connections to Congress. Within 30 minutes I got connected to a Congressional staffer, who then connected me to three Senators’ energy policy advisors. These folks used the analysis as part of a much larger effort to lobby moderate Senators.

Later that day, I went to lunch with a friend. Halfway through lunch my phone started buzzing with text messages. Elon Musk had shared my analysis on Twitter, writing, “This would be incredibly destructive to America!”

That was enough to boost it to the top trending news posts on Twitter for most of the day. The New York Times then published a story at the top of their home page about how Elon Musk had come out against the bill citing my post and another from Princeton’s Jesse Jenkins.

The next day, Sen. Brian Schatz (Hawaii) used the data in his floor speech arguing against the bill. The staffers that I spoke to told me that they used it in conversations with moderate Republicans.

Ultimately, some of the worst clean energy provisions of the bill—including a new tax on solar and wind farms—were taken out. Modeling from Princeton shows that those last minute changes will be enough to cut 65 millions of CO2 emissions every year compared to the initial version.

That small win was without a doubt the result of work by hundreds of people, including the public servants in Congress who stayed up all night fighting to make a very bad bill a little less bad. Playing a small role in that effort was one of the highlights of my career.

I’m learning to trust in the process more

As I shared in a post from earlier this year, I’ve always doubted the impact that newsletters and social media posts can really make in the world. But between these moments of doubt, has come proof that stories do matter.

One entrepreneur, who left his job at Google and eventually raised $33 million to build a better heat pump, told me that one of my stories encouraged him to start looking into heat pumps in the first place. A Congresswoman told me that my story encouraged her to launch an investigation into a company spreading disinformation online. Countless others reached out to tell me about the action they were taking in their homes and in their communities.

Over time, I’ve come to trust in the unpredictable process of all this more. And I’ve come to see my work as something like an extended and asynchronous conversation with the world around me—one from which I can learn and then share what I learn with others.

Where I want to take this newsletter going forward

Since starting this newsletter, I’ve always planned to write about more than climate change and clean energy. Gradually over the coming months, I plan to do that.

The main reason that I want to write about more topics is simple: I’m endlessly curious about the world around me. I love learning and this newsletter is first and foremost an excuse for me to learn.

The second reason is a bit more abstract. Ever since the 2024 election, I’ve had this nagging feeling that what I write is largely reaching the wrong audience. Positioning this as a climate newsletter has attracted wonderful people committed to making the world better—many of whom I’ve had the pleasure to meet. But it tends to attract the “die-hards” who have generally already made up their mind about climate change and clean energy.

I’m not naive enough to think that my writing is ever going to appeal to people who think climate change is some great hoax. Studies on changing people’s minds who have been swayed by disinformation are pretty bleak. But one thing those studies tend to suggest is that the most effect way to combat misinformation is by “pre-bunking” myths and basically inoculating people to the disease that is bullshit.

My hope is that writing a newsletter that is wider in focus will help me reach a broader audience who might not have already made up their mind about the topics I write about. And because I think persuasion is a two-way street, I hope to learn from this broader audience, to have my own mind changed and opened to more ideas and worldviews.

What I’m planning to write about

I’d love to share exactly what I plan to write about in 2026, but the truth is that I’ve never been good at planning more than a month in advance. I’ve created many editorial calendars over the years and I’ve stuck to exactly zero of them. But here’s some of what I’ve been thinking about:

I want to explore how we can make the essential things in society more affordable and sustainable—energy, housing, healthcare, childcare. These are the costs that squeeze families and shape elections. My hope is to explore this from a non-dogmatic and nuanced perspective as I did in my story, “Be Careful of the Lessons You Learn In Texas.”

I’ve also become much more interested in technology since starting this newsletter. Many of the stories that I’ve written have required me to learn about the history of various technologies like solar, wind, and batteries. I don’t think new inventions and innovation can solve every problem, but they can certainly help.

One major subcategory of technology that I plan to write about is artificial intelligence. I’ve written stories about the environmental costs of AI and the degree to which data centers are and aren’t raising electricity prices. I plan to continue this coverage, but also expand into other topics like the impact of AI on society more broadly. As an example, I’m currently working on a story about how AI will impact labor markets.

