Anthropic Locks In $11.6B Cloud Deal With Akamai

Anthropic just signed one of the largest cloud infrastructure deals in tech history. On September 25, 2026, Akamai Technologies announced an $11.6 billion, seven-year agreement to supply CPU computing capacity to Anthropic — the company behind Claude. The deal can expand to $20 billion in total commitment if Anthropic opts into additional spend, making it potentially the biggest single infrastructure partnership in AI.

This is not the first Anthropic-Akamai deal. Back in May 2026, the two companies signed a $1.8 billion agreement that sent Akamai’s stock up more than 20 percent in a single day. This new agreement is more than six times larger. The scale of the expansion says something about how fast Anthropic’s compute requirements are growing, and about how seriously enterprise AI adoption has accelerated in just four months.

What the Deal Actually Covers

The agreement focuses on CPU capacity — not the GPU clusters most people associate with AI training. That distinction matters. Akamai’s distributed infrastructure will handle the production-side workloads: inference, retrieval, orchestration, and the long-running agent tasks that businesses are deploying at scale. GPU-heavy training runs are expensive and infrequent. CPU-intensive inference happens constantly, every time a Claude deployment answers a question, completes a workflow, or processes a document.

Akamai will invest approximately $5.5 billion in capital expenditure to fulfil the contract and has already flagged a $1.7 billion increase in 2026 capex to secure supply chain components including memory. That is real infrastructure being built specifically to run Anthropic’s models.

As part of the deal, Akamai issued Anthropic a warrant to purchase non-voting convertible Series B preferred stock representing up to 7.7 million shares of Akamai common stock — roughly 5 percent of shares outstanding — at an exercise price of $111.33 per share. Approximately 2 percent vests on the first payment, with additional tranches unlocking as Anthropic spends more on the platform.

Why This Matters Beyond the Dollar Number

Enterprise AI deployment is no longer a GPU story. The expensive training phase happens once or occasionally. But once a model is deployed — answering customer questions, generating reports, running voice AI workflows, processing invoices — it runs on CPU-heavy infrastructure continuously. Companies like Akamai are positioning to own that layer.

For Anthropic specifically, this deal is about capacity certainty. Signing a seven-year, $11.6 billion agreement is a bet that Claude will be running in production for hundreds of thousands of enterprise customers for the foreseeable future. It is infrastructure for a world where AI agents are embedded into business operations, not occasionally used as tools.

That shift from experimentation to embedded operation is already visible in the data. In August 2026, Salesforce reported that its Agentforce platform crossed $1.2 billion in annual recurring revenue. Microsoft’s Copilot has been installed in well over 700,000 enterprise seats. The inference demand those deployments generate is enormous.

What This Means for Business

AI infrastructure costs are going to keep compressing. When Anthropic locks in a seven-year commitment of this size, it is negotiating wholesale compute rates that feed through to API pricing. Businesses accessing Claude through Anthropic’s API benefit from the scale economics of deals like this one, even if they never see the contract.

CPU infrastructure is the unsexy bottleneck. Most AI press covers GPUs, model releases, and benchmarks. The actual production constraint for large-scale enterprise AI is memory bandwidth, network throughput, and the CPU capacity to run inference continuously. Akamai’s deal is a signal that smart infrastructure companies have identified this gap and are moving to fill it.

Vendor concentration is building fast. A handful of agreements — Microsoft’s OpenAI tie-up, Google’s Anthropic investment, and now this Akamai deal — are shaping which infrastructure platforms AI runs on at scale. Businesses evaluating their own AI vendor choices should watch how these partnerships evolve, because they will influence latency, pricing, and availability for years.

Seven-year contracts reflect genuine conviction. When a company signs a deal with a seven-year initial term — in an industry that moves quarterly — it is expressing real confidence in long-term demand. The fact that this is Akamai’s largest contract in history, and that it followed a $1.8 billion deal from just four months earlier, shows how dramatically the trajectory of AI enterprise adoption has shifted in 2026.

For businesses trying to understand where to place their own AI bets, the movement of capital at this scale is one of the clearest signals available. The infrastructure is being built for a sustained era of AI-embedded business operations. The question is no longer whether AI will be central to how companies run — it is whether individual businesses will be positioned to take advantage of it.

Enterprise DNA helps businesses move from AI experimentation to embedded, operational AI. If you want to understand how AI agents can be deployed in your business — not as demos, but as working infrastructure — start the conversation with our team.