Dario Amodei's AGI prediction: “powerful AI” by 2026–27?

What Amodei actually forecasts

In his October 2024 essay Machines of Loving Grace, Amodei sketched what he calls “powerful AI”: systems smarter than Nobel-level experts across most fields, able to operate autonomously over long horizons — millions of them, running faster than humans. His much-quoted summary: “a country of geniuses in a datacenter.” On timing, he has said it could arrive as early as 2026, and has since referenced the 2026–27 window — always with explicit caveats that the timing could slip and the term “AGI” obscures more than it reveals.

Why his framing is distinctive

Amodei's contribution to the timeline debate is definitional discipline: instead of arguing about “AGI,” he specifies capabilities (expert-level breadth, long-horizon autonomy, massive parallelism). That maps almost exactly onto the strictest row of the AGI definition table — and onto the bar this scorecard uses for the 2027 verdict. In effect, Amodei and Aschenbrenner describe similar destinations on similar clocks, from different institutional seats.

How it compares

What the evidence says so far

Amodei's window is now live: mid-2026 sits inside his “as early as” range. The scorecard's read applies unchanged — capability near the top of the skilled-human range (~83% GDPval, ~80% SWE-Bench Pro), but nothing resembling an autonomous country of geniuses. If the 2026–27 window closes without long-horizon autonomous systems, his forecast joins the missed-aggressive-calls column; if it doesn't, he called it earlier than almost any lab leader.

Frequently asked questions

Amodei avoids the term AGI but has said “powerful AI” — systems beyond Nobel-level experts with long-horizon autonomy, “a country of geniuses in a datacenter” — could arrive as early as 2026–27, with explicit uncertainty.

Amodei's shorthand from Machines of Loving Grace (2024): millions of AI systems, each smarter than top human experts across fields, operating autonomously and faster than humans — his definition of transformative, powerful AI.

They're closely aligned: Amodei's as-early-as 2026–27 window overlaps Aschenbrenner's 2027, and both use a strict autonomy-centric bar. The main difference is that Amodei attaches explicit uncertainty rather than a single focal year.

The window is live but the destination is undemonstrated: capability benchmarks are strong as of mid-2026, but long-horizon autonomous systems — the core of his definition — haven't appeared. It resolves within roughly the next 18 months.