Codex Hits 4M Weekly Users, OpenAI Pushes It Into Big Enterprise with Accenture and PwC

OpenAI dropped a couple of interesting numbers and a big partnership play for Codex this week. First, the product now has 4 million weekly active users. That’s a solid number for a developer tool that’s still relatively young, and it tells me they’re seeing real stickiness beyond the initial curiosity wave.

But the bigger news is Codex Labs. This is OpenAI’s formal push to get Codex into the hands of large enterprises — not just individual developers tinkering with it, but actual teams inside Accenture, PwC, Infosys, and a handful of other global systems integrators. The idea is to help these firms deploy and scale Codex across the full software development lifecycle, not just code completion or small snippets.

This is the right move. The early adopters for AI coding tools have mostly been startups and individual devs. Enterprises move slower, have compliance requirements, need training, and want guarantees around security and data handling. A self-serve API alone wasn’t going to cut it for a Fortune 500 company. Partnering with consultancies that already have deep relationships and integration expertise is how you actually get code assistants into legacy banking systems and government contracts.

I’m curious to see how much of this is about customization. Codex is already powerful out of the box, but enterprises have their own codebases, style guides, and internal libraries. If Codex Labs can help Accenture or PwC fine-tune the model on client-specific code without leaking data, that’s the kind of thing that will justify the price tag.

That said, I don’t think this is a slam dunk. Enterprise sales cycles are long, and consultancies love to rebrand everything as their own IP. There’s a risk that the partners end up wrapping Codex in so many layers of process and consulting fees that it loses the speed advantage that makes it useful in the first place. OpenAI will need to keep the product experience clean even as it gets institutionalized.

Still, 4M WAU is a real milestone. For context, that’s more than many standalone developer tools have ever hit. If even a fraction of those convert to paid enterprise seats through these partnerships, it’s a meaningful revenue stream. And it signals that OpenAI is serious about making Codex more than a demo toy — they want it embedded in how software gets built, from spec to deployment.

I’d keep an eye on how quickly these partners ship actual case studies. That will tell me whether Codex Labs is a real product play or just another press release partnership.