Akamai introduced Akamai Inference and Akamai Agent Sandbox, managed model serving and isolated agent execution on Akamai Cloud.
Akamai launched two products for enterprise AI on Tuesday, a managed inference service and an isolated execution environment for AI agents, as the cloud company moves further from content delivery into the infrastructure behind AI applications.
The Cambridge, Massachusetts company said Akamai Inference brings managed model serving to Akamai Cloud, while Akamai Agent Sandbox provides isolated execution for agents. Together they let enterprises build and tune their own or hosted open-weight models, ground them in their own data, run agents inside enforced boundaries, and secure the applications, agents and infrastructure involved, all on one platform.
Akamai said the foundation underneath the new services has already secured multiyear contracts collectively worth more than $14 billion, and that its network spans more than 4,350 edge locations alongside dense GPU and CPU clusters in core regions. The company presents the launch as the latest step in an AI strategy built on serving, running and securing AI workloads close to users and their data.
Akamai's own announcement gives the product names, the $14 billion contract figure and the edge-location count, and the announcement itself is the source for the claim that agents run inside enforced boundaries. Akamai has not published pricing, customer names or performance benchmarks for the two products in the announcement, so this report states neither and treats the contract figure as a cumulative multiyear total, not quarterly revenue.
The launch matters because the hardest part of putting agents into production is not the model but the boundary around what the agent can reach and the cost of serving it reliably. Akamai is betting that enterprises will pay for a single platform that handles serving, data grounding and agent isolation rather than assembling those pieces from separate providers. The test will be whether customers adopt the new services at a pace that turns the multiyear contract base into repeatable AI revenue.