Did AI capex hit trillion-dollar scale, as predicted?

The prediction

Situational Awareness argued that the race to AGI would drive an unprecedented capital build-out — trillion-dollar compute clusters and an accelerating capex curve as labs and nations competed for compute.

What happened

Investment has run ahead of his projections. Of all the essay's quantitative bets, this is the one graded Exceeded — the build-out of AI infrastructure has matched the aggressive trajectory and then some. Compute concentration, mega-clusters, and the surrounding energy and capital commitments have all tracked or outpaced the curve he drew.

The honest caveat: revenue

We deliberately do not count AI revenue as a vindicated prediction. By spring-2026 reporting, revenue is tracking meaningfully behind the run-rate the essay sketched — the most generous third-party figure was around $60B, with leading labs annualizing well below that. Capex has exceeded; monetization has not kept pace. Flagging this guards against grading the scorecard in Aschenbrenner's favor.

Why it matters

The capex curve is the clearest real-world confirmation that the actors with the most information are behaving as if the timeline is short. But the revenue gap is the live question underneath it: spending like AGI is near doesn't prove it is.

Frequently asked questions

Yes — this is his most clearly vindicated call, graded Exceeded. Capital expenditure on AI infrastructure has run ahead of even his aggressive trillion-dollar-scale projections.

The scorecard doesn't rule either way, but it flags the gap: capex has exceeded Aschenbrenner's projections while revenue has lagged — the most generous third-party 2026 figure was ~$60B, below the run-rate the essay implied.

Because it isn't tracking to the essay's sketch. As of spring 2026, AI revenue is running meaningfully behind, so counting it would inflate the scorecard in Aschenbrenner's favor.