I should say that my writing about AI will likely diverge from other climate-focused and liberal writers. I’m concerned about the harms associated with AI. But I’m much more optimistic than many of my friends and fellow writers about the technology. I think with the right policies AI can bring massive benefits to all of us. I suspect this perspective will be frustrating to some readers who strongly dislike AI. I hope I can bring you a different perspective and also learn from yours.

Finally, I plan to continue to write about politics and policy—with a primary focus on the US. I think few events will carry more consequence for Americans and the world than the midterm elections. So I plan to write about the debates and issues in the run up to that election.

Thank you

To everyone who has read this newsletter over the years and supported my work, thank you. As always, you can say hello, share what’s on your mind, or suggest story topics and ideas by responding to this email or commenting below.

Lastly, if you aren’t interested in following this next phase of Distilled, I totally understand. I won’t be offended if you unsubscribe. And if you are a paid subscriber and want to cancel or want a refund on your annual subscription, just reply to this email. Regardless, I appreciate your support up until now.

Happy New Year!

- Michael

Here’s what’s in today’s newsletter:

• A new tool to explore clean energy and data center projects

• The man who killed offshore wind in America

• Battery prices fell again in 2025 according to a new analysis.

• A new paper shows that clean energy technology costs don’t change in the way that most forecasters assume—and sometimes technologies get more expensive the more you make them.

• Ford is pivoting out of the EV business and into… data centers.

• A supersonic plane startup is getting in on the data center boom too by turning its jet engines into gas turbines to run OpenAI’s GPUs.

A new tool to explore clean energy and data center projects

This week, my company Cleanview launched a free tool that lets you explore thousands of operating and planned solar, wind, battery, and data center projects in every state across the US.

Here are a few examples of what you can find using the tool:

• An interactive map of every planned data center in the US

• The largest operating and planned solar projects in each state

• Data showing how many battery storage projects have been built in each state

Give it a try and let me know what you think.

Try it out

The man who killed offshore wind in America

3 years ago, I heard a story about a man, David Stevenson, who sent a letter to 35,000 residents on the coast of New Jersey, Delaware and Maryland telling them that a planned offshore wind farm would destroy their property values (which was false).

One weekend, he paid a pilot thousands of dollars to fly an aerial banner over the beach telling people the wind turbines would destroy their beautiful views. He ran Facebook ads, wrote op-eds in newspapers, and did everything he could to tell people that Skipjack Wind would ruin life as they knew it.

Virtually all of this claims were false.

I love trawling through public documents so I started following the paper trail to learn more about this man. I learned that his nonprofit was funded by the American Fuel and Petrochemical Manufacturers and part of a network of groups started by the Koch Brothers.

I tried to get in touch with him but he never answered my calls. So after a few weeks of reporting and research, I published my story in this newsletter.

Immediately it blew up and went viral, racking up millions of views. A group of researchers at Brown University wrote a paper inspired by it. I still get calls and emails about it to this day.

This week, Canary Media published a story by Clare Fieseler profiling this man. It’s a beautifully written, rich, human story.

Clare told me that after reading my story she called David and he picked up. Over the next few years, she had a series of conversations with him to understand his motivations.

Reading the story this morning, I couldn’t help but feel sad.

What started as a 35,000 person snail mail campaign soon blossomed into one of the most damaging disinformation campaigns in the country. Fox News ran many segments with the lies that David initially spread. Donald Trump soon took up many of his talking points. Now many blue-collar communities that were banking on offshore wind to revive their coastal economies are seeing infrastructure projects and federal funding cancelled.

David wasn’t the only person responsible for the collapse of the offshore wind industry in America. But he was one of the most influential.

I highly recommend reading Clare’s story.

Read more

In the state, more solar permits are now denied than are approved. Recently Hanover County officials approved a 2.4 GW data center on the same night they denied a 20 MW solar project.

At least 10 Virginia counties have enacted outright bans or severe restrictions on solar, with Mecklenburg County—the state’s ninth-largest county—passing an ordinance in April 2025 that removed utility-scale solar as an allowed land use entirely.

