Next-gen model training could hit 5x pretraining compute if 300k GB200 rumor holds
scaling01 · x · 2026-09-07
scaling01 estimates next-gen frontier model training compute: 1-2x pretraining with 100k GB200s, but closer to 5x if the rumored 300k GB200s materialize. On top of that, RL may add another 2-5x of pretraining-scale compute.
Related event: KOL Tests Astra, Sees No AGI; Speculates GPT-6 Uses Up to 5x Compute(3 posts)→
More from Infra
- FP8/FP4 Quantization Delivers Only ~1.5x and 2x Real Speedups, Far Below Theoretical Gains — scaling01 · 2026-09-07
- Naura demos key etch process for 64-layer 3D DRAM without EUV, selectivity above 500:1 — pstAsiatech · 2026-09-07
- Netherlands builds 'Dutch AI' by finetuning Qwen 3.5 27B in subsidized datacenter — teortaxesTex · 2026-09-07
- This Week's AI Must-Reads: OpenAI's Research Acceleration Report and Broadcom's $16.7B AI Chip Quarter — VibeMarketer_ · 2026-09-07
- On-device Android agent with Gemma 4 E2B hits 2.6 tok/s live vs 11 tok/s on replay — HowDevelop · 2026-09-07
- Dev wrote Marlin-style FP4/FP8 kernels for RTX 3090 — NVIDIA declined to upstream them — QuixiAI · 2026-09-07