Akamai launches AI inference & agent sandbox tools

Akamai has launched two AI products, Akamai Inference and Akamai Agent Sandbox, expanding its cloud platform for AI applications and agents.

The services are aimed at organisations moving AI systems from trials into production, where model serving, latency, data access and agent security become harder to manage. They add managed model serving and isolated execution environments to Akamai's existing cloud infrastructure.

Akamai Inference provides managed access to open-weight and custom models through APIs designed to work with existing OpenAI-compatible applications. Customers can choose either serverless, pay-per-use inference or dedicated deployments, depending on how they want to run workloads.

At launch, the model catalogue includes 12 open-weight models ranging from 8 billion to 2.8 trillion parameters. Available models include NVIDIA Nemotron 3 Ultra, OpenAI gpt-oss-120b and Google Gemma 4, alongside embedding models for retrieval-augmented generation.

The service also connects with managed databases, object storage, and private and public cloud environments, allowing customers to use enterprise data while Akamai handles orchestration, model execution and authentication. Requests run across its global network to reduce latency and maintain service levels.

Agent controls

The second product, Akamai Agent Sandbox, focuses on the growing use of AI agents that can act on behalf of users or generate and run code. It gives developers a separate environment in which that code can run away from the agent runtime.

According to Akamai, the service runs agents on Akamai Functions, while code written by those agents runs in its own isolated environment. Delivered as an open-source component of the Spin framework, it uses WebAssembly isolation to limit compute, memory, outbound connections and system access.

Developers can also choose different infrastructure approaches for agent workloads, including dedicated CPU systems, NVIDIA-accelerated infrastructure, managed Kubernetes and serverless execution across edge and core environments.

Adam Karon, Chief Operating Officer and General Manager of Akamai Cloud, said customers had pushed the company further in developing its AI offering. "We have been implementing an AI infrastructure strategy based on the notion that AI inference will represent the majority of enterprise focus and investment. Our customers proved that thesis and pulled us further," Karon said.

"They want to build, run and protect AI on the same platform. Akamai Inference and Akamai Agent Sandbox add the inference and execution layers that turn our AI infrastructure into a full AI stack, and our security portfolio protects every layer of it. Akamai is uniquely positioned to bring our scale and distribution to address scaling real-time inference and agentic workloads," he said.

GPU expansion

Alongside the software announcements, Akamai has expanded its dedicated GPU estate to include NVIDIA Blackwell Ultra GPU clusters. These systems are intended for enterprises that want to build, tune and self-host models on infrastructure under their own control.

Its GPU estate now also includes NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs for low-latency inference. Akamai added that each eight-GPU Blackwell Ultra node includes 288 GB of HBM3e memory per GPU, or 2.3 TB in total.

The company linked the infrastructure expansion to workloads that require larger memory footprints, including test-time scaling, chain-of-thought reasoning and longer-context agent-based tasks. Customers using dedicated GPU clusters can control model versions and deprecation timelines while avoiding rate limits and shared-tenancy queues.

The announcement adds to Akamai's broader effort to position its cloud business around AI workloads running from centralised regions to edge locations. Its cloud network spans more than 4,350 edge locations and supports both CPU- and GPU-based workloads, alongside storage, networking and security services.

Akamai also said these additions sit on infrastructure that has already secured multiyear contracts for Akamai Cloud worth more than $14 billion collectively. It presented that figure as evidence of demand for a platform that combines compute, networking and security for AI-related deployments.

Jon Alexander, Senior Vice President of Cloud Technology at Akamai, said customers were looking for different ways to put AI into operation. "Enterprises should not have to choose between speed and control when they put AI into production," Alexander said.

"Some teams want to ship an AI feature this quarter on a managed platform. Others need dedicated GPUs to run frontier-class open models on their own terms. Every team running agents needs a boundary around the code those agents write. Akamai Cloud now supports all of them on one platform, with security and resilience built in," he said.