Can AI justify $3 trillion in infrastructure spend?
Original: Can AI answer the $3 trillion question?
Why This Matters
A $3 trillion revenue gap tests whether AI infrastructure investment can deliver returns before market confidence erodes.
Sequoia's David Cahn estimates the AI industry must generate $3 trillion in revenue to justify 2026's $1.5 trillion infrastructure spend. Meanwhile, Apollo economist Torsten Slok warns that if hyperscalers miss 2028 cash-flow targets, the fallout could trigger a broader economic recession.
Sequoia partner David Cahn, who in 2023 calculated that $200 billion in revenue was needed to justify AI infrastructure spending tied to Nvidia's $50 billion GPU revenue, has updated his figures for 2026. He now estimates total AI infrastructure spending at $1.5 trillion, requiring the industry to generate $3 trillion in revenue to break even — a figure he calls likely an underestimate, citing rising memory costs and growing use of inference-specific chips. On the revenue side, Anthropic is reported to have reached $60 billion ARR, while OpenAI earned $13 billion in 2025, claiming $20 billion ARR as of November 2025. Apollo chief economist Torsten Slok notes that Google, Meta, Microsoft, and Amazon are all projecting sharp free-cash-flow acceleration by 2028, implying they expect a return on their chip investments. However, Slok flags key risks: enterprises are shifting to cheaper open-weight models, often from Chinese developers, and token prices are falling. OpenAI CEO Sam Altman stated its latest model is 54% more token-efficient on coding tasks — a benefit for users but a potential headwind for revenue. Slok warns that if hyperscalers miss cash-flow targets, the market reaction could be severe enough to tip the economy into recession and push the S&P 500 into a correction.