The same dynamic is playing out across the US.

Officials in states like Ohio and Indiana have successfully courted some of the country’s largest planned data center facilities from Google, Meta, and Amazon—with AEP forecasting that data center demand in Ohio alone will increase from 600 megawatts to 5,000 megawatts by 2030. Yet both states have lost gigawatts of potential power generation capacity in 2025 due to project cancellations.

In Ohio, 6 GW of capacity has been cancelled in 2025—more than AEP’s entire data center pipeline by 2030. In Indiana 5 GW of capacity has been cancelled, equivalent to two of the world’s largest data centers that are under construction in the state.

Ohio has blocked more clean energy projects (221) than any state in the country. Virginia ranks 4th with 57 projects blocked. In 2025, more than a dozen projects were blocked in each state.

In Indiana, 44% of proposed data centers are located in counties that restrict renewable energy development, according to a Heatmap analysis.

Growing electricity demand like this while blocking projects that can supply new power will have one inevitable result: higher electricity prices. At some point, something has to give.

Read the whole report

Over the last few weeks, I analyzed data to see how many power projects have been cancelled this year. I found that 1,891 power projects with a combined 266 GW of generation capacity have been cancelled in 2025—equivalent to roughly one-quarter of America’s entire current electricity generation capacity.

You can read the full report by clicking the button below:

Read the report

Electricity demand is growing faster than it has in decades in the United States. Data centers, manufacturing reshoring, and electrification are driving massive growth in power consumption. But at the moment America needs more electricity generation, the country’s infrastructure developers are cancelling projects at an alarming rate.

Over the last few weeks, we’ve been analyzing data to see how many power projects have been cancelled this year. In our latest report, we found that 1,891 power projects with a combined 266 GW of generation capacity have been cancelled in 2025—equivalent to roughly one-quarter of America’s entire current electricity generation capacity.

Clean energy projects have been hit hardest, accounting for 93% of project cancellations in 2025. Utility-scale solar alone saw 86 GW cancelled, while battery storage projects lost 79 GW and wind projects shed 54 GW.

The Trump administration and their chaotic policies were responsible for many of these project cancellations. But the analysis showed that there’s much more causing project cancellations and preventing the buildout of clean energy than federal policy.

While each region and grid faces unique challenges, five patterns emerged across the 1,891 cancelled projects that we analyzed:

Local opposition is killing projects where electricity demand is growing fastest

Virginia—home to the country’s largest data center market—lost 6.7 GW of potential capacity. In Ohio, where elected officials have courted hyperscale data centers, the state has blocked more clean energy projects than any other. In Indiana, another data center hotbed, 44% of proposed data centers are located in counties that restrict renewable energy development.

A failure to build transmission lines is resulting in high interconnection costs

High-voltage transmission construction has plummeted from 4,000 miles in 2013 to just 322 miles in 2024—less than one-tenth of the 5,000 miles per year needed. In MISO, average interconnection costs for cancelled projects hit $753,116 per MW—roughly half of a typical project’s total capital cost. Louisiana and Missouri, with costs exceeding $900k per MW, lost 26 GW of capacity combined in 2025.

Battery storage economics are worsening from market saturation and tariffs

“Battery cannibalization”—where growing capacity drives down the prices batteries earn—is making it harder to build profitable battery projects. Texas battery revenues crashed 70% from $192/kW in 2023 to $55/kW in 2024, while ancillary service prices fell 90%. California saw similar declines, with revenues falling from $103/kW in 2022 to just $53/kW in 2024. Meanwhile, Trump’s tariffs have increased the cost of batteries by 56-69%.

Trump’s War on Wind killed gigawatts of potential capacity

Since taking office, the Trump administration has systematically dismantled offshore wind development through sweeping policy actions, funding cancellations, and construction halts. These moves led to massive waves of project cancellations in New York (13 GW), New England (11 GW), and California (2.5 GW).

