Anthropic and Akamai Ink Landmark $11.6 Billion Compute Partnership to Fuel Long-Term AI Infrastructure Growth

To meet the immense technical requirements of the contract, Akamai is preparing for a period of significant investment. The company projected that it will undertake approximately $5.5 billion in capital expenditures directly related to this agreement. This capital will likely be used to build out data center capacity, procure hardware, and enhance the "Gecko" (Generalized Edge Compute) initiative, which aims to bring full cloud computing capabilities to the very edge of the network.

Rapid infrastructure investment

The $11.6 billion deal with Akamai is the latest chapter in what has become an unprecedented infrastructure-building spree for Anthropic. Over the last 24 months, the company has moved with remarkable speed to secure the compute capacity necessary to compete with rivals like OpenAI and Google. As Anthropic ramps up its development timelines for future frontier models and expands its global customer base, it has consistently sought out multi-billion-dollar deals to ensure it is never throttled by a lack of hardware.

This strategy of diversification is evident in the variety of partnerships Anthropic has secured. In May, the company made headlines by announcing a deal with SpaceX to utilize the entire compute capacity of the Colossus 1 data center in Tennessee. Hosted by Elon Musk’s xAI, this facility provided Anthropic with access to more than 300 megawatts of new compute capacity and approximately 220,000 Nvidia GPUs. That deal was seen as a move to secure the raw "training" power needed for the next generation of large language models.

However, training is only one side of the AI equation. Once a model is trained, it must be deployed to users around the world with low latency and high reliability—a process known as inference. This is where the Akamai partnership becomes critical. By utilizing Akamai’s distributed AI infrastructure, Anthropic can ensure that Claude is responsive and available to enterprise clients regardless of their geographic location.

The Akamai deal also follows a major partnership expansion with Google and Broadcom. That agreement focused on utilizing "multiple gigawatts" of Google’s proprietary Tensor Processing Units (TPUs). By balancing its reliance on Nvidia GPUs (via SpaceX/xAI), Google TPUs, and now Akamai’s distributed CPU and cloud infrastructure, Anthropic is building a resilient, multi-vendor hardware stack that mitigates the risks associated with supply chain bottlenecks or platform lock-in.