The CNCF is graduating projects faster than ever. AI agents are helping with the due diligence.

OpenAI wasn’t always the poster child for giant AI models. Back in 2018, the company was using Kubernetes to spread workloads across its own data centers and cloud capacity from AWS and Azure, long before ChatGPT and the current AI boom made the whole sector feel inevitable.

That history is exactly why Jonathan Bryce sees open source as central to the AI era. The CNCF executive director told The New Stack that tools like Kubernetes make it possible to handle the kind of sudden, uneven demand AI services create. Without that flexibility, companies would be stuck with a single compute setup, and progress would be a lot slower.

Bryce’s bigger point is that the relationship runs both ways. AI companies such as OpenAI had to rely on open-source software because they were building systems and training runs that nobody had done at that scale before. Now those same companies are contributing back to the tools that helped them get there, which then filters into other industries, including financial services and manufacturing, as they build out their own AI infrastructure.

The CNCF is also feeling the AI effect in its own process. Bryce says the foundation has been moving projects through graduation faster than ever over the last 18 months, partly because it has created agents to help with due diligence at different stages. Not every project is riding the AI wave, but the pace is clearly changing, and that is one reason the group is pushing toward KubeCon in Salt Lake City this November.

For Bryce, the meeting matters because the hard problems have changed. Software now has to deal with AI queries at agent speed, not human speed, while non-deterministic agents need guardrails and companies are trying to juggle multiple compute providers and hardware architectures at once. The old open-source playbook is still there, but it has been put on a much hotter burner.