Global AI compute additions double every 9.1 months, topping 6.3M H100e per quarter: Epoch
dl_weekly · x · 2026-11-12
Epoch AI's data insight: new global AI computing capacity doubles roughly every 9.1 months. Quarterly additions grew from 0.83M H100-equivalents in Q1 2024 to 6.33M in Q2 2026.
Capacity spans six chip designers — Nvidia, Google, AMD, Amazon, Huawei, Cambricon. Nvidia shipped 3.95M H100e in Q2 2026; Google reached 1.6M as its TPUs scale. Growth has slowed from 6×/year in 2023 to 2.5×/year; evidence of further deceleration isn't statistically significant and may reflect a generational lull, with Nvidia's Rubin chips not yet recognized in revenue.
More from Infra
- Five IndiaAI Mission-empanelled cloud providers lack data centres, barred from compute tender — kingslyj · 2026-11-12
- Startup turns Apple silicon into elastic cloud compute for AI agents, touting 512GB M5 Ultra clusters — awnihannun · 2026-11-12
- NVIDIA takes $4.5 billion H20 charge after new export licensing requirements — heyshrutimishra · 2026-11-12
- LangChain releases fireside chat with Modal CEO on sandboxes and inference at Interrupt NYC — LangChain · 2026-11-12
- SteerDistill: teacher-guided quantization keeps W4A16 agents within 1 point of BF16 — _akhaliq · 2026-11-12
- Arm CEO: hyperscaler server share tripled to 50%+, 60% by end of 2026 — Beth_Kindig · 2026-11-12