| LKE: Start simple | LKE Enterprise: Scale with confidence |
|---|---|
| Free control plane (HA option $60/cluster) | Dedicated HA control plane, flat $300/month |
| Up to 250 nodes | Up to 500 nodes |
| Calico CNI | Cilium CNI (better performance) |
| Standard NodeBalancers (Load Balancer) | Premium NodeBalancers, VPC isolation, and enhanced security |
| Access Akamai’s growing fleet of GPU nodes | Same GPU node pools, plus clusters to 5,000 GPUs |
OVERVIEW
Run Kubernetes. Not the control plane.
Deploy, manage, and scale containerized and AI workloads without worrying about control plane uptime, node failures, or version upgrades. LKE is a CNCF-certified, production-ready Kubernetes environment on Akamai Cloud. It provides GPU node pools and a pre-integrated open source AI stack within a single cluster.
Built for builders who need portability and control
HOW IT WORKS
Choose the right managed Kubernetes tier for your team
One cluster, five layers
Build on the Akamai global network, with CDN delivery, edge security, and $0.005/GB egress. Add GPU and CPU compute, and Block and Object Storage. Run LKE or LKE Enterprise on top, then App Platform for the AI stack, then your workloads.
AI WORKLOADS
AI workloads on LKE
Run inference, RAG pipelines, fine-tuning, and agents on the same managed Kubernetes platform as the rest of your applications
Seamless migration, pre-integrated tooling, zero friction
Empowering engineers, accelerating developers
FEATURES
Key LKE-managed Kubernetes capabilities
Simplify Kubernetes operations with built-in security, automated scaling, and high availability across distributed cloud environments, with the flexibility to support AI and inference workloads.
Features
- Spin up production-ready clusters in minutes using Cloud Manager, CLI, or API
- Add App Platform for LKE in one click for a pre-integrated CNCF stack: KServe, Kubeflow, Istio, Knative
- Keep clusters aligned with the latest Kubernetes releases through automated, zero-downtime version updates
- Add or remove nodes automatically as workloads change, with built-in data encryption for every node pool
- Move data at $0.005/GB egress, with Akamai CDN available on the same platform
- Handle everything from dev/test to production-grade, multi-team, and AI workloads
- Add NodeBalancers when you need them, or upgrade for advanced routing and higher throughput
- Get VPC isolation and enhanced network security on LKE Enterprise to meet compliance and multi-tenant requirements
- Choose your CNI: Calico on LKE for simplicity or Cilium on LKE Enterprise for performance and observability
Customer Stories
Frequently Asked Questions (FAQ)
Frequently Asked Questions (FAQ)
Kubernetes versions: See latest available versions
API: Full Kubernetes API compatibility
Regions: Availability
Compute: Shared, Dedicated, High Memory, GPU
Storage: Persistent volumes via Block Storage
Networking: 40 Gbps inbound, free ingress
Yes. You can migrate existing clusters using the same Kubernetes version and manifests.
LKE Enterprise includes a 99.9% SLA on the control plane and worker nodes. LKE provides best-effort uptime with standard Akamai support.
Yes: your costs begin and end with the resources you deploy (nodes, storage, load balancers). But if you choose the HA control plane, there is a cost associated with it.
LKE uses Calico for pod networking. LKE Enterprise uses Cilium, which adds eBPF performance and observability, along with VPC isolation. Network policies are available on both tiers.
LKE and LKE Enterprise track upstream Kubernetes on separate release schedules, so the versions available on each tier can differ. See the current list in the documentation. Automated, minimal downtime upgrades are included on both.
GPU node pools use NVIDIA RTX PRO™ 6000 Blackwell Server Edition® and RTX 4000 Ada, with up to 8 GPUs per node and dedicated vCPU cores. App Platform for LKE adds a pre-integrated open source AI stack, and published guides cover LLM inference, RAG pipelines, and fine-tuning.
Yes. LKE supports scalable AI inference workloads. It features native GPU support, GPU-aware autoscaling, and enterprise-grade networking and storage. Because LKE is CNCF AI Conformant, workloads meet recognized industry standards without the complexity or cost of hyperscale providers.
CNCF certification means your Kubernetes environment meets a rigorous, vendor-neutral standard for reliability, compatibility, and interoperability. In practice, it means your workloads, tools, and APIs behave consistently whether you'’re running on LKE or anywhere else in the cloud native ecosystem. For AI teams specifically, CNCF AI Conformance means LKE has been validated for the workload patterns and infrastructure requirements that modern AI deployment demands. You get the confidence of a proven standard without vendor lock-in.
Explore resources to deploy AI on Kubernetes
See if you qualify for up to US$5,000 in cloud credits
We’re here to help you with large-scale deployments, migrations, app optimization, architecture assessments, security and reliability audits, and much more. For support or account issues, please visit our support page.
