A stark new calculation from Silicon Valley has reignited the debate over whether the artificial intelligence boom can ever pay for itself. David Cahn, a partner at Sequoia Capital, has updated his famous 2023 analysis and now estimates that AI infrastructure spending in 2026 alone will reach $1.5 trillion — meaning the industry must generate $3 trillion in revenue to justify the investment. That figure, he warns, is likely an underestimate given rising memory costs and the shift to specialised chips.
The numbers on the revenue side remain modest by comparison. Anthropic is thought to have hit $60 billion in annual recurring revenue, while OpenAI reported $13 billion in 2025 and is believed to be earning more this year. Yet the gap between spending and income remains vast. Torsten Slok, chief economist at Apollo Global Management, notes that the hyperscalers — Google, Meta, Microsoft, and Amazon — are banking on a major cash-flow acceleration by 2028. If that doesn't materialise, he warns, "a slower payoff wouldn't just be a sector problem, it would risk tipping the economy into recession."
For UK businesses, the implications are immediate. Many are already turning to cheaper open-weight AI models, often from Chinese developers, rather than expensive frontier systems. OpenAI's latest model is 54% more token-efficient on coding tasks, which cuts costs for users but squeezes revenue for companies that build and sell AI tokens. The UK's Information Commissioner's Office (ICO) is watching closely, as is the EU with its AI Act, which imposes transparency and risk-management rules that could affect how British firms deploy these technologies.
The risk of a market correction is real. Slok points out that so much of the S&P 500's recent gains are tied to a handful of AI-focused tech stocks. If the promised returns fail to appear, a sell-off could spill over into UK pension funds and portfolios that hold US equities. The Bank of England has already flagged AI as a potential source of financial instability in its latest Financial Stability Report.
For now, the debate is not whether AI will transform industries — it clearly is — but whether the scale of investment is rational. Cahn's original 2023 challenge to entrepreneurs was to build products that could generate real revenue from all that infrastructure. Three years on, the question remains unanswered, and the stakes have never been higher.