Enterprise AI Agent Spending Drives $3 Trillion Data Center Boom
Morgan Stanley projects global data center construction costs will reach $3 trillion by 2028 as enterprises surge spending on artificial intelligence agents.
Morgan Stanley projects that global data center construction costs will reach nearly $3 trillion by 2028, driven by a surge in enterprise spending on artificial intelligence agents. This massive infrastructure investment is fueling record growth for hardware designers and data center operators who are scaling to meet agentic AI workloads.
Arm Holdings reported record quarterly revenue of $1.49 billion, with management targeting the largest share of data center CPUs by 2030. Similarly, Nvidia is diversifying its revenue streams through the Vera CPUs, which are expected to generate $20 billion this year. The company is also preparing the Vera Rubin platform, specifically designed for the complex requirements of AI agents. Analysts project Nvidia's total revenue will increase 81% this year to $391 billion.
On the infrastructure side, IREN is aggressively expanding its capacity, with 480 megawatts coming online in 2026. The company is planning new sites in Spain and Australia to support this growth, backed by long-term cloud contracts with Microsoft and Nvidia. These expansions highlight a broader industry trend where cloud providers and hardware giants are partnering with infrastructure firms to secure the power and space necessary for the next generation of AI.