Akamai has signed a seven-year, $11.6 billion cloud infrastructure agreement with Anthropic that puts distributed CPU capacity at the center of one of the largest AI infrastructure commitments announced this month. In the company’s announcement, Akamai said the agreement will support Anthropic’s growing CPU workload requirements across its distributed cloud platform.
A CPU-Heavy AI Infrastructure Deal
The agreement is notable because the workload emphasis is not limited to GPUs. Akamai specifically says Anthropic will use its infrastructure for accelerating CPU workloads at scale, showing how general-purpose compute remains essential around model serving, orchestration, preprocessing, agents, networking and the wider software stack that surrounds accelerator clusters.
That distributed model overlaps with a broader trend BitcoinVersus.tech has been tracking. Civo’s planned 40-site UK AI network and Spectrum’s 1,000-site distributed AI strategy both show compute moving outward from a small number of centralized hyperscale campuses.
The Commitment Could Approach $20 Billion
Akamai says the initial $11.6 billion commitment can expand by as much as another $9 billion over the seven-year term, taking the potential relationship to roughly $20 billion. The company also issued Anthropic warrants tied to the expansion of the relationship, including an initial portion connected to the current commitment.
The scale fits into Anthropic’s much larger infrastructure buildout. BitcoinVersus.tech recently reported on Anthropic’s broader AI infrastructure commitments, which increasingly span multiple cloud providers and hardware architectures rather than a single compute partner.
Akamai Plans Billions in Supporting Capex
Serving the contract will require major physical investment. Akamai estimates approximately $5.5 billion in capital expenditures related to the $11.6 billion commitment and expects about $1.7 billion of additional 2026 spending to secure and pre-purchase critical supply-chain components, including memory.
AI Infrastructure Is Becoming More Distributed
The deal reinforces a change in how AI infrastructure is being assembled. Training still rewards massive centralized clusters, but inference and agent workloads increasingly benefit from compute placed closer to users, APIs and regional data. Akamai’s existing content-delivery footprint gives it a different physical starting point than a traditional hyperscaler.
That does not eliminate the need for giant GPU campuses. Instead, the emerging architecture looks layered: hyperscale accelerator clusters for the heaviest model work, regional compute for serving and orchestration, and edge capacity for latency-sensitive applications. Anthropic’s agreement with Akamai is another sign that the AI infrastructure race is expanding from GPUs and megawatts into geography, CPUs, memory and network placement.
BitcoinVersus.Tech Editor’s Note:
We volunteer daily to ensure the credibility of the information on this platform is Verifiably True. If you would like to support to help further secure the integrity of our research initiatives, please donate here: 3C9o19EH5HSiwEPyCTmEKzxhNCbo2X6TTb
BitcoinVersus.tech is not a financial advisor. This media platform reports on financial subjects purely for informational purposes.
Leave a comment