Grid operators cleaned up their interconnection queues

Starting in 2023, ISOs implemented sweeping reforms to eliminate speculative projects from their queues. New cluster processes, higher deposits, and compressed timelines resulted in hundreds of project cancellations. In MISO, 62% of 2025’s 73.4 GW in cancellations came from a single study cycle. SPP saw 27 GW cancelled in August 2025 alone following interconnection study results. While reforms improved queue quality, they accelerated project cancellations.

Cancelled projects will result in higher power prices and less economic growth

These project cancellations will affect nearly every American. The 266 GW of lost capacity threatens higher electricity prices for households and businesses as supply fails to keep pace with surging demand—a dynamic already visible in PJM’s double-digit annual price increases.

Beyond higher bills, the cancellations represent a massive economic loss. We estimate that the 1,891 projects cancelled this year would have generated $400 billion in investment—capital that would have flowed disproportionately to rural communities in need of economic revitalization.

The current trajectory is unsustainable: America is simultaneously approving unprecedented electricity demand while cancelling the generation needed to meet it, creating policy incoherence that threatens grid reliability, affordability, and the country’s competitiveness in the global AI race.

You can read the full report by clicking the button below:

Read the report

Here’s what’s in today’s newsletter:

• Developers have canceled 1,891 power projects this year with a combined capacity of 266 GW, with clean energy accounting for 93% of cancellations.

• Why Trump’s plan for AI dominance requires renewables

• 18,000 reasons that it is so hard to build semiconductor fabs and wind farms in America

• Trump raised 15 times as much from the 100 richest Americans in 2024 than he did during his first presidential campaign, in 2016.

• How OpenAI’s latest data center is skirting zero-carbon electricity requirements in New Mexico and how it could save $132 million per year with solar power

• AI is getting cheaper, very fast. The cost of Google’s flagship AI models have fallen by 97% in the last 18 months.

Developers have canceled 1,891 power projects this year

I spent most of this summer telling anyone who would listen to me that Trump’s energy policies posed huge risks to our country’s economy and power grid. I wrote more than 50 posts—some of which were cited on the floor of the US Senate and shared by Elon Musk. I talked to countless reporters across the country. And I dedicated most of my writing on this newsletter to the topic.

The gist of my argument was this: Trump’s policies would force many developers and investors to cancel their projects leaving the US with an electricity shortfall in a few years. Tariffs would raise prices too quickly and introduce uncertainty; illegally canceling funding and forcing wind developers to stop construction would create a chilling effect; and phasing out clean energy tax credits would kill thousands of projects.

Over the last few weeks, I’ve been analyzing data to check in on the state of many of these at-risk energy projects. I found that 1,891 power projects with a combined 266 GW of generation capacity have been canceled in 2025—equivalent to roughly one-quarter of America’s entire current electricity generation capacity.

Clean energy projects have been hit hardest, accounting for 93% of project cancellations in 2025. Utility-scale solar alone saw 86 GW canceled, while battery storage projects lost 79 GW and wind projects shed 54 GW.

The Trump administration and their chaotic policies were responsible for many of these project cancelations. But the analysis showed that there’s much more causing project cancelations and preventing the buildout of clean energy than federal policy. Tomorrow I’m planning to publish the full results of the analysis in a report for Cleanview. I’ll share it with Distilled readers here.

I shared an early preview of the report with about a dozen reporters. If you’re a journalist and want me to send you a preview before publication, you can respond to this email or get in touch here.

Trump’s plan for AI dominance requires renewables

On a related note, Bloomberg published a story this week showing that Trump’s plan for AI dominance is threatened by his own attacks on solar and wind power:

The Trump administration is moving to fast-track the construction of power-hungry data centers as a matter of national security. At the same time, it’s adding roadblocks for new solar and wind farms.

But the two policies could be at odds: Hindering renewable energy projects risks slowing the AI boom — and could exacerbate rising electricity prices, a slew of data suggests.

