Key Takeaways

  • Anthropic committed $11.6 billion over seven years, with options that could lift the agreement to $20.6 billion.
  • The deal focuses on CPU capacity, showing that AI infrastructure demand extends beyond GPU-intensive model training.
  • Akamai expects roughly $5.5 billion in related capital spending and could issue Anthropic a warrant equivalent to about 5% of its common stock.

Akamai has secured a seven-year cloud computing commitment from Anthropic valued at $11.6 billion, giving the infrastructure provider a major anchor customer as it expands its position in artificial intelligence. According to Morningstar, this is the second cloud agreement between the companies, marking an operational shift in Akamai’s effort to turn its globally distributed infrastructure into an AI computing platform.

The agreement can grow substantially. Anthropic has the option to add as much as $9 billion in further commitments, taking the potential total to $20.6 billion. That expansion is not assured, but the structure gives Anthropic room to acquire more capacity as its models, products and customer base develop.

This is not simply another contract for scarce AI accelerators. The arrangement centers on Anthropic’s growing CPU requirements, highlighting the extensive computing work surrounding generative AI. Model training attracts attention because of its appetite for GPUs. Production inference, data processing, orchestration, safety testing and supporting services also consume large amounts of conventional compute, memory, storage and network capacity.

That distinction matters for Akamai. Its historical strengths lie in distributed infrastructure, content delivery, edge computing and network security rather than in operating the biggest centralized GPU clusters. CPU-oriented AI workloads could play more naturally to that footprint, particularly when applications benefit from processing capacity closer to users or across multiple regions.

The contract follows an earlier Anthropic agreement reported in May 2026. Reuters reported that the prior cloud deal was worth $1.8 billion. Taken together, the agreements suggest a relationship moving beyond a limited capacity purchase toward a longer-term infrastructure alignment.

It is also a capital-intensive bet. Akamai estimates roughly $5.5 billion in capital expenditure connected with the new commitment. That includes about $1.7 billion of additional spending during 2026 for supply-chain components, including memory. Hardware procurement, data-center preparation and power availability will therefore influence how efficiently Akamai converts the contracted demand into revenue and cash flow.

The commercial ties could extend into ownership. A warrant associated with the agreement could give Anthropic rights equivalent to approximately 5% of Akamai’s common stock. Such an arrangement can align incentives, but it also introduces potential dilution and creates a relationship that investors will assess differently from an ordinary customer contract. How dependent will Akamai’s AI expansion become on one fast-growing customer?

Investors initially responded favorably to the announcement. SiliconANGLE reported that Akamai shares jumped more than 20% following news of the agreement. The reaction reflects both the contract’s scale and the possibility that Akamai can compete for AI workloads alongside AWS, Microsoft Azure, Google Cloud and newer infrastructure specialists such as Nscale.

Still, execution is the harder part. Akamai will have to procure equipment, manage energy and memory constraints, deploy capacity on schedule and maintain performance across a distributed environment. Kubernetes-based orchestration could support workload portability, although moving large AI systems between providers remains more complicated than shifting conventional applications.

For enterprise technology leaders, the deal broadens the infrastructure conversation. AI capacity planning is no longer only about securing premium GPUs from the largest hyperscalers. CPU fleets, memory, regional availability, networking and operational governance can carry comparable weight once AI systems reach production. Akamai’s agreement with Anthropic puts a very large number on that quieter part of the stack, and it gives Akamai a credible opportunity to prove that distributed cloud infrastructure has a substantial role in the next phase of AI deployment.