Akamai signed a seven-year, $11.6 billion cloud computing agreement with Anthropic to supply CPU-based distributed infrastructure for AI workloads, with an option to expand to roughly $20 billion and a warrant granting Anthropic up to 5% equity in Akamai.

The $11.6 billion topline figure made the headlines, but the more unusual piece of the Akamai-Anthropic agreement is the equity warrant attached to it - the first time Akamai has ever put a warrant on a cloud contract ^[1]. The warrant covers nonvoting preferred stock convertible into 7.7 million common shares, or up to about 5% of Akamai's outstanding stock, struck at $111.33 per share; roughly 2% vests once Anthropic makes its first payment, with another 1% unlocking for every additional $3 billion Anthropic commits, up to the full 5% ^[2]. Akamai CEO Tom Leighton called it 'a serious step, but I think in this case it made sense to do,' framing the dilution as the price of locking in the revenue ^[3]; in a separate interview he added that the arrangement should help Akamai's margins over time. That structure effectively lets Anthropic buy into the infrastructure it depends on rather than simply renting it, aligning both companies' incentives around Anthropic actually spending the full $11.6 billion rather than walking away early - a customer-financed-capex playbook that is becoming a recognizable pattern across large AI infrastructure deals.

Unlike CoreWeave's Nvidia GPU fleets or the hyperscalers' AI accelerators, Akamai's contribution is CPU-based distributed infrastructure - the company frames the deal as meeting Anthropic's growing CPU compute needs through Akamai Cloud's distributed AI infrastructure and software, not through GPU racks ^[4]. That is a distinct lane from Anthropic's other major compute commitments: more than $100 billion committed to AWS Trainium capacity with up to five gigawatts through Project Rainier, plus a multi-gigawatt TPU expansion with Google and Broadcom ^[5]. Oppenheimer read the deal as validation that Akamai's distributed cloud model - built originally for content delivery, not AI training - can still win frontier AI workloads ^[6], effectively repositioning Akamai as a specialized compute layer rather than a head-on competitor to the GPU-cloud pack. Developers discussing the deal online made a similar point from the infrastructure side: CDN operators like Akamai run large CPU fleets that are typically bottlenecked by network capacity rather than compute, leaving spare processing power that a CPU-workload deal like this one is well suited to absorb - the same underlying logic that led Cloudflare to build its Workers platform on top of its own edge network.

Nine sell-side firms raised price targets after the announcement. Piper Sandler moved to Overweight with a $158 target, up from $125, arguing the deal 'drastically changes the financial profile' of the company and could push Cloud Infrastructure Services past Security as Akamai's largest business by late 2028 ^[7]. Guggenheim went furthest, raising its target to $225 ^[6], while Morgan Stanley projected 2027 revenue growth accelerating to around 16% on the back of the bookings momentum ^[8]. The one dissenting note comes from JPMorgan, which held a Neutral rating despite raising its target to $167, pointing out that Anthropic alone now accounts for roughly 93% of the $14.4 billion in new deals Akamai has signed so far in 2026 ^[8]. Akamai's own securities filing adds a further caveat: the $11.6 billion is a commitment, not guaranteed revenue, and depends on Akamai meeting delivery and service-availability requirements ^[1]. The stock market's own reaction underlines how company-specific this repricing was: Akamai's shares jumped roughly 15% while the broader cloud sector, and even direct AI-infrastructure peers, barely moved.

The relationship between the two companies is only four months old in its current form: Bloomberg first reported a $1.8 billion, seven-year Akamai computing deal in May 2026, without initially naming Anthropic as the counterparty ^[9]. The new agreement is more than six times that size ^[10], with an option that could push the total toward $20 billion. It sits inside a much larger pattern of Anthropic diversifying its compute supply chain beyond any single vendor - alongside the AWS and Google/Broadcom commitments, Anthropic also signed a multi-year GPU deal with CoreWeave in April 2026 to run Claude workloads at production scale ^[11]. Akamai itself is spending roughly $5.5 billion in capex to build out the promised capacity ^[4], with revenue not expected to show up until the second half of 2027 and ramping toward a $1.7 billion annual run rate by the end of 2028 ^[1]- meaning the capital outlay comes well before Anthropic's payments do.