Akamai Technologies shares jumped as much as 20% in after-hours trading Thursday after the cloud computing company announced an $11.6 billion, seven-year agreement with artificial intelligence firm Anthropic.
The contract will support Anthropic’s CPU computing requirements using Akamai Cloud’s distributed AI infrastructure and software. The massive commitment builds on more than $2.8 billion in multi-year Cloud Infrastructure Services agreements Akamai announced earlier this year, strengthening the company’s position in the rapidly expanding AI infrastructure market.
As part of the Anthropic deal, Akamai issued a warrant allowing the AI company to purchase non-voting convertible Series B Preferred Stock. The warrant represents 7.7 million Akamai common shares on an as-converted basis, equivalent to as much as approximately 5% of the company’s outstanding common stock. The exercise price is $111.33 per share.
Approximately 2% of Akamai’s outstanding common stock is expected to vest based on Anthropic’s initial $11.6 billion commitment. Another roughly 3% could vest if the agreement expands by as much as an additional $9 billion during the seven-year warrant period. Each additional $3 billion in cloud services purchases would trigger vesting equivalent to approximately 1% of Akamai’s outstanding shares.
“Anthropic is advancing the AI revolution and we are thrilled they chose Akamai’s capabilities for building and operating AI infrastructure at scale,” Akamai co-founder and CEO Tom Leighton said.
Akamai estimates total capital expenditures associated with the $11.6 billion contract at approximately $5.5 billion. The company expects the agreement to have no impact on its 2026 revenue guidance.
However, Akamai expects 2026 capital expenditures to increase by approximately $1.7 billion as it moves to secure and pre-purchase critical supply chain components, including memory.
The Anthropic agreement highlights growing demand for computing capacity as AI companies expand infrastructure to support increasingly resource-intensive models and workloads.
