Anthropic Signs $11.6B Akamai Cloud Deal — Shares Pop 20% as AI Infrastructure Race Heats Up

Rows of blue-lit server racks inside a modern data center

Anthropic just signed the biggest check in Akamai’s history — and Wall Street noticed.

The Biggest Check Akamai Has Ever Cashed

On Thursday, Akamai Technologies announced a seven-year, $11.6 billion cloud services agreement with Anthropic, locking in the AI lab as one of the largest compute customers on the planet. According to Reuters, Akamai shares surged as much as 22% in extended trading after the announcement — a monster reaction for a company most people still think of as a content-delivery network.

The deal isn’t for GPUs, either. It’s for CPU capacity. Anthropic will run its growing CPU workloads across Akamai’s distributed cloud, which spans thousands of points of presence worldwide. With agents running longer, multi-step tasks and coding workloads exploding, plain old CPUs are suddenly the hottest commodity in AI infrastructure.

The Warrant That Could Make Anthropic a 5% Owner

Here’s the part that turns a big contract into something much more interesting. Akamai issued Anthropic a warrant to buy roughly 7.7 million shares of non-voting convertible Series B preferred stock at $111.33 a share — potentially about 5% of Akamai’s common stock outstanding.

About 2% of that vests with this initial commitment. The rest unlocks if the deal expands by up to another $9 billion, with roughly 1% vesting for every additional $3 billion Anthropic spends. The Wall Street Journal reports the total potential commitment could reach around $20 billion. And because the shares are non-voting, Anthropic gets the financial upside of Akamai’s rise without touching the boardroom.

Why CPUs? Because AI’s Hunger Isn’t Just GPUs Anymore

This deal is a flashing sign that the AI infrastructure story has a second act. The GPU land grab got the headlines, but agentic AI — models that reason, call tools, and iterate for hours — burns through CPU compute at a staggering rate. Akamai is betting its globally distributed network is the right shape for that workload, running AI applications from core data centers out to the edge.

It’s not cheap for Akamai, either. The company estimates about $5.5 billion in capital spending tied to the agreement, including a $1.7 billion increase in 2026 capex to pre-purchase critical components like memory. Management says there’s no revenue contribution in 2026 and full-year guidance is unchanged — this is a long game.

It’s also a pattern. This agreement builds on a $1.8 billion Anthropic–Akamai commitment from May, and it lands just weeks after Anthropic agreed to spend a reported $45 billion renting AI cloud capacity from Nscale’s West Virginia campus. The labs are buying compute like it might run out.

Why This Matters

The AI arms race has a second front, and it’s infrastructure. OpenAI and Anthropic spent the week slashing model prices — yesterday’s story — while quietly committing tens of billions to the compute underneath. Akamai, the company that spent two decades making websites load faster, is now positioning itself as a premier infrastructure provider for secure AI applications, in CEO Tom Leighton’s words.

For everyone else, the message is simple: the companies selling shovels in this gold rush are getting paid in billions, and sometimes in equity. As eMarketer analyst Jacob Bourne put it, infrastructure investment will keep expanding to support increasingly compute-intensive agentic workloads — even amid growing concerns about agentic AI itself. The potential stake, he noted, “can be seen as a vote of confidence in the durability of AI-driven cloud demand.”

Meanwhile in AI Tech: OpenAI Slashes AI Costs in Half With GPT-6 Sol and Luna — Hours After Anthropic’s Counterpunch | DeepSeek Secures NVIDIA H200 Deal: New Chip Strategy Exposed

Written by
Ryan covers artificial intelligence and enterprise tech — from foundation models and AI chips to the business of machine intelligence. He tracks model releases, funding rounds, and the policy moves shaping the AI industry.