For this story, Bloomberg used Cleanview’s data to show the risks that Trump’s anti-renewable policies pose to the US power grid. As they write:

Renewable energy so far remains the fastest and cheapest option to add power to the grid. Nearly 80% of the planned power plant capacity in the pipeline is tied to renewable sources

18,000 reasons it’s so hard to build a chip factory in America

One of the Biden administration’s signature achievements was passing the CHIPS Act, which convinced companies like TSMC to build semiconductors in the US instead of countries in Asia like Taiwan. But these companies are learning why it's so difficult to build large infrastructure projects in America.

Per The New York Times:

In Taiwan, TSMC and its suppliers build facilities in dedicated industrial zones that generally require one permit from a central authority. In Arizona, they must negotiate municipal, county, state and federal regulations, requiring thousands of approvals.

TSMC has said that building fabs in the US takes twice as long as it does in Taiwan. Once the factory is complete, manufacturing chips is then 50% more expensive.

The company was required to gain permits from city and county authorities to comply with state and federal regulations. In many instances, regulations for its industry did not exist at the local level, so TSMC had to convene a team of experts to craft its own language and gain approvals.

“We ended up establishing 18,000 rules, which cost us $35 million,” Mr. Wei said.

Many permitting processes are designed with good intentions. They prevent companies from polluting the local air and water, killing endangered species, and harming people in surrounding communities. But many of them are Kafkaesque.

Linde, a company that pumps pristine air into TSMC fabs, erected a plant next door to the factories, including a $45 million pipeline. Linde needed 150 permits for the pipeline alone, Mr. Planck said. A permit was required to pile extra dirt left from grading the site. Each permit entailed as many as 15 inspections.

The same permitting labyrinths are slowing the energy transition. Earlier this year, I wrote about a wind project in California that has been under development since George Bush was president.

In 2004, Iberdola, Tule Wind’s developer, submitted an interconnection request and began the process of development. Because the project was set to be located on both state and federal land, it required both a NEPA and CEQA environmental impact report. Those took two years. After the state and federal government issued their reports, the county of San Diego had to issue its own report. Then, after that, the state’s public utility commission had to approve the power purchase agreement between the developer and the state’s utility. And then, finally, the State Lands Commission, which controlled a small strip needed for phase two of the project had to weigh in. By this point, it was 2016—12 years after the project was first proposed. But the real saga was only just getting started.

Each state review and decision opened up a new opportunity for litigation from any member of the public. In 2014, a local group began launching a series of lawsuits targeting the project. At first they claimed that the electromagnetic frequency (EMF) from the wind turbines would cause cancer. When that didn’t work, they argued that the turbines would kill eagles. Their lawsuits were ultimately unsuccessful at stopping the project, but they added years to the development timeline.

20 years later, the second phase of the project is still not complete. It’s expected to come online in December 2027.

Again, it’s good that large infrastructure projects are regulated. But the permitting process shouldn’t be this difficult—especially for critical projects that would make our world better like wind farms.

Political donations by billionaires

Per The Washington Post:

Overall, billionaires have rallied behind Trump’s Republican Party. More than 80 percent of the federal campaign spending by the 100 wealthiest Americans in 2024 went to Republicans, The Post found. Trump himself raised 15 times as much from the 100 richest Americans in 2024 than he did during his first presidential campaign, in 2016.

Why isn’t this data center using solar?

OpenAI and Oracle recently announced plans to build a massive data center in New Mexico called Project Jupiter. The project’s developers convinced the New Mexico legislature to create a loophole allowing them to avoid the state’s 2045 zero-carbon electricity mandate by building a “microgrid.” Rather than connect to the grid, they plan to build a 900 MW gas power plant.

But the project is located in one of the sunniest parts of the country and could immediately power 40% of its operations with nearby solar, according to an analysis by Ember’s Dave Jones and Kostantsa Rangelova. That would save $132 million per year and prevent 1.5 million tons of unnecessary emissions.

AI is getting cheaper—fast

One of the things I’ve learned writing about clean energy is the importance of learning curves. Forecasters failed to anticipate the clean energy revolution, in part, because they didn’t expect costs to fall so quickly.

Like solar, wind, and batteries, AI is on a learning curve. It’s easy to miss this given the insane amounts that companies are spending on AI infrastructure. But you can see it clearly in the data. The cost of Google’s flagship models have fallen by 97% in the last 18 months.

