Headline: How AI is changing open source

Key facts at a glance

• Open source has transformed from a development movement to a strategic control plane for AI infrastructure.
• Corporate contributors dominate: Red Hat, Microsoft, and Google lead CNCF contributions, with Red Hat at 194,699 contributions in 2025.
• Kubernetes is the de facto operating system for AI, with 66% of organizations using it for inference workloads.
• OpenTelemetry, Cilium, and Kubeflow are rising fast as critical infrastructure projects.
• Nvidia is investing in open source scheduling and orchestration (e.g., KAI Scheduler, Kubeflow) to optimize chip usage.
• 98% of organizations now use cloud-native techniques, and 82% run Kubernetes in production.

Open source has quietly undergone a radical transformation. While headlines focus on proprietary AI models and the latest breakthroughs from OpenAI or Google DeepMind, the real revolution is happening in the plumbing—the open source infrastructure that powers AI workloads in production. This is not the open source of the early 2000s, driven by volunteer developers and ideological debates. Instead, it is a cold, calculated, and highly strategic investment by the world’s largest technology companies. The goal is not just to give away code but to set the standards, control the interfaces, and shape the operational assumptions that everyone else must adopt.

According to the Cloud Native Computing Foundation (CNCF), its 2025 survey found that 98% of organizations have adopted cloud-native techniques, and 82% of container users run Kubernetes in production. GitHub’s Octoverse report for the same year recorded 1.12 billion contributions, more than 180 million developers, and 518.7 million merged pull requests. The Apache Software Foundation (ASF) reported 9,905 committers across 295 projects and 1,310 software releases in fiscal year 2025. These numbers are staggering, but they tell only part of the story. The real shift is in who is contributing and why.

The corporate takeover of open source

In 2025, Red Hat led CNCF contribution activity with 194,699 contributions. Microsoft followed with 107,645, and Google with 91,158. Independent contributors still mattered—landing fourth with 52,404—but the center of gravity is unmistakable. These companies are not acting out of charity. They are investing in the long-term control of the infrastructure layers that their products and services depend on. This is a classic land-grab strategy: shape the substrate, and you get leverage over everything built on top of it.

Red Hat’s dominance is easily explained. Its OpenShift platform is built around Kubernetes. By pouring engineering resources into Kubernetes and related CNCF projects, Red Hat ensures that its platform remains relevant and that the ecosystem evolves in ways that benefit OpenShift. Microsoft’s second-place position is more surprising to those who remember the company’s historical hostility to open source. But Microsoft has made a strategic pivot: its Azure cloud relies heavily on open source technologies, and contributing to projects like Kubernetes and OpenTelemetry helps Microsoft ensure that those technologies work well with Azure and that Microsoft has a voice in their direction.

Google, the birthplace of Kubernetes, remains a top contributor. But its contributions are now more focused than ever. Google is heavily involved in Cilium, a project at the intersection of networking, observability, and security. Cilium’s journey report shows that the number of contributing companies rose 90% after it joined the CNCF, from 533 to 1,011, while individual contributors jumped from 1,269 to 4,464. Google, Datadog, and Cloudflare all expanded their contributions as the project matured. Cilium sits at the critical junction of distributed workloads, latency sensitivity, and cost management—exactly the areas where AI workloads push infrastructure to its limits.

Observability as a strategic battlefield

OpenTelemetry has become one of the fastest-rising CNCF projects, with a 39% rise in commits in 2025 and a contributor base that grew from 1,301 to 1,756 in a single year. This is not a coincidence. Observability is the ability to understand and debug complex distributed systems, and with AI models running in production, observability becomes mission-critical. Companies like Microsoft and Splunk are among the top contributors to OpenTelemetry. Their goal is not to build a community-driven tool but to create a standard that others will be forced to adopt, thereby locking in their own observability solutions. This is the new reality: open source is a battlefield for setting defaults.

Nvidia: From hardware to open source orchestration

Perhaps the most telling example of this strategic shift is Nvidia. Nvidia has so much cash that it could single-handedly fund entire ecosystems, yet it chooses to contribute to open source projects rather than buy them outright. Nvidia ranked 14th in Kubernetes contributions in the past two years with 5,892 contributions. It open sourced the KAI Scheduler, a Kubernetes-native GPU scheduler derived from its Run:ai acquisition. Nvidia also describes itself as a key contributor to Kubeflow, the machine learning toolkit built on Kubernetes. By investing in the scheduling, orchestration, and workflow layers, Nvidia ensures that its GPUs are the easiest to use in the most popular open source AI stacks. This is not charity; it is a strategic move to make Nvidia hardware indispensable.

AI needs open infrastructure

The rise of AI is making open infrastructure more important, not less. CNCF reports that 66% of organizations hosting generative AI models now use Kubernetes for some or all inference workloads. Kubernetes is increasingly called the de facto operating system for AI. This is because most enterprises do not want to build their entire future on opaque, proprietary infrastructure that they cannot inspect, influence, or customize. Open source provides transparency and flexibility, but it also provides a layer of control for the companies that invest in it.

The old narrative of open source as a fringe alternative or a developer-led morality play is dead. It has been replaced by a new narrative: open source is the control plane for the next generation of computing. The companies that shape that control plane will dominate the AI era. They are not contributing out of altruism but out of a hard-nosed understanding that whoever sets the standards for infrastructure will reap the rewards of the applications built on top of it.

Open source grew up and became dull. We are all better for that. The romance of the early days is gone, but in its place is something far more powerful: a mature, strategic ecosystem where the most important decisions about AI infrastructure are made in open, collaborative communities. The winners will not be the ones with the best models but the ones who control the platforms that run them.

Source: InfoWorld News