Artificial general intelligence
An AI system that matches competent humans across essentially all cognitive work, including open-ended multi-week projects, learning new skills on the job and acting autonomously in the world.
State of the art
Frontier agents complete multi-hour software tasks at ~50% reliability (METR, 2025–2026); no system sustains weeks-long autonomous work or learns persistently from it.
- Reliability collapses on tasks longer than a few hours at high success thresholds
- No persistent learning from experience after deployment
- Weak grounding in physical and social context outside text and code
Who’s working on it
- OpenAI ↗Startup
Building Stargate with Oracle, SoftBank and MGX, with ~7 GW planned by Sept 2025 toward 10 GW by 2028; states AGI as its mission
- Google DeepMind ↗Company
Released Gemini 3 (Dec 2025); Demis Hassabis frames AGI as the lab's goal
- Anthropic ↗Startup
Claude models led METR long-task time-horizon measurements in 2025-2026
- Demis HassabisResearcherGoogle DeepMind
Google DeepMind CEO and 2024 Nobel laureate (AlphaFold); directs Gemini and the lab's AGI research
- Meta Superintelligence LabsCompany
Meta AI division formed in 2025 under Alexandr Wang to build frontier general models
- METR ↗Academic lab
Measures the time horizon of AI tasks, which doubled about every 7 months from 2019 to 2025 and is the main yardstick for agent autonomy
What to invest in
| Kind | Investment | Scale |
|---|---|---|
| research | Long-horizon agent reliability, memory and self-correction research with open evaluations | $$$$ > $1B |
| infrastructure | Realistic multi-week task benchmarks and sandboxed environments | $$ $10M – $100M |