Just 300K hours of robot data exist; Bessemer sees $3B spend in two years
0xSammy · x · 2026-10-17
Khala Research's report argues robotics is still waiting for its "ChatGPT moment": LLMs learned from the internet, robots need costly physical demonstrations.
- Bessemer estimated only 300K hours of robot manipulation data globally in April; training may require 100M+ hours
- Bessemer projects $3B+ in robot-data spending over two years
- Open networks use crypto incentives to coordinate contributors across simulation, teleoperation, and human video
- Axis Robotics reports 7M+ trajectories from 220K+ contributors, lifting LIBERO-Plus task success from 83.9% to 88.8%
Related event: Robotics still awaiting its 'ChatGPT moment' as data gap hits 100x(2 posts)→
More from Venture
- CoreWeave CEO Calls Stock Sell-Off Irrational as Founders Cash Out $3.3B — GaryMarcus · 2026-10-17
- Investor Brad Gerstner talked up AI stocks while quietly cutting exposure to a 3-year low — SumitGup · 2026-10-17
- Dev ships AI-coded Steam game Nekomancer with a few hundred prompts, 1,500 commits — ssh4net · 2026-10-17
- Polymarket odds for 2026's #1 AI model: OpenAI 55%, Meta 16%, Google marked Yes — nian2326076 · 2026-10-17
- Onchain AI inference surging: Venice hits 500B tokens/day as Bittensor expands subnets — 0xJeff · 2026-10-17
- Chamath's $200B Anthropic Argument: Balance Sheets May Beat Smarter Models — CurieuxExplorer · 2026-10-17