Rumor: Nvidia to pour massive compute into Arete-27B model
generativist · x · 2026-11-09
Account valhalladev claims Nvidia is about to dedicate substantial compute behind Arete-27B, its in-house 27B-scale model. Unverified for now; if true, it signals Nvidia doubling down on training its own frontier models.
More from Infra
- Startup Mirendil claims a Kimi K3-level model trained by one researcher for under $10M — Ok_Warthog6007 · 2026-11-09
- SkyFi thermal satellite imagery reveals which data centers are actually powered up — bradneuberg · 2026-11-09
- Epoch AI: 2027 chip supply could run 30–170M frontier AI agents, or 1.9B cheaper ones — Dr_Singularity · 2026-11-09
- White paper: moving beyond the AI compute arms race as big tech faces historic spending losses — ReportProfessional57 · 2026-11-09
- LLMs max out hardware: former xAI exec reports failing PSUs, tripping breakers — MParakhin · 2026-11-09
- Local LLM reality check: $3K rig with 32GB GPU handles simple tasks only — 2C_ornot2C · 2026-11-09