As demand for computing power from large AI models explodes exponentially, power supply has become a core pain point constraining the industry's development. On November 8, according to media reports, leading AI company Anthropic reached a major deal worth as much as $3.4 billion in September this year with Texas-based equipment supplier OnePWR Solutions, fully securing power supply for future data centers through direct procurement and taking over orders.

Massive Investment Locks In Major Order: 1.7 GW of Heavy-Duty Generator Sets to Be Deployed

According to the agreement between the two parties, the total installed capacity Anthropic is purchasing is as high as 1.7 gigawatts, covering a large number of heavy-duty generator sets. At the same time, the deal includes not only the hardware equipment itself but also comprehensively covers the engineering and construction capabilities required to install the generators.

It should be noted that this transaction is purely a commercial procurement and supply chain cooperation. It does not involve acquiring OnePWR itself, nor does it involve any employees transferring to Anthropic. Industry analysts pointed out that this is the largest single purchase of power generation equipment by an AI company known to date, highlighting the extreme urgency among current leading large-model developers in their energy planning.

From Mere Customer to Directly Taking Over Orders: Locking In Supply Chain Assurance in Advance

As the specific deal progressed, Anthropic initially planned to purchase only as a regular customer of OnePWR.

However, in the face of an increasingly tight global power equipment supply chain and rapidly growing demand for data center expansion, Anthropic ultimately decided to directly take over the company's previous order for power generation equipment placed with well-known construction machinery giant Caterpillar. Through this more direct and decisive intervention, Anthropic successfully bypassed the usual lengthy lead times and completely locked in critical power supply resources in advance, laying a solid energy foundation for its continuously expanding AI infrastructure in the future.