Anthropic Inks Massive $3.4 Billion Deal for Data Center Power Before Blockbuster IPO
Every artificial intelligence (AI) boom story hits a similar wall at scale. AI labs can order chips and build data centers, but they also need electricity to run them, and the equipment that generates power is in short supply. In response to this problem, Anthropic (ANT) recently did something different. Instead of waiting for someone else to supply the power, the company went shopping for the equipment itself.
AI developers typically lean on utilities and specialist power companies for electricity, per a Financial Times report. But that business model which is under strain, given that demand for gas turbines and generators is soaring. Manufacturing lead times now stretch several years. So, AI companies are using capital to cut out intermediaries and take direct control of production.
Inside Anthropic's $3.4 billion Generator Deal
Anthropic has reportedly bought commitments for 1.7 gigawatts of heavy-duty generators for $3.4 billion. OnePWR Solutions, an equipment supplier based in San Antonio, Texas, sold the commitments. According to Financial Times, the deal is the largest known single purchase of power generation equipment by an AI lab.
Anthropic did not acquire OnePWR. However, the deal included the engineering and construction capabilities needed to install the generators.
OnePWR ordered Caterpillar (CAT) generators two years ago and required financing to complete the transaction. Anthropic reportedly first planned to be a OnePWR customer, but changed course and took over the orders instead.
Several AI giants, including SpaceX (SPCX) and Alphabet (GOOGL), have invested heavily in turbines and energy-related acquisitions. Owning power infrastructure helps companies "better protect themselves against disruptions beyond their control," said Arash Nazhad, Managing Director and Head of Clean Energy and Power at Stifel.
In its amended S-1 filing, Anthropic revealed that it has nearly $900 billion in computing commitments through 2045. The company has leaned heavily on partners for power, chips, and more, while paying a premium to lock down capacity and meet demand. Buying generators directly gives Anthropic more control over the electricity it needs.
What Anthropic Has Told Investors
Anthropic’s prospectus opens with a letter from its founders, which reads like a pitch with a conscience. According to the founders, they want investors to grasp both AI’s potential and risks, as well as Anthropic’s “unique approach to building a company and operating in the world" and the ways in which its actions "may differ from that of a purely profit-maximizing company.”
Anthropic's story started in 2020, when four of the founders coauthored a paper on the scaling laws of AI. The thesis was straightforward enough: Give an AI model a larger size, more computing power, more data, and more training, and it gets much smarter.
“The phenomenon is somewhat analogous to Moore’s law for computer chips, under which the computing power of these chips doubled every 12 to 18 months for several decades,” the letter reads. But Anthropic's founders argue the payoff from machine cognition will be "vastly greater." The thesis is that the curve will keep climbing until it reaches “a country of geniuses in a data center.”
In the letter, the founders also note that Anthropic will often “make forward investments (especially in compute) that may seem irresponsibly large at first glance.” They liken this approach to that of drug makers, which spend heavily up front to earn returns later.
The Risks Anthropic Admits
The founder letter points to areas where Anthropic sees significant upside from AI. That includes advances in medicine, economic growth both domestically and globally, an increase in global reach, and improvement in government services.
However, Anthropic's leadership also outlines four categories of risk in the letter. These include the misuse of AI technology, loss of control, authoritarian abuse, and economic disruption. Speaking on that last risk factor, the letter outlines Anthropic's $200 million commitment to “partner with governments to run studies on different ways of responding to changes in the economy.”
The letter’s boldest passage describes a time that Anthropic held back. According to the founders, Anthropic judged that its Claude Mythos Preview model carried significant cybersecurity risk. In response, Anthropic delayed the launch, then offered the capabilities as Claude Fable 5 after adding safeguards. “The delay meant forgoing substantial revenue and ceding ground to competitors," the letter reads.
Still, the founders note that each Anthropic employee supported more than $6.6 million in trailing 12-month revenue as of Sept. 30, 2026. Meanwhile, all seven of the AI lab's founders have pledged to donate 80% of their personal equity in the company to charity.
The Key Takeaway
Looking forward, Anthropic faces both opportunity and obstacles. For example, Blackstone (BX) recently warned investors that Anthropic could receive a junk credit rating instead of an investment-grade one, given rising debt levels to support its capital expenditures.
The latest $3.4 billion generator deal shows how far the AI race has moved. It is no longer just about models and chips, but about who controls the power that keeps them running.
On the date of publication, Aditya Raghunath did not directly have a position in any of the securities mentioned in this article. All information and data in this article is solely for informational purposes. For more information please view the Barchart Disclosure Policy here.