Akamai secures $11.6B CPU cloud deal with Anthropic

Akamai has entered a seven-year, $11.6 billion cloud services contract with Anthropic, marking one of the largest CPU-focused infrastructure commitments in the AI sector. The agreement highlights the often-overlooked demand for general-purpose compute in AI operations, alongside the more widely discussed need for specialized accelerators like GPUs. While the deal does not disclose specific workloads, locations, or service-level guarantees, it underscores the scale of infrastructure required to support large AI models beyond training and inference hardware.

Financial and operational implications

The $11.6 billion commitment is fixed, but the contract includes an optional expansion of up to $9 billion, contingent on future negotiations. Akamai expects to spend approximately $5.5 billion in capital expenditures to support the deal, with $1.7 billion allocated for 2026 alone. This upfront investment in components like memory and hardware precedes revenue recognition, placing procurement and deployment risks squarely on Akamai. The company has not provided margin forecasts or a detailed spending schedule, leaving investors to assess the deal’s profitability based on hardware costs, utilization rates, and Anthropic’s consumption pace.

The financial structure extends beyond traditional cloud services. Akamai issued Anthropic a warrant for up to 7.7 million common shares, equivalent to roughly 5% of outstanding stock, at an exercise price of $111.33 per share. Approximately 2% of the warrant is expected to vest immediately, with the remaining 3% tied to the optional $9 billion in additional purchases. This equity component may strengthen customer retention but complicates the deal’s economics, as potential dilution adds an ownership cost alongside capital expenditures.

Capacity and competition

For hosting and data center operators, the deal signals a shift in the AI infrastructure conversation. While GPU availability dominates industry discussions, the Anthropic agreement highlights the critical role of CPU fleets, memory supply, and network design in supporting AI workloads. Akamai’s infrastructure spans thousands of points of presence, but the company has not clarified whether Anthropic’s workloads will leverage this distributed network or be concentrated in specific locations. The lack of transparency around deployment sites, reserved capacity, and service commitments leaves other cloud buyers uncertain about how this large tenant might affect their own access to resources.

The deal also raises questions about Akamai’s broader cloud strategy. The company has disclosed over $2.8 billion in other multiyear cloud infrastructure commitments this year, but the Anthropic agreement is significantly larger. While a single major customer can justify capacity investments, it also exposes Akamai to concentration risk if Anthropic’s requirements change. The company has not shared unit economics, deployment timelines, or a breakdown of the workload mix, leaving operators to speculate about the deal’s long-term impact on the cloud services market.

What to watch

The immediate focus will be on Akamai’s capital spending and its ability to meet Anthropic’s demands without disrupting service for other customers. Hardware delivery schedules, memory availability, and power constraints at data center sites will be critical factors in determining the deal’s success. Additionally, the optional $9 billion expansion and the vesting of Anthropic’s warrant could shape Akamai’s financial performance and customer relationships in the coming years. Operators should monitor whether this model—combining cloud services with equity incentives—becomes a trend among large-scale infrastructure providers.