OpenAI's storage playbook: 12x lower read latency for 1.2B users across 40 regions
OpenAIDevs · x · 2026-10-24
OpenAI's dev team walks through its storage architecture with Azure Cosmos DB serving 1.2B+ users in 40 regions: broad replication reserved for small critical datasets like auth metadata, in-region caching for hot settings/app state, and a Rust rewrite to cut proxy overhead — together taking global median read p50 from 120ms to 10ms, with a chart showing each drop.
Related event: OpenAI cuts global read latency from 120ms to 10ms with storage overhaul(3 posts)→
More from Infra
- NVIDIA makes 8-node DGX Spark clusters official: 97.4% scaling efficiency on 900M training — TheZachMueller · 2026-10-24
- One NVIDIA Vera Rubin rack would set you back at least ~$7M — LinusEkenstam · 2026-10-24
- PS5 Pro sells out in Toronto as memory shortage hits consumer hardware — Rasmic · 2026-10-24
- Open-source kapitsa engine boosts Mac local agent throughput 23% with tree speculative decoding — DotaMate · 2026-10-24
- Nvidia compute unlock boosts RTX 3080 20GB prefill by up to 29%, decode unchanged — No-Refrigerator-1672 · 2026-10-24
- Baseten unveils Project Beacon to build safety tooling into the inference layer for open models — baseten · 2026-10-24