Anthropic’s $200 Billion Google Commitment

Anthropic has committed to spend $200 billion with Google Cloud over five years, according to a report cited by Reuters, in a deal that reportedly covers cloud services and chip capacity (Reuters, 2026). The same report said the agreement includes multiple gigawatts of tensor processing unit capacity expected to come online from 2027.

Access to compute is no longer a background requirement that companies sort out after building a model; it is becoming the centre of strategy, capital allocation and competitive advantage. The reported commitment would amount to more than 40 percent of the cloud revenue backlog Google recently disclosed to investors, which shows how directly one major AI customer can now shape the business outlook of a cloud provider.

The April agreement reportedly involved Google and Broadcom, and the planned TPU capacity is expected to begin coming online from 2027. That timing suggests that the biggest AI firms are already reserving future infrastructure well ahead of demand peaks, rather than waiting to see how the market settles.

Anthropic declined to comment to Reuters, while Google directed questions back to Anthropic. Even with that caution, the reported size of the deal points to the shape of the next phase of competition. Companies with the deepest model talent will still matter, but so will companies with the balance sheet, supplier access and long horizon needed to secure compute before everyone else does.