Akamai Launches AI Inference Service and Agent Sandbox for Enterprise AI Deployment

Akamai Expands AI Services with Inference and Sandbox Solutions

Akamai Technologies has launched two foundational AI services that transform its cloud infrastructure into a complete AI stack1

1

. Akamai Inference provides managed access to open-weight AI models through OpenAI-compatible APIs, while Akamai Agent Sandbox offers a secure execution environment for AI agents2

2

. These services address enterprise demand for flexible yet reliable serving infrastructure and enable organizations to integrate AI with proprietary data sources under strict governance protocols2

2

.

The AI Cloud platform now delivers a connected AI environment spanning dense GPU and CPU clusters in core regions to more than 4,350 edge locations3

3

. Adam Karon, Chief Operating Officer and general manager of Akamai Cloud, stated that the company's AI infrastructure strategy centers on inference representing the majority of enterprise focus and investment1

1

. The momentum behind this strategy is evident in multiyear contracts collectively worth more than $14 billion that Akamai Cloud has already secured3

3

.

Akamai Inference Delivers Managed Model Access

Akamai Inference gives enterprises managed access to 12 open-weight models ranging from 8 billion to 2.8 trillion parameters4

4

. The initial catalog includes NVIDIA Nemotron 3 Ultra, OpenAI gpt-oss-120b, and Google Gemma 4, alongside embedding models for retrieval-augmented generation pipelines1

1

. These RAG pipelines connect open-weight models with Akamai's managed database services to ensure responses are grounded in enterprise data2

2

.

The AI inference service offers customers two deployment options: serverless, pay-per-use inference or dedicated deployments5

5

. The service handles model execution, orchestration, and authentication while providing access to managed databases, object storage, and connections to private and public cloud platforms1

1

. By using OpenAI-compatible APIs, developers can integrate the service with existing applications without rewriting code1

1

. Requests run across Akamai's global network to reduce latency and support high availability3

3

.

Agent Sandbox Provides Isolated Execution for Autonomous AI

Akamai Agent Sandbox addresses a pressing need as autonomous AI agents take on more software coding responsibilities2

2

. The sandbox provides an isolated environment for executing code written by AI agents, separating it from the agent's own runtime1

1

. Agents run on Akamai Functions while their generated code executes in an isolated environment using WebAssembly-based isolation1

1

.

The secure execution environment for AI agents enforces limits on compute, memory, outbound network connections, and system capabilities4

4

. Jon Alexander, senior vice president of cloud technology at Akamai, emphasized that every team running agents needs a boundary around the code those agents write1

1

. The sandbox is being released as an open-source component of the Spin framework, with plans to add capabilities governing what AI agents can access and which systems can interact with them1

1

.

NVIDIA Blackwell Ultra Strengthens GPU Infrastructure

Akamai has expanded its dedicated GPU fleet with NVIDIA Blackwell Ultra clusters and is preparing data centers for NVIDIA Vera Rubin platform deployment in 20271

1

. The GPU fleet now includes NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs for low-latency inference5

5

. Each eight-GPU Blackwell Ultra node provides 288 GB of HBM3e memory per GPU, totaling 2.3 TB across the node4

4

.

This infrastructure targets workloads involving test-time scaling, chain-of-thought reasoning, long-context reasoning, and agentic work5

5

. Dedicated GPU clusters give enterprises control over model versions and deprecation timelines without rate limits or queueing associated with shared infrastructure1

1

. Akamai allows developers to choose architectures fitting their workload, with NVIDIA-accelerated and dedicated CPU infrastructure, managed Kubernetes, and Akamai Functions for serverless deployments across edge and core locations5

5

. Both services are currently available in beta, with the inference beta including serverless deployment options4

4

.