ChipsSeptember 11, 2026via OpenAI Blog

Rapidly scaling online storage to serve over 1 billion ChatGPT users

Why it matters

Infrastructure deep-dive: OpenAI is publishing its storage architecture for serving ChatGPT at scale—a rare window into the compute and backend buildout required to run the world's largest AI application.

Key signals

  • Habitat platform serves 1 billion ChatGPT users
  • 22M requests per second throughput
  • Evolved from Python library to globally distributed system
  • OpenAI engineering blog (part one of series)
  • Published September 2026
  • Habitat evolved from Python library to globally distributed storage platform
  • Serves 1 billion ChatGPT users
  • Handles 22M requests per second
  • Published as engineering deep-dive (part one of series)
  • Infrastructure/systems story, not model or product feature

The hook

22M requests per second. Here's how OpenAI's Habitat storage platform evolved to serve a billion ChatGPT users without melting.

Learn how OpenAI evolved Habitat from a Python library into a globally distributed storage platform serving 1 billion ChatGPT users and 22M requests per second.

The week's key stories, every Friday.

ONE BRIEFING · EVERY FRIDAY · FREE

Free. Unsubscribe anytime.