Anthropic disclosed $895.4 billion in aggregate compute and infrastructure purchase commitments in an amended Form S-1 filing Monday, escalating the stakes for its planned IPO. The company generated $33.24 billion in revenue during the nine months ended Sept. 30—a 1,200 percent increase year-over-year—but posted a $104 billion net loss, mostly from fair-value adjustments to convertible notes and derivative liabilities.
Third-quarter revenue hit $17 billion, up 1,127 percent annually but only 47 percent sequentially. That marks a sharp deceleration: Q2 grew 144 percent, Q1 grew 132 percent. Compute and infrastructure expenses jumped $14.4 billion, a 329 percent increase over the same nine-month period last year.
The math is unforgiving. Anthropic is locking in $895B in infrastructure spend while its sequential growth rate collapsed from triple-digit percentages to mid-40s. For a five-year-old company preparing a road show, the core risk is binary: either demand justifies this scale—or the capital becomes stranded.
The timing adds pressure. Microsoft, Google, and Amazon reported aggressive AI spending last week, signaling that hyperscaler cash flows are already being consumed by compute buildout. Wall Street analysts worry rising interest rates will make debt financing prohibitive for data center operators and emerging cloud providers, potentially crimping Anthropic's ability to source capacity at expected terms.
Anthropics procures compute primarily from Amazon, Google, and SpaceX, plus neoclouds. Its dependency on third-party infrastructure—rather than owning its own—shifts risk but doesn't eliminate it. If hyperscalers deprioritize Anthropic allocations during a buildout crunch, the company's committed roadmap becomes notional.
OpenAI, Anthropic's chief rival, confidentially filed its prospectus in June with a target 2027 debut. Both companies frame the IPO as validation of the AI model's economic viability. For Anthropic, that pitch hinges on proving that $895B in committed spend drives unit economics strong enough to justify the risk.
