Anthropic inks $3.4 billion deal for gas-powered data centers—the largest of its kind—to secure AI autonomy ahead of planned IPO
Ahead of its estimated $1.3 trillion IPO, Anthropic secured power autonomy in the largest power generation equipment acquisition from an AI company.
In late September, investment bank Piper Sandler announced that it had advised OnePWR Solutions, a San Antonio-based equipment supplier, on the sale of its power generation platform to an "AI frontier lab" for $3.4 billion. The recipient of this deal was Anthropic, a prominent artificial intelligence company. According to the Financial Times, this marks the largest direct investment in power generation equipment by an AI company.
The $3.4 billion agreement involves 1.7 gigawatts of powerful natural gas generators from OnePWR Solutions, taking over commitments previously made with Caterpillar. Anthropic also assumes responsibility for the engineering and construction of these installations. The generators are anticipated to be available within the next two years.
Anthropic is currently preparing for its initial public offering (IPO), having filed its S-1 registration statement with the SEC in September. The filing reveals that Anthropic has committed approximately $900.9 billion to computing and infrastructure, with about $451.6 billion projected to be spent before 2030. The company is aiming for a $1.3 trillion valuation, offering 825 million shares at a price range of $70 to $78.
The S-1 filing highlighted concerns related to Anthropic's dependency on natural gas for power, citing potential volatility in prices due to limited supply capacity. The company acknowledged that procuring electricity and power equipment can take years due to high prices, tariffs, and limited production capacity. Furthermore, Anthropic noted that some data centers may rely on dedicated natural gas-fired generation, exposing them to risks such as fuel price fluctuations and pipeline capacity constraints.
Written by urgent.news from Fortune's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.