Financial Times has a provocative question about Anthropic: how much revenue would the company need to justify its enormous AI computing commitments? The headline number is roughly $600 BILLION a year. Not because Anthropic has announced that target. And not because it has to pay $1 trillion tomorrow. It's a hypothetical calculation about the scale of returns investors might eventually expect. And the math is worth looking at. THE $1 TRILLION COMPUTE BET According to Anthropic's IPO prospectus, the company has approximately $900.9 billion in compute and infrastructure purchase commitments and leases. • About $70.7 billion is due within 12 months of September 30, 2026. • About $451.6 billion is expected before the end of 2030. • Most of the remainder extends from 2031 to 2045. Broader arrangements without fixed dollar amounts could make the total exposure higher. These are FUTURE commitments across many years, not $900 billion already spent or a single loan coming due. SO WHERE DOES $600 BILLION COME FROM? Here is the Financial Times-style thought experiment: 1. Round the computing commitments to $1 trillion. 2. Assume investors want a 20% annual after-tax return on that amount: $200 billion in annual net profit. 3. Assume a 17% tax rate. That requires about $241 billion in pretax profit. 4. Assume a 40% pretax operating margin. Divide $241 billion by 0.40. THE RESULT: APPROXIMATELY $602 BILLION IN ANNUAL REVENUE. That's an illustrative scenario, not an official financial target, a cash-flow forecast, or a rigorous calculation of return on invested capital. Future purchase commitments aren't the same as capital invested on day one. HOW BIG IS $602 BILLION? For perspective: • Anthropic's October 2026 revenue run rate: approximately $81 billion. • Alphabet's trailing 12-month revenue through September 2026: approximately $471 billion. • Apple's fiscal 2026 revenue: approximately $479 billion. A $602 billion annual business would be around 7.4 times Anthropic's October run rate, and considerably larger by revenue than either Alphabet or Apple today. And remember: a run rate extrapolates recent sales activity. It is not the same as revenue actually earned over a full year. WHAT ABOUT MARGINS? Anthropic reported a 15% ADJUSTED operating margin in Q3 2026. That's not a 15% net profit margin. Its GAAP operating margin in that quarter was around negative 3%, which is why comparing current profitability with a hypothetical 40% operating-margin scenario requires care. The company is scaling fast. But its compute costs are enormous, and long-term contracts can become risky if customer demand doesn't keep pace. IS $600 BILLION REALLY NECESSARY? No. The number depends entirely on the assumptions. Change the required return, taxes, operating margin, investment timing, or financing structure, and the result changes. It also ignores important realities: contracts are spread over almost two decades, capacity may generate revenue over time, and compute efficiency and prices will change. But the thought experiment makes one point very clearly. The AI infrastructure race isn't just about building smarter models. It's about building businesses large and profitable enough to support extraordinary amounts of compute. AND HERE'S THE PERFECT ENDING. While everyone debates the economics of a trillion-dollar infrastructure commitment, Dario and Daniela Amodei were ringing the Nasdaq opening bell for Anthropic's IPO on November 19, 2026. This is not merely an ambitious business plan. It's one of the biggest financial experiments in the history of technology. SOURCES Financial Times: https://www.ft.com/content/ea8434c7-59d9-42cf-afd4-97cf42544c59 Anthropic IPO prospectus: https://www.sec.gov/Archives/edgar/data/2130357/000162828026074944/anthropicpbc-424b4.htm Anthropic IPO factsheet: https://ipo.anthropic.com/documents/us/factsheet.pdf Nasdaq IPO announcement: https://www.nasdaq.com/newsroom/anthropic-marks-ipo-at-nasdaq Can AI demand grow fast enough to justify these commitments, or are we confusing unprecedented model adoption with guaranteed long-term profits? #Anthropic #AIInfrastructure #AIEconomics