The same trend holds across virtually every AI lab and model. As Epoch AI writes:

The price to achieve GPT-4’s performance on a set of PhD-level science questions fell by 40x per year. The rate of decline varies dramatically depending on the performance milestone, ranging from 9x to 900x per year.

I worry that many AI skeptics are missing this fact and making the same mistake that renewable skeptics made a decade ago, underestimating the trajectory of the technology and its impacts on society.

More weekly briefings

I hope you’ve enjoyed reading the first six editions of this new format as much as I’ve enjoyed writing them. You can read the first five editions below. And if you aren’t already a paid subscriber, you can sign up here:

Subscribe now

Here’s what’s in today’s newsletter:

• Natural gas power is making a comeback in the US. In the last year, gas power capacity under construction has grown by 64%.

• Another forecast shows AI power demand accelerating. Nearly a quarter of newly proposed data centers are 500 MW or larger.

• Climate hawk and Canadian Prime Minister Mark Carney wants to build oil pipelines. A map of the country’s pipeline infrastructure shows why.

• Two new research papers suggest a novel way to build more housing in America: Give NIMBY’s and renters cash.

• Why Slow YouTube is my favorite new genre on the internet.

Natural gas power is slowly making a comeback in the US

As I wrote in my last newsletter, one of the best ways to see the future of America’s electric grid is to look at what kind of power projects are under construction.

So what does the construction pipeline tell us about the future of natural gas power? That after a decade of steady decline, gas is making a comeback. In the last year, gas capacity under construction has grown 64% to 8,691 MW. Gas capacity across all stages of development we track at Cleanview has more than doubled in the last year.

I wrote up a more detailed analysis on construction pipelines—including how much solar, wind, and battery capacity the US is currently building—for Cleanview. You can check it out here.

Another day, another big AI power demand forecast

Read more

Here’s what’s in today’s newsletter:

• Clean energy and grid-scale batteries account for 88% of all under-construction power projects in the US.

• America has been building less transmission lines each year. But that could be changing based on some big recent announcements.

• Wall Street’s AI darlings got hammered last week. Some stocks fell by as much as 17%, despite Nvidia beating expectations last quarter.

• Many New Deal programs took months to implement. The Inflation Reduction Act took years. A group of former Biden staffers recently published a report explaining why.

• Progress against extreme poverty is slowing. In the next decade there will likely be more people living in extreme poverty than there are today—a sharp reversal.

• Iowa City made its buses free. Now ridership has surpassed pre-pandemic levels by 18 percent.

• We’ve already reached peak gas-combustion vehicle. Since 2019 internal combustion engine car sales have fallen by 31% globally.

• Utilities are forecasting 260% more data center power demand than they were just a year ago, according to a new report.

• Will oil and gas consumption rise over the next few decades? If you read any mainstream news stories on the topic over the last month, you may have been misled.

America is still building a lot of clean energy

One of the best ways to see the future of America’s electric grid is to look at what kind of power projects are under construction. Here’s a chart I made last week to peer into that future:

There are currently 668 power projects under construction in the US with a combined capacity of 72.5 GW, according to our project tracker at Cleanview.

Solar leads the construction pipeline with 39% of capacity, followed by wind (26%), batteries (23%) and natural gas (12%). That means clean energy and storage account for 88% of all under-construction projects.

Clean energy in America faces a lot of challenges right now. But everyday—despite all the headwinds—we’re building more clean energy projects.

I wrote more about this in Cleanview’s newsletter recently.

America is building transmission lines

There’s a saying that I like that goes “There’s no transition without transmission.” Every rapid energy transition scenario requires building a lot of high-voltage transmission lines to bring electricity from the windy plains and sunny deserts to populations centers.

The problem is that America isn’t really building any transmission infrastructure.

Between 2012 and 2016, an average of 2,000 miles of high-voltage transmission lines were built in the United States. Between 2017 and 2021, that number fell to 700 miles. Last year, just 322 miles of high-voltage lines were built.

That might be changing soon though.

Read more