Akamai Cloud Keeps Strengthening the Foundation (Updated August 2026)

AI applications don’t rely on accelerated compute alone. They depend on networking, storage, Kubernetes, identity, observability, and edge services all working together as a cohesive platform. As customers move from experimentation to production, those capabilities become just as important as the models themselves.

That’s where much of our work focused during Q2. Across networking, identity, observability, Kubernetes, and storage, we continued strengthening the platform that underpins AI inference and agentic applications. This post highlights the major platform improvements we delivered during Q2 2026.

Enterprise teams rarely deploy infrastructure within unconfigured environments. Networks have already been designed. Organizations established network address plans years ago. Hybrid connectivity is already in place. Security teams have already defined how traffic moves between environments.

Our networking updates in Q2 focused on fitting into those environments rather than asking customers to redesign them.

Akamai Cloud now supports configurable IPv4 ranges for VPCs. Customers can define custom ranges, including nontraditional address spaces, while built-in safeguards help prevent routing conflicts and maintain network isolation.

For teams operating hybrid or multicloud environments, this update removes one of the more common migration hurdles. Existing addressing schemes remain intact, allowing workloads to communicate with existing infrastructure without renumbering networks or introducing translation layers.

Cloud interfaces and firewall templates are now generally available, simplifying network setup for compute instances.

• Cloud firewalls can now be attached directly to interfaces
• Preconfigured templates simplify common deployment patterns
• Networking policies are managed from a single location

The result is a more consistent security posture from the moment the new infrastructure is deployed.

Access management changes as organizations grow. Once manual processes that work for a handful of engineers become operational, compliance challenges scale across hundreds of users.

As applications become more capable, operating them becomes more complex. Visibility into the platform is essential for understanding what changed, when it changed, and how those changes affect production systems.

Managed Kubernetes should reduce operational work, not create more of it.

Q2 brought Kubernetes v1.34 to new and existing LKE Enterprise clusters, along with updated CoreDNS and storage components. Customers now receive the latest platform capabilities without having to plan or coordinate upgrades themselves.

We also introduced automatic kube-reserved resource allocation, helping protect worker nodes from resource exhaustion during periods of high demand. Kubernetes API audit logs are now available for LKE Enterprise clusters, improving auditing, compliance, and operational troubleshooting.

Additionally, we added full support for the Kubernetes Gateway API and made Tekton optional for existing CI/CD pipelines. We also continued improving workflows for installation and upgrades.
 

We upgraded the Tokyo Object Storage endpoint (jp-tyo-1.linodeobjects.com) from E2 to E3 architecture, requiring no configuration changes from existing customers.

For workloads serving users across the Asia-Pacific region, the upgrade provides higher throughput and lower latency while preserving compatibility with existing applications.

Managed PostgreSQL Databases powered by Aiven now support PGBouncer connection pooling, improving efficiency for applications that establish large numbers of short-lived database connections, a common pattern for modern APIs and AI applications.

Building a platform for AI inference and agentic applications requires more than adding new AI capabilities. It requires making every layer of the platform work together, from networking and identity to Kubernetes, storage, and observability.

That’s where much of our investment went in Q2 2026. Although these improvements may not be as visible as introducing a new service, they’re what make production applications easier to deploy, easier to operate, and easier to scale.

That’s the update for this quarter. We’ll continue to share our journey in future blog posts, highlighting not only what we've delivered, but how each investment moves Akamai Cloud closer to the ultimate platform we're building for the next generation of AI applications.

In the meantime, you can try Akamai Cloud for yourself.