~90% of frontier lab compute now goes to post-training and inference: Social Capital
soumitrashukla9 · x · 2026-10-08
Social Capital shares a chart showing that roughly 90% of frontier AI lab compute now goes to post-training and inference rather than pretraining. The resharer says he's cited the chart at least five times since seeing it, calling the shift hugely consequential — a signal that the industry's center of gravity is moving away from raw pretraining scaling.
Related event: Frontier Labs Shift ~90% of Compute to Post-Training and Inference(5 posts)→
More from Infra
- Mapping the firm-power stack behind Google's massive nuclear deal: CEG, TLN, VST, Hubbell, Bloom — demian_ai · 2026-10-08
- Why temp=0 LLM inference still isn't deterministic: floating-point order and parallelism — ducha_aiki · 2026-10-08
- NVIDIA releases CUDA 13.4 with new features detailed in official blog — NVIDIAAI · 2026-10-08
- India and AI in Four Numbers: 38,000 Public GPUs and the Chip Rally It Missed — AlpineRupee · 2026-10-08
- Liquid AI's 600M Decision Model Runs in the Browser via WebGPU at 45ms per Query — FinancialAd1961 · 2026-10-08
- Surface Laptop Ultra launches Oct 16 starting at $2,599 with Nvidia RTX Spark — tomwarren · 2026-10-08