Akamai Technologies has launched two managed services aimed at helping enterprises move AI applications and autonomous agents into production: Akamai Inference is for serving and tuning models, while Akamai Agent Sandbox is for running agent-generated code within enforced boundaries.
The new additions extend Akamai Cloud from infrastructure into a fuller AI stack, combining managed model serving, data connectivity, security and distributed compute. Akamai said its cloud platform is supported by multiyear contracts collectively worth more than US$14 billion.
Inference closer to users
Akamai Inference provides managed access to open-weight and custom models through OpenAI-compatible application programming interfaces, allowing existing applications to connect without rewrites. Customers can choose serverless, pay-per-use inference or dedicated deployments.
At launch, the model catalogue includes 12 open-weight models ranging from 8 billion to 2.8 trillion parameters, including NVIDIA Nemotron 3 Ultra, OpenAI gpt-oss-120b and Google Gemma 4, alongside embedding models for retrieval-augmented generation.
The service connects to managed databases and object storage, as well as private and public cloud platforms, so models can use enterprise data. Akamai said it handles orchestration, model execution and authentication, while running requests across its global network to reduce latency and support availability.
The distributed model may be particularly relevant where users, data and regulatory requirements are spread across markets. Akamai’s compute continuum extends from GPU and CPU clusters in core regions to more than 4,350 edge locations.
Boundary for autonomous agents
Akamai Agent Sandbox addresses a growing operational concern: AI agents increasingly take actions on users’ behalf and write and execute their own code. The service runs agents on Akamai Functions, while code generated by an agent runs in a separate isolated environment.
Delivered as an open-source component of the Spin framework, the sandbox uses WebAssembly isolation to limit compute, memory, outbound connections and system capabilities. Akamai said this gives developers clearer control over where agents run and what they can reach.
The company also plans to add capabilities to govern what agents can access and what can access them. It will support alternative isolation approaches, including microVMs and policy-enforced containers, alongside NVIDIA-accelerated and dedicated CPU infrastructure, managed Kubernetes and serverless execution.
“Enterprises should not have to choose between speed and control when they put AI into production,” said Jon Alexander, senior vice-president of cloud technology at Akamai. “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.”
