Akamai Adds Pair of Managed AI Services to Securely Access Custom Models
Techstrong.ai, Tuesday, October 27th, 2026
Akamai launched two managed AI services in beta, Akamai Inference and Akamai Agent Sandbox, for running open-weight models and AI agents.
Akamai introduced two managed services in beta: Akamai Inference, which provides access to 12 open-weight models from 8 billion to 2.8 trillion parameters through an OpenAI-compatible API, including NVIDIA Nemotron 3 Ultra, OpenAI gpt-oss-120b and Google Gemma 4.
The second, Akamai Agent Sandbox, lets organizations run AI agents in a secure serverless runtime built on WebAssembly technology from its Fermyon acquisition.
The services run across more than 4,350 points of presence using CPU and NVIDIA GPU clusters, with Vera Rubin support planned for 2027.
Akamai says demand is rising as organizations deploy custom models to cut costs and reduce reliance on frontier model providers.