AI’s Infrastructure Bill Comes Due

The biggest infrastructure story comes from Anthropic and Akamai.

Akamai says, Anthropic has committed about $11.6 billion over seven years for dedicated cloud capacity and managed support, focused on CPU workloads.

And Akamai Filing says the payments depend on delivery and service availability requirements.

Akamai also estimates about $5.5 billion of related capital spending.

The strategic signal is bigger than a vendor contract.

Frontier Labs are spreading compute needs across more suppliers, and Akamai is positioning its distributed cloud as part of the AI infrastructure stack.

Microsoft introduced a new co-pilot organized around home, code, and autopilot.

Home combines chat and longer-running co-work tasks.

Code lets users describe internal apps, dashboards, and workflows in natural language.

Autopilot is designed as a persistent cloud-hosted agent that can run recurring work.

Microsoft says Home and Code will begin rolling out through its Frontier program in the coming weeks, while Autopilot expands to Private Preview at the end of September.

Long-running agent work and frontier models will use usage-based billing.

Microsoft has not published the usage prices.

Lambda announced plans Sunday for a new data center in Mays County, Oklahoma.

The company estimates the project could generate $500 million in taxes over 10 years, create up to 1,000 construction jobs, and support up to 100 permanent positions.

Lambda says it will pay the facility's energy costs and use closed-loop cooling.

Those are company projections and commitments, not completed outcomes.

The announcement shows how power costs, water use, and community benefits are becoming part of the sales pitch for AI infrastructure.