DeepMind researcher: global AI compute growing at least 10x per year
141_1337 · reddit · 2026-11-20
A Google DeepMind researcher estimates global effective compute stock is growing at least 10x per year when multiplying hardware growth, investment, and efficiency gains — a factor of 10 billion over a decade. The writeup projects that a cognitive workload costing $1B in compute tokens today would cost about $1,000 in ten years, with inference efficiency alone cutting demand 4x annually. A recent Epoch study finds benchmark-matched consumer prices falling over 13x per year since 2023. If trends hold, you could run more high-quality AI agents than there are people on Earth for each agent you run today.
Related event: DeepMind Researcher Estimates Global Effective Compute Grows 10x Per Year(2 posts)→
More from AGI Musings
- "Human" is Cambridge's word of the year; Mustafa Suleyman calls for humanist AI subordinate to humanity — TheTuringPost · 2026-11-20
- Robotics eval shop bets 2027 brings the ChatGPT 3.5 moment for robot models — paigeinsf · 2026-11-20
- 'What I Believe About AI': a one-page credo on opportunity, cost, and human worth — MazMansoor · 2026-11-20
- Pushing back on Bostrom: existing people's lives can't be gambled for descendants' forever — AndyMasley · 2026-11-20
- Rob LeClerc: AI is rapidly changing the investment thesis for almost every company — robleclerc · 2026-11-20
- Gary Marcus stands by his take, 16 months on: AGI is neither impossible nor imminent — GaryMarcus · 2026-11-20