Will there be superintelligence in the 2030s?

What he predicted

Superintelligence is the end-state of Situational Awareness's argument. The sequence is explicit: AGI around 2027 → an intelligence explosion (2027–29) as AI automates AI research → superintelligence in the early 2030s, with capabilities far beyond any human. It is the most consequential claim in the essay and, by construction, the last to resolve.

Why it's Pending, not Open or Wrong

This prediction sits at the end of a dependency chain. Its precursors haven't fired: a system that can autonomously do AI research (the AGI-2027 bar) is undemonstrated, and the intelligence explosion that would have to precede superintelligence is itself pending. With neither the trigger nor the timeline in play, there is no evidence to grade it for or against — the definition of Pending on this scorecard.

What to watch

The leading indicators are upstream. If AGI-2027 resolves and a credible intelligence explosion begins, the superintelligence clock effectively starts and this verdict moves to Open. Until then it remains the most speculative and least testable claim on the board — and the one most dependent on every earlier prediction holding.

Frequently asked questions

Too early to tell. Aschenbrenner predicted superintelligence in the early 2030s, following AGI and an intelligence explosion. As of mid-2026 neither precursor has occurred, so the prediction is graded Pending.

AI with capabilities vastly beyond the best humans across essentially all domains. In Situational Awareness it arrives in the early 2030s, produced by an intelligence explosion in which AI rapidly improves AI.

Because it depends on earlier steps — AGI by 2027 and an intelligence explosion — that haven't happened. Without those, there's no evidence to grade it either way.