OpenAI details how it scaled Habitat storage for ChatGPT growth

OpenAI said it has rebuilt its internal storage platform, Habitat, to support rapid growth in ChatGPT and other products. The company said Habitat now handles more than 70 million requests per second, serves more than 1 billion people each week, and manages over 500 petabytes of data across nearly 40 regions. The post explains how Habitat began in mid2024 as a Python library tied to a single database and later became a standalone service. OpenAI said moving the storage logic into a service helped it centralize deployment, observability, access control, and data security while reducing the coordination burden across many product teams. OpenAI also described the engineering tradeoffs of running a highthroughput Python service at this scale, including issues such as asyncio scheduling delays, connection pooling behavior, and downstream load management. The company said it later rewrote Habitat in Rust in Q2 2026, and that the new version now handles most production traffic with lower CPU and memory use.