aiqbench analysis
The filing shows how far frontier AI labs are committing to external compute, with multi-year contracts running into the 2030s that are largely non-cancellable even if usage drops. For AI readers, that makes these commitments a useful signal of demand expectations and of how dependent model development is on a few hyperscale cloud providers. The main open questions are how the spending will be financed and what happens to these obligations if revenue growth slows.
Anthropic Discloses Massive Computing Power Purchase Commitments, Plans to Buy at Least $114.5 Billion in Services from Google
In a prospectus filed on October 26, Anthropic disclosed large computing power and infrastructure agreements with multiple technology companies, with spending commitments extending through 2036. The company has committed to spend at least about $114.5 billion with Google from April 2026 to July 2033; at least $110 billion with Amazon from May 2026 to April 2036; and at least $44.9 billion with Microsoft from November 2026 to May 2033. In principle, these commitments will not be canceled due to reduced actual usage, except in certain cases such as specific defaults. (Jiemian)
Read the full article at 36KrRelated coverage
Investing in the public benefit
Anthropic · Oct 18, 2026
Safety is marketing for Anthropic, not sacrifice
LessWrong · Oct 28, 2026
Anthropic commits $10 million to Canadian AI research
Anthropic · Jul 14, 2026
Anthropic’s S-1 makes clear it’s on the hook for billions in compute. Its customers can walk at any time
Fortune · Oct 26, 2026