Akamai stock jumps 27% on $1.8B Anthropic cloud deal in AI pivot

Akamai Technologies saw its stock surge 27% in a single day after disclosing a $1.8 billion, seven-year cloud infrastructure deal with a customer identified by Bloomberg as Anthropic. This marks the largest contract in the company’s 28-year history and its biggest single-day rally ever. For a firm long known for accelerating web content delivery, this agreement represents a definitive pivot into AI infrastructure, validating a strategy years in the making.

The deal anchors a quarter where Akamai’s cloud infrastructure services revenue jumped 40% year-over-year to $95 million, even as its legacy content delivery business declined 7%. Investors are now pricing the company not for its past two decades but for what this single commitment suggests about its future. The central question remains: does this contract signal a true transformation or a dangerous concentration of risk?

The $1.8 billion commitment is the largest in Akamai’s history. Revenue from the deal is expected to start in the fourth quarter of 2026, contributing roughly $20 to $25 million in that period. The seven-year term offers a level of revenue visibility that Akamai’s legacy CDN business, which operates on shorter cycles and faces relentless price compression, has never provided.

This agreement follows a $200 million, four-year cloud services deal Akamai signed in February with another unnamed U. S. tech company, which will use a multi-thousand NVIDIA Blackwell GPU cluster. Together, the two contracts represent $2 billion in committed cloud revenue from customers that didn’t exist for Akamai two years ago.

Anthropic has already secured all of SpaceX’s Colossus 1 data centre capacity, adding over 300 megawatts and more than 220,000 NVIDIA GPUs to its compute footprint. The Akamai deal extends the same logic: Anthropic is buying compute capacity from every available provider as demand for Claude outpaces supply. CEO Dario Amodei said the company experienced “80x growth” in annualised revenue and usage in the first quarter of 2026 and is “working as quickly as possible” to secure more computing resources.

Founded in 1998 at MIT to solve web congestion, Akamai spent two decades operating the world’s largest content delivery network across more than 4,000 locations in 130 countries. That business made the company indispensable but also became commoditised. Under CEO Tom Leighton, who moved from chief scientist to the top role in 2013, the company diversified first into cybersecurity, which now accounts for 55% of revenue at $590 million per quarter, growing 11% year-over-year. The second pivot into cloud computing began with the $900 million acquisition of Linode in 2022 and is now producing the growth investors had been waiting for.

Leighton told CNBC the deal validates Akamai’s “different approach” and noted a “very strong pipeline of major enterprise customers, including some that have very large cloud needs.” At NVIDIA’s GTC event in March, Akamai announced it would deploy thousands of NVIDIA RTX PRO 6000 GPUs and build what it called the “industry’s first global-scale implementation of NVIDIA’s AI Grid,” pushing AI inference closer to end users to reduce latency and cost.

Anthropic’s decision reflects a defining constraint in the current AI infrastructure market: demand for compute exceeds the capacity of any single provider. The company already runs Claude across Google tensor processing units, Amazon’s custom chips, and NVIDIA hardware. It has signed with SpaceX for data centre capacity and is exploring building its own chips. With run-rate revenue surpassing $30 billion, custom silicon remains years away. In the interim, Anthropic is buying capacity wherever it can find it. Akamai’s distributed edge network, originally built for CDN traffic, offers something centralised hyperscale data centres cannot: low-latency inference for real-time enterprise applications.

Nebius acquired Eigen AI for $643 million to optimise inference performance, betting that the most valuable layer in AI infrastructure is not raw compute but efficiency. Akamai’s pitch to Anthropic rests on a similar premise: that distributed inference at the edge is more efficient for certain workloads than centralised processing.

Akamai reported first-quarter revenue of $1.074 billion, up 6% year-over-year. Adjusted earnings per share were $1.61. Cloud infrastructure services revenue hit $95 million, up 40%. Security revenue was $590 million, up 11%. Delivery and other revenue was $389 million, down 7%. The cloud segment represents less than 9% of total revenue. The $1.8 billion deal, at roughly $257 million per year, would more than double the segment’s current annual run rate, transforming cloud from a promising division into the company’s primary growth engine.

For the full year, Akamai forecasts revenue of $4.45 to $4.55 billion and adjusted earnings of $6.40 to $7.15 per share. The guidance does not yet reflect the full impact of the Anthropic contract, which begins in the fourth quarter. Analysts will spend the next two quarters trying to determine whether this deal is a one-off or the first in a series.

Anthropic launched a marketplace for Claude-powered enterprise software and committed $100 million to the Claude Partner Network, signalling ambitions that extend well beyond model development into infrastructure and services. The scale of expansion explains the compute hunger behind the Akamai deal. But a $1.8 billion contract with one customer concentrates risk as much as revenue. Anthropic’s annualised revenue has grown from roughly $900 million in late 2025 to a reported $30 billion run rate. Growth at that pace creates demand for infrastructure but also conditions for a correction if the curve flattens. Akamai’s stock gained 27% on the announcement. Sustaining that valuation depends on whether Anthropic’s growth trajectory holds for seven years.

Leighton says more is coming. Akamai survived the dot-com crash, navigated the commoditisation of its original business, and spent a decade building a cybersecurity franchise before the market rewarded it. The AI cloud deal is the latest reinvention of a company that has been reinventing itself since 1998. The difference this time is that the reinvention depends on one customer’s continued appetite for compute, and on the assumption that demand for AI inference at the edge will grow as fast as the demand for AI itself.