Glostarep

How OpenAI Used Ory to Scale Identity Management to 900 Million Users

How OpenAI Used Ory to Scale Identity Management to 900 Million Users

In the history of software, few milestones are as staggering as 900 million weekly active users. To put that in perspective, OpenAI’s user base now rivals the populations of entire continents, all interacting with a high-compute, stateful AI environment.

At that scale, identity management becomes infrastructure. It is no longer just a login box. It is, instead, the gatekeeper of system stability, data privacy, and global accessibility, and it needs to scale accordingly.

When OpenAI launched ChatGPT, they didn’t just release a product; they triggered a global shift in computing. ChatGPT reached 1 million users in just five days. Traditional identity providers, however, weren’t built for this. Most rely on rigid database schemas that don’t distribute well globally. Latency compounds quickly. And critically, they offer limited control over the user experience or the roadmap.

When OpenAI faced the challenge of managing millions of identities, they didn’t choose a traditional, monolithic “Identity-as-a-Service” provider. Instead, they leveraged Ory to build a bespoke, hardened, and infinitely scalable identity system.

The results of OpenAI’s identity scaling with Ory speak clearly. OpenAI successfully navigated what Benjamin Billings, Engineering Manager at OpenAI, described as a “success disaster”, the explosive growth from 200 million to over 400 million weekly active users within just a few months. The migration replaced the legacy login system with zero disruption to active sessions.

Beyond raw scale, the move also resolved a common enterprise dilemma. OpenAI gained full control over its identity data and user interface. Previously, that visibility was partly lost when using third-party full-stack providers. Now, the team tracks traffic and behavior in-house, enabling precise optimization.

The engineering velocity gains were equally significant. Rather than spending months building and maintaining a custom identity stack, OpenAI relied on Ory’s battle-tested Ory Hydra engine and responsive support. That freed the team to focus on its core mission. Billings described the partnership plainly: “You’ve been very responsive. You’ve done changes that we’ve asked for in a very timely manner.”

Moreover, the Ory-powered architecture avoids the vendor lock-in trap. A small team of security engineers can manage nearly a billion identities, using hardened, audited open-source that the world’s best researchers trust, while delivering tokens and validating sessions at the edge of the internet.

As OpenAI approaches the 1 billion weekly user mark, the identity layer remains one of the most critical, yet invisible, components of the platform. That invisibility is, of course, the point.

Leave a Comment

Your email address will not be published. Required fields are marked *