Definition
Kubernetes, often written K8s, is an open-source system for running containerised applications across a fleet of machines. You describe the desired state, for example "run three copies of this service and keep them healthy", and Kubernetes places the containers, restarts them if they crash, scales them when traffic rises and rolls out new versions without downtime. It was created at Google, released in 2014 and is now governed by the Cloud Native Computing Foundation.
In a company, Kubernetes is the operating layer that large engineering teams use to run dozens or hundreds of services in production. Every major cloud offers a managed version, such as Amazon EKS, Google GKE and Azure AKS, which removes the hardest part of running it. Kubernetes gives teams a consistent way to deploy anything, from web services to AI inference, on any cloud, and it is the reason microservices architectures are operationally feasible at all.
According to the CNCF Annual Cloud Native Survey published in January 2026, 82% of container users run Kubernetes in production and 66% of AI adopters use it to scale inference workloads. The misconception is that this makes it the right choice for everyone. Kubernetes is complex, and a small team can lose months to it. Most companies below a certain size are better served by a PaaS or a simple managed container service, and should adopt Kubernetes when their operational pain justifies the learning curve.
In practice
A SaaS company with 40 engineers and 60 services moved to managed Kubernetes and cut deployment incidents sharply, because every service was deployed, monitored and scaled the same way. A 5-person team that copied the setup spent a quarter on infrastructure it did not need yet.
Why it matters
Kubernetes is a powerful tool with a real cost in engineering time. When your CTO proposes it, the right question is not "is it good" but "what operational problem do we have today that this solves".
Frequently asked questions
- Do I need Kubernetes?
- Probably not yet, if you are a small team with a handful of services. A PaaS or a managed container service is simpler and cheaper in engineering hours. Kubernetes starts to pay off when you run many services, need fine-grained scaling, or want the same deployment model across several clouds.
- What is the difference between Docker and Kubernetes?
- Docker packages an application into a container and runs it on one machine. Kubernetes orchestrates many containers across many machines: it schedules them, keeps them running, scales them and updates them. Docker is the shipping box; Kubernetes is the port that moves thousands of boxes.