Anthropic CEO Cautions: Measured AI Spending to Avoid Bankruptcy if Revenue Timings Miss

Anthropic, developer of the Claude chatbot, is scaling more conservatively than hyperscalers amid significant uncertainty in AI revenue timing. Co-founder and CEO Dario Amodei warned that even a one-year miscalculation in revenue growth could lead to bankruptcy if the company commits trillions to data centers prematurely. “We’re buying compute at a level comparable to the largest players, but not at their YOLO pace,” Amodei said in an interview. “If the country of geniuses arrives in 2028 instead of 2027, and our revenue is still only $800B, the capital burden could be ruinous.” In November 2025, Anthropic announced $50B in AI infrastructure spending, starting with data centers in Texas and New York. The company projects 10x annual revenue growth, with 2026 revenue around $10B and potential $1T by 2027, which would justify massive compute spending. Amodei contrasts this with Amazon’s $200B, Alphabet’s $185B, and Meta’s $135B in 2026, emphasizing Anthropic’s focus on enterprise customers and realistic production constraints. The company is proceeding cautiously while still investing heavily in compute capacity.