US-based Akamai Technologies has entered into a seven-year, \$11.6 billion cloud services agreement with generative AI developer Anthropic. The deal comes amid expanding demand for distributed infrastructure capable of powering rapidly growing AI inference workloads and CPU tasks.
- Key Takeaways: Akamai and Anthropic’s $11.6 Billion Landmark Deal
- Deal Structure and Warrants: Mechanisms of Equity Granting and Tiered Vesting
- CPU Workload Surge: The Shift from GPU Training to Distributed Inference
- Supply Chain & Capital Expenditure: $5.5B Investment and Partnerships with Lenovo and Jabil
- Multi-Cloud Strategy & Vendor Risk: Anthropic’s Five-Provider Infrastructure
- Future Outlook & Key Watchpoints: Scaling AI Infrastructure and Market Challenges
Key Takeaways: Akamai and Anthropic’s $11.6 Billion Landmark Deal
On September 24, 2026, U.S. cloud and content delivery giant Akamai Technologies announced a seven-year, \$11.6 billion cloud services agreement with leading generative AI developer Anthropic. Under the agreement, Akamai will supply its distributed AI cloud infrastructure and software to support Anthropic’s rapidly expanding CPU workload processing needs. Financial markets reacted with overwhelming optimism, sending Akamai’s shares surging 16% to 22% in after-hours trading. The relationship originally came to light in May 2026, when Akamai disclosed a \$1.8 billion, seven-year contract with an undisclosed “frontier model developer.” This latest announcement reveals that the scale of the partnership has expanded exponentially in just a few months. It represents a historic infrastructure transaction symbolizing the surging global demand for large-scale AI capabilities. Key components of the deal and the investment overview are summarized in the figure below.

Deal Structure and Warrants: Mechanisms of Equity Granting and Tiered Vesting
This agreement goes beyond a standard cloud service contract, incorporating a structure designed to deepen the capital and strategic alignment between the two companies. In addition to the baseline \$11.6 billion commitment, mutual extension options allow for up to \$9.0 billion in additional contracts, bringing the potential total contract value to approximately \$20.0 billion. In connection with the transaction, Akamai issued warrants to Anthropic to purchase Series B Non-Voting Convertible Preferred Stock. These warrants represent approximately 7.7 million shares of common stock (equivalent to roughly 5% of Akamai’s outstanding common stock) at an exercise price of \$111.33 per share. Vesting occurs in tranches: approximately 2% vests in connection with the initial \$11.6 billion contract, with the remaining 3% vesting in ~1% increments for every additional \$3.0 billion in spend commitments. If fully exercised, the warrants would represent an equity investment of approximately \$857 million in Akamai.
Technical Background and Physical Infrastructure Procurement
CPU Workload Surge: The Shift from GPU Training to Distributed Inference
Demand for computational resources across the AI industry is broadening from traditional GPU-centric model training phases toward inference processing and workload execution leveraging CPUs and edge networks. Founded in 1998 by MIT Professor Tom Leighton and Danny Lewin, Akamai has long optimized internet traffic as a pioneer in Content Delivery Networks (CDNs). In March 2022, the company acquired cloud provider Linode for approximately \$900 million, fusing edge infrastructure with cloud computing capabilities. Akamai’s distributed cloud platform features thousands of Points of Presence (POPs) worldwide, providing an environment built for low-latency, resilient execution of advanced AI applications. The network architecture of this distributed AI inference infrastructure is illustrated below.

Supply Chain & Capital Expenditure: $5.5B Investment and Partnerships with Lenovo and Jabil
To physically support this massive \$11.6 billion computing demand, Akamai has launched extensive infrastructure builds and supply chain procurements. Total capital expenditures (Capex) associated with the contract are estimated at approximately \$5.5 billion. While maintaining its full-year 2026 revenue guidance, Akamai announced an additional ~\$1.7 billion in Capex during 2026 to secure critical components, such as memory, in advance. On the physical server and hardware front, Akamai entered into a seven-year Master Products and Services Agreement (Lenovo MPSA/SOW) on September 23, 2026. Furthermore, on September 24, Akamai issued a Build Request to electronics manufacturing services (EMS) giant Jabil to procure \$1.7 billion worth of memory components, ramping up hardware readiness at the physical layer.
AI Industry Impact and Future Outlook
Multi-Cloud Strategy & Vendor Risk: Anthropic’s Five-Provider Infrastructure
Anthropic experienced explosive growth in the first quarter of 2026, with usage and revenue surging 80-fold on an annualized basis, prompting a shift toward a multi-provider strategy rather than relying on a single cloud vendor. Currently, the upstream infrastructure supporting Anthropic’s Claude models spans five providers: Amazon AWS, Google Cloud, Microsoft Azure, Akamai, and xAI’s “Colossus 1” supercomputer located in Memphis, Tennessee. With computing resources distributed across multiple clouds and data centers, third-party vendors offering legal AI and enterprise software face new risk management and governance imperatives, including data residency compliance, service-level agreements (SLAs), and failover routing control.
Future Outlook & Key Watchpoints: Scaling AI Infrastructure and Market Challenges
This landmark agreement marks a decisive turning point for Akamai as it transforms from a traditional CDN provider into a trusted AI cloud infrastructure provider. Moving forward, key watchpoints include the pace of Anthropic’s cloud consumption relative to the baseline \$11.6 billion commitment toward the \$20 billion maximum ceiling, as well as Akamai’s execution in completing its \$5.5 billion buildout and server deployment on schedule. In addition, industry observers will closely monitor U.S. regulatory developments concerning export controls on frontier AI models, alongside global supply chain constraints in memory and semiconductors that could influence both companies’ growth trajectory and deployment speed.
Reference Pages
- 【8-K – SEC.gov】https://www.sec.gov/Archives/edgar/data/1086222/000119312526401048/d288154dex991.htm
- 【Akamai Press Release】https://www.akamai.com/newsroom/press-release/akamai-announces-11-6-billion-multi-year-agreement-with-anthropic-to-support-growing-demand
【#Akamai #Anthropic #CloudAI #AIInfrastructure #CPUInference #SupplyChain #USStocks #Warrants】


コメント