Anthropic Signs $11.6 B Cloud Deal with Akamai, Marking Biggest Contract in Akamai History

Anthropic will spend $11.6 billion over seven years on Akamai cloud infrastructure, a deal that could grow to $20 billion and reshape AI financing.

abstract glowing orb with rings and wave nodes
AI-generated illustration
On this page
  1. What happened
  2. Why it matters
  3. The bigger picture
  4. What happens next

Anthropic announced a multi‑year agreement to purchase $11.6 billion worth of cloud infrastructure from Akamai, the largest contract in Akamai’s history. The deal, which hinges on service‑level targets and includes a warrant that could lift the total value to about $20 billion, underscores the growing demand for general‑purpose CPUs to power AI agents.

What happened

  • Anthropic committed to spend $11.6 billion over seven years on Akamai’s cloud services.
  • The agreement is conditional: Akamai must meet delivery and availability benchmarks, and either party may terminate under defined circumstances.
  • Akamai expects to recognize $150 million to $300 million of revenue in 2027, with annual revenue reaching roughly $1.7 billion by the end of 2028.
  • To support the new capacity, Akamai plans to invest about $5.5 billion in infrastructure and add $1.7 billion to this year’s capital spending for components such as memory.
  • The contract includes a warrant giving Anthropic the right to buy non‑voting preferred stock convertible into 7.7 million common shares—about 5 % of Akamai’s outstanding equity—at $111.33 per share.
  • An initial 2 % of the warrant vests after the first payment; each additional $3 billion spent unlocks roughly another 1 % of equity, allowing the total deal size to rise to $20 billion if Anthropic’s spend expands.

Why it matters

The size of the commitment marks a dramatic escalation from the $1.8 billion deal reported in May, showing Anthropic’s aggressive “compute gobbling” strategy.

  • CPU focus: While much AI infrastructure attention centers on GPUs and specialized accelerators, the contract highlights a less‑hyped segment—CPUs, the versatile chips that run everyday code and web traffic.
  • Financial structure: Instead of the usual model where cloud providers or chipmakers receive equity stakes in AI labs, Akamai is granting Anthropic a potential equity position that grows with spend. This reverses the typical supplier‑to‑customer investment flow and mirrors a similar warrant‑based arrangement AMD used with OpenAI.
  • Revenue outlook for Akamai: The deal will not affect Akamai’s top line this year, but the projected revenue stream could reshape its earnings profile, moving the company further into AI‑related services.

The bigger picture

Anthropic’s financing pattern reflects a broader trend where AI developers secure long‑term capacity from multiple cloud and hardware partners.

  • Companies such as Amazon, Google, Microsoft and AMD have previously invested in Anthropic while supplying chips or cloud resources, creating a web of reciprocal stakes across the industry.
  • Akamai’s entry into AI‑focused cloud services, coupled with a sizable capital outlay, signals that traditional content‑delivery networks are positioning themselves to capture a slice of the AI infrastructure market.
  • The warrant mechanism may become a template for future deals, offering suppliers a way to align incentives with customers’ spending without direct equity purchases.

What happens next

Akamai’s executives indicated that revenue from the agreement should begin flowing in the second half of 2027, scaling to an annual pace of about $1.7 billion by 2028.

  • The company will allocate roughly $5.5 billion to expand its cloud capacity and an additional $1.7 billion for component procurement, positioning it to meet Anthropic’s performance requirements.
  • As Anthropic reaches spending milestones, the warrant could vest further, potentially increasing Akamai’s dilution but also deepening the financial link between the two firms.
  • Both parties retain the right to terminate the contract under specified conditions, leaving the ultimate duration and total value open to future negotiation.

The agreement illustrates how AI developers are locking in massive compute resources while cloud providers experiment with novel financing structures. Whether the warrant model gains broader adoption will depend on how closely the parties meet their performance targets and on the evolving economics of AI workloads.