As the U.S. changed presidential administrations this week, AI infrastructure dominated the old and the new. 

One of Joe Biden’s final acts before leaving office was signing an executive order focused on making federal sites available for private sector companies to build data centers and clean energy facilities, as well as speeding up the time it takes to obtain permits and connect new facilities to the grid. Days later, one of Donald Trump’s first acts was announcing “The Stargate Project,” billed as a four-year, $500 billion investment in AI infrastructure from OpenAI, SoftBank, Oracle, and MGX that will “secure American leadership in AI, create hundreds of thousands of American jobs, and generate massive economic benefit for the entire world.” 

You’d be right to have some questions about Stargate. But one thing is clear: AI infrastructure is taking center stage across the tech industry, power sector, and political landscape. 

AI infrastructure gets political 

When you get into the details, the Stargate announcement starts to look like a bunch of previously announced deals and loose plans being cobbled together into a “new” initiative to give Trump an early political win, as tech investor and commentator M.G. Siegler breaks down well in his Spyglass newsletter, calling it a ”catch-all of convenience for many of the parties involved.” 

Where exactly all those billions will come from is not yet clear (even Elon Musk sniped the announcement, posting, “They don’t actually have the money”). The Texas site mentioned in the announcement seemingly refers to a data center that Oracle has already been building for quite some time. And the timing with Trump’s return to office is awfully convenient, as is the venture’s four-year timeline. 

But Stargate does reveal how AI infrastructure is emerging as a top priority, funding magnet, and political target in America. 

Data center and energy projects sweep the U.S.

Regardless of the political positioning of Stargate, big AI-related infrastructure projects are already well underway across the United States, with the cloud hyperscalers and savvy real estate developers locking up locations with excess power capacity. Meanwhile, utility companies are scrambling to find power to supply the new data centers.

AWS this month announced plans to invest at least $11 billion into cloud and AI infrastructure in Georgia, which has become a hot spot for big tech’s infrastructure with Microsoft, Google, Meta, and X all building in the state. The company previously said it’s spending a similar $11 billion in Indiana, and in November, announced it’s pursuing three nuclear projects in Virginia and Washington state.

Microsoft similarly kicked off the new year by announcing plans to invest about $80 billion into constructing data centers for AI in the fiscal year 2025 alone, with over half of it taking place in the U.S. Meta recently announced plans to build its largest AI data center yet—a $10 billion facility—in Louisiana, and OpenAI executives said the company plans to build its own data centers in the U.S. Midwest and Southwest. 

Yesterday, the Wall Street Journal reported that South Carolina and its state-owned power provider ​​Santee Cooper are working to reboot a giant nuclear project to meet the growing energy demands posed by AI—yet another showing of how AI is driving a resurgence of interest in nuclear power. The state is seeking proposals and betting that the project will be of high interest to the many tech companies scrambling to build data centers and infrastructure to power them. A similar story is playing out in Wisconsin, where a developer is working to turn a 1,000 acre farm into a data center with hopes that a tech company eager to power its AI ambitions will scoop it up. The site is less than an hour from a Microsoft data center that’s currently under construction. 

The implications of this AI infrastructure build out stretch beyond simply more capable AI models for business and consumers. States are competing to attract AI infrastructure, hoping it will boost local economies. AI may push the U.S. to modernize its electrical grid, but it poses environmental concerns, such as the massive amounts of water data centers require for cooling (not to mention issues around nuclear waste disposal). There are regulatory challenges and national security concerns, especially as AI comes to be seen as a critical national technology and point of geopolitical tension. Infrastructure sometimes gets a reputation as being boring, but this is anything but. 

And with that, here’s more AI news. 

Sage Lazzaro
sage.lazzaro@consultant.fortune.com
sagelazzaro.com

AI IN THE NEWS

Microsoft is no longer OpenAI’s exclusive cloud provider. A new deal gives Microsoft the “right of first refusal” but allows OpenAI to buy cloud services from rival providers. The change was disclosed as part of the Stargate announcement. OpenAI has previously blamed product delays on a lack of compute available from Microsoft. You can read more from CNBC.

Google invests another $1 billion in Anthropic. That brings Google’s total investment in the company to $3 billion, while Amazon has also invested $8 billion. The growing partnerships—as well as Microsoft’s with OpenAI—had been probed by the FTC under Lina Kahn. Now that she’s stepping down in the coming weeks, tech giants looking to strike big AI deals will likely face less scrutiny. You can read more in the Financial Times. 

Databricks closes its $10 billion round. Meta joined as a “strategic investor,” with Temasek and QIA (Qatar’s sovereign wealth fund) also joining the round. The deal values the data analytics company—which has gained prominence in the AI boom as it offers companies a way to combine and standardize data they need to leverage AI—at $62 billion. You can read more in TechCrunch. 

FORTUNE ON AI

The AI trend that could send software stocks soaring in 2025 —by Greg McKenna

Dell’s employees had 800 different gen AI ideas. Here’s how the PC maker winnowed the list to —by John Kell

OpenAI’s Stargate may be tech’s biggest gamble ever, but here’s what’s really at stake —by Sharon Goldman

AI CALENDAR

Feb. 10-11: AI Action Summit, Paris, France

March 3-6: MWC, Barcelona

March 7-15: SXSW, Austin

March 10-13: Human [X] conference, Las Vegas

March 17-20: Nvidia GTC, San Jose

April 9-11: Google Cloud Next, Las Vegas

EYE ON AI NUMBERS

$1.75 million

That’s how much OpenAI spent on political lobbying in 2024—a nearly sevenfold increase from the $260,000 it spent in 2023, according to a filing reported on by MIT Technology.

I wrote in this newsletter a few weeks ago that the company recently tripled the size of its lobbying team to 12 members, scooping up D.C. insiders from across the political spectrum. Now that we have a benchmark number for its spend, it will be interesting to see how that compares to spend in 2025 under a new administration that has already proved to be friendly to the AI industry. In addition to the Stargate announcement, one of Trump’s other first acts was reversing former President Joe Biden’s executive order that would require companies to submit safety test results to the government before releasing new models.