Model Race Becomes Chip Race: AI Labs Turn to Custom Hardware
atShruti · x · 2026-07-21
The current AI model race is increasingly becoming a chip race. Because AI labs are running out of power faster than they can add compute capacity, major players like Google and OpenAI are designing custom chips tailored specifically around their models.
This strategy aims to extract more AI performance per watt. By optimizing hardware for their specific architectures, they hope to achieve up to a 10x speedup, even without fundamentally changing the core model design.
More from Infra
- Encrypted GPT-2 speedrun challenges FHE inference: best is 3.031 s/token, goal sub-1s — tensorqt · 2026-10-27
- Open-source Latch compiles PyTorch into training megakernels, 2.75x faster on H100 — sguada · 2026-10-27
- NVIDIA claims HGX Rubin NVL8 delivers 10x tokens per megawatt vs B200 — thursdai_pod · 2026-10-27
- NVIDIA Claims HGX Rubin NVL8 Delivers 10x More Tokens per Megawatt Than B200 — thursdai_pod · 2026-10-27
- Together lets multiple LoRA post-training experiments share the same GPU capacity — togethercompute · 2026-10-27
- Commoditization dilemma: either human labor or GPUs become the commodity — _AustinCalvert_ · 2026-10-27