Infrastructure for AI is Finally Getting a Standard

By Puja Abbassi

On November 11, 2025

Over the past year, AI has rapidly transitioned from experimental models to critical production use cases, powering customer-facing applications and internal copilots. However, while AI models evolved at an incredible pace, the underlying infrastructure often lagged, characterized by bespoke configurations, opaque tooling, and significant vendor lock-in.

This landscape is now changing. At KubeCon North America in Atlanta, Chris Aniszczyk, CTO of the CNCF, officially announced the launch of the Kubernetes AI Conformance Program. This initiative also introduced the first set of certified Kubernetes-based platforms.

We are proud to announce that Giant Swarm is among these pioneering platforms, having received official recognition for its readiness to run AI/ML workloads in a standardized, cloud-native manner.

Why This Standard Matters

The CNCF AI Conformance Program is the culmination of months of dedicated effort by the AI Conformance Working Group. Its core objective is to define a consistent set of capabilities, APIs, and configurations that a Kubernetes cluster must offer to reliably and efficiently execute AI/ML workloads.

This standardization is crucial because AI/ML workloads, particularly those involving generative AI, have swiftly become central to the strategic initiatives of numerous organizations. Until now, there has been no shared baseline to assess whether a Kubernetes platform could consistently support these demanding workloads at scale. This new initiative provides much-needed clarity, reducing uncertainty for teams developing AI infrastructure and laying a robust foundation for interoperability, portability, and ecosystem-wide growth.

According to research cited by the CNCF, an impressive 82% of organizations are already building custom AI solutions, with 58% leveraging Kubernetes to support these workloads. As AI becomes increasingly integral to business strategy, teams require shared, open standards to mitigate fragmentation and ensure consistent performance. This certification program directly addresses that need, offering a common foundation across diverse vendors, clouds, and frameworks.

What We’ve Seen in the Field

At Giant Swarm, we’ve observed firsthand how AI/ML workloads have transformed from niche use cases to critical components of modern product development. This shift has accelerated dramatically with the advent of generative models, which introduce new infrastructure requirements, particularly concerning GPU sharing, distributed scheduling, and robust governance.

For years, we have collaborated closely with our customers to integrate these advanced capabilities into the Giant Swarm Kubernetes platform. This encompasses everything from GPU-aware scheduling and storage tuning to customized security policies and comprehensive observability for model pipelines. In many instances, we’ve successfully extended existing platform engineering patterns to meet these evolving needs without reinventing fundamental solutions.

Betting on Standards and Helping Build Them

One of Giant Swarm’s foundational principles is to adopt community standards when they exist and actively contribute to shaping them when they don’t. This commitment is why we have contributed to significant initiatives such as the CIS Benchmarks for Kubernetes, Cluster API, and now, this crucial AI conformance standard. We firmly believe that such community collaboration is indispensable for constructing enduring infrastructure.

“We believe in open standards, this certification validates our platform's readiness for AI, and it gives our customers the confidence to build the future.”

— Timo Derstappen, Giant Swarm Co-Founder and CTO

A Strategic Choice with Broad Industry Support

For teams investing heavily in AI, one of the most challenging decisions is where to place their infrastructure bets. Concerns about choosing the wrong stack, scalability limitations, or adaptability issues are common.

This conformance program directly addresses those concerns. It demonstrates that Kubernetes-based platforms are not only viable but are already being utilized by leading companies such as Bloomberg, Zalando, OpenAI, NVIDIA, and Apple. With strong backing from major cloud providers and the CNCF, Kubernetes is increasingly establishing itself as the de facto foundation for modern AI infrastructure. This robust ecosystem support helps mitigate risks and empowers teams to move forward with confidence.

Where We Go From Here

At their core, AI/ML platforms are an extension of traditional developer platforms. They necessitate the same fundamental principles: self-service capabilities, robust governance, scalability, and unwavering reliability. Many of the features we’ve meticulously built, such as multi-cluster fleet management, central observability, GitOps workflows, and hardened security baselines, are directly applicable to this new class of workloads.

We view AI as the next frontier in platform engineering, and we are enthusiastic about collaborating with both our customers and the wider community to define its future. For us, this certification marks just one significant milestone. The real work continues: supporting teams actively building AI platforms today and ensuring they are well-prepared for whatever comes next.