When does AGI arrive? The predictions compared, 2027 to never
October 6, 2026 · 23 min read
Kurzweil holds to 2029, Amodei extrapolates to 2026 or 2027, Altman expects very capable systems by the end of the decade, and Bostrom, Yampolskiy and chip analysts give their own dates and doubts. Seeing the forecasts and their reasoning together helps readers weigh the confidence behind each.
The rules, numbers and steps worth keeping in front of you.
- Forecasts cluster from 2026 to the early 2030s.
- They forecast different things: a Turing test, Nobel-level systems, or all human tasks.
- Kurzweil: 2029 for an hours-long Turing test, held since 1999.
- In 1999 experts averaged 100 years and 20 percent said never; a recent poll said 2030.
- Amodei: powerful AI around 2026 or 2027, by eyeballing the curve; not a scientific prediction.
- Amodei: effects take five to ten more years; data or chip supply could derail it.
- Altman: capable systems by the end of the decade; progress feels continuous from inside.
- Bostrom: superintelligence possible within a few years, probably somewhat later.
- Yampolskiy: capability scales with compute, safety does not, so any early date is too early.
- Lambert: after 2030; discount company heads for their incentives.
- Patel: capability by 2027 or 2028, but cost slows deployment.
- Early dates rest on extrapolating curves in compute, spending and benchmarks.
- Possible brakes: data, chips, energy, a spending ceiling, reasoning and long tasks.
- Arrival of a capability is not arrival of its impact.
This is the short form. The full insight has every expert's reasoning, their own words and the sources.