Anthropic Commits $11.6 Billion to Akamai Cloud Infrastructure in Record Deal

Artificial intelligence lab Anthropic has agreed to spend $11.6 billion over seven years on cloud infrastructure provided by Akamai Technologies. Disclosed in regulatory filings, the agreement represents the largest contract in Akamai's corporate history, surpassing a previous $1.8 billion arrangement between the two firms reported earlier in the year. The commitment is designed to support Anthropic's expanding central processing unit computing workloads across Akamai's distributed cloud network.

The massive agreement features a unique equity structure that reverses conventional tech funding models. Rather than an infrastructure vendor investing capital into an AI developer, Akamai issued Anthropic a warrant for nonvoting Series B preferred stock equivalent to 7.7 million common shares, representing roughly 5% of Akamai's outstanding common stock at an exercise price of $111.33 per share.

The warrant mechanism ties Akamai's equity distribution directly to Anthropic's ongoing spending milestones. Approximately 2% of Akamai's outstanding common stock will vest when the initial payment under the $11.6 billion commitment is executed. The remaining 3% is tied to potential agreement expansions, with every additional $3 billion committed to cloud services unlocking roughly 1% of the equity stake.

This tiered expansion could push the total contract value to approximately $20 billion if fully realized. According to securities filings, the underlying commitment is not entirely unconditional and depends on Akamai meeting specific delivery and service availability metrics, while either party retains termination rights under defined conditions.

To fulfill the capacity requirements, Akamai anticipates incurring approximately $5.5 billion in total capital expenditures over the life of the agreement. The company is immediately adding $1.7 billion to its current year capital spending budget to pre-purchase critical supply chain components, specifically memory hardware, through manufacturers such as Lenovo and Jabil.

Akamai executive leadership told investors that revenue generation from the contract will begin in the second half of 2027, yielding an estimated $150 million to $300 million that year. The company projects that revenue under the agreement will reach an annual run rate of approximately $1.7 billion by the end of 2028.

Wall Street financial institutions responded aggressively to the disclosure, sending Akamai shares surging between 17% and 21% in trading sessions following the announcement. Major investment firms revised their outlooks for the company, with Piper Sandler raising its price target from $125 to $158 and characterizing the transaction as a potential catalyst shifting Akamai from a value asset into a hypergrowth profile.

BofA Securities raised its price target for Akamai to $185, while UBS adjusted its target to $148 alongside a Neutral rating. Additionally, the master services agreement allows Anthropic to terminate individual project plans under specific conditions following a material outage.

Analysts noted that the transaction highlights intense competition for non-GPU computing assets. While much of the artificial intelligence boom has concentrated on specialized graphics processing units, Anthropic's commitment emphasizes growing enterprise demand for general-purpose CPUs capable of executing complex code execution and web navigation tasks required by autonomous AI agents.

The agreement builds on more than $2.8 billion in other multiyear Cloud Infrastructure Services commitments secured by Akamai across its broader enterprise customer base earlier in the year. Financial analysts expect revenue ramp-ups to progress through fiscal 2027 and 2028, with market participants closely monitoring the execution of component pre-purchases and the formal scheduling of Anthropic's first vesting milestone.

Because the arrangement is primarily CPU-based, it delivers enhanced cash gross margins and greater revenue per megawatt compared to GPU workloads. Market participants also noted that newly established sites generally need 60 to 90 days to achieve full revenue capacity, a ramp-up phase that can temporarily affect margins.