He took it as a challenge, asking entrepreneurs to come up with AI products and services to make use of, and generate revenue from, all that infrastructure. Fast-forward to today, adding up three years of hyperscaling, and Cahn\u2019s got a new number<\/a> on AI infrastructure spending for 2026: $1.5 trillion.<\/p>\n All told, he calculates that the AI industry will have to earn $3 trillion to justify all those chips and other data center expenditures. And that\u2019s probably an underestimate \u2014 the rising costs of memory and the increasing use of exotic or inference-specific chips will drive that number up. \u201cRecently,\u201d he writes, \u201cthe required revenue per GW of CapEx has sharply increased due to these bottleneck dynamics and rising costs of construction.\u201d<\/p>\n On the other side of the ledger, Anthropic is thought to have hit $60 billion in ARR<\/a>, while OpenAI reportedly earned $13 billion<\/a> in 2025 (although in November 2025, it said it was at $20 billion ARR<\/a>) and is presumably making more this year. But there\u2019s clearly a large gap to be closed.<\/p>\n Someone minding that gap is Torsten Slok, the chief economist at Apollo, the giant asset manager. In a recent note<\/a>, he points out that the hyperscalers \u2014 Google, Meta, Microsoft, and Amazon \u2014 are all predicting massive accelerations in their free-cash flow in 2028. That is, they expect to see the payback from all those chips they bought.<\/p>\n What if they don\u2019t? Slok notes a risk we\u2019re currently seeing across AI usage: More organizations turning to cheaper open weight models, often Chinese, not those built by the frontier labs, and overall token prices falling. OpenAI\u2019s latest model, per CEO Sam Altman, is 54% more token efficient<\/a> on coding tasks. That\u2019s good for users fretting about the cost of their AI agents, but it may be bad for companies building token factories should users not wildly increase their overall token usage with them.<\/p>\n Slok worries that if hyperscalers don\u2019t meet their cash-flow goals, the market reaction could be severe \u2014 Just something to keep in mind as you\u2019re herding your AI agents toward cheaper tokens.<\/p>\n![\"\"](\"https:\/\/www.europesays.com\/ai\/wp-content\/uploads\/2026\/07\/Jul09_Chart1_v2.jpg\")
Image Credits:Torsten Slok\/Apollo<\/p>\n![\"\"](\"https:\/\/www.europesays.com\/ai\/wp-content\/uploads\/2026\/07\/Jul09_Chart3_v2.jpg\")
Image Credits:Torsten Slok\/Apollo<\/p>\n
\u201cwith so much riding on so few names,\u201d he writes, \u201ca slower payoff wouldn\u2019t just be a sector problem, it would risk tipping the economy into recession and the S&P 500 into a correction.\u201d<\/p>\n