Is the AI capex a bubble? The spend-vs-revenue gap

The spend is real — and record-breaking

Of all Aschenbrenner's quantitative bets, the capex call is the most clearly vindicated. Capital expenditure on AI infrastructure — mega-clusters, chips, energy — has run ahead of even his aggressive trajectory, graded Exceeded. The actors with the most information are spending as if the timeline is short.

The gap that fuels the "bubble" worry

Here's the honest caveat the scorecard flags: revenue is lagging the spend. By spring-2026 reporting, the most generous third-party AI-revenue figure was around $60B, with leading labs annualizing well below that (OpenAI ~$25B+, Anthropic ~$19–20B). Spending like AGI is near doesn't prove it is — and that spend-vs-revenue gap is exactly what a bubble looks like early.

So is it a bubble?

The mature read: the capability trends underneath the spend are real, so this isn't a pure speculative mania. But the monetization gap is genuine, and if it doesn't close, the capex could look overbuilt in hindsight. It's not "bubble vs. no bubble" — it's a race between capability compounding and revenue catching up.

Frequently asked questions

Not obviously, but the risk is real. Capex has exceeded trillion-dollar projections and underlying capability is on track — not pure mania. However, revenue is lagging the spend, which is the classic early-bubble signal to watch.

Not yet fully. As of spring 2026 the most generous AI-revenue figure was ~$60B, with leading labs annualizing below that — behind the run-rate the record capex implies.

Revenue catching up to the spend would confirm it's justified; a stall in capability gains or revenue while capex keeps climbing would suggest overbuild.