Global spending on data centers is entering a period of extraordinary expansion as demand for artificial intelligence services accelerates infrastructure investment. BofA Global Research estimates that spending will reach $1 trillion in 2026 before more than doubling to $2.1 trillion by 2028, highlighting the scale of capital required to support the next generation of AI computing.

AI Becomes the Main Engine of Data Center Expansion

The global data center market, valued at approximately $275 billion in 2022, could reach $2.93 trillion by 2030, according to BofA. The brokerage expects AI to remain the primary force behind that expansion, forecasting a 37% market-wide compound annual growth rate from 2025 through 2030.

The rapid development of generative AI and other computing-intensive services requires increasingly powerful infrastructure. As AI models become more sophisticated, companies need greater computing capacity, networking capabilities and specialized power and cooling systems.

The scale of projected investment demonstrates that AI’s economic impact extends well beyond semiconductor manufacturers and software developers. It is also creating a broad infrastructure cycle involving construction, power generation, cooling and electrical equipment.

Infrastructure Suppliers Become Major AI Beneficiaries

Chipmakers and server developers have been among the biggest beneficiaries of the AI-driven data center buildout, particularly in the United States. However, the spending cycle is also generating significant demand for industrial companies producing generators, construction equipment, switchgear and chillers.

BofA estimates data center construction costs at approximately $49 million per megawatt, with IT-related expenses representing 78% of total spending. That allocation illustrates how much of the investment is concentrated in computing equipment rather than the physical building itself.

Neoclouds, or AI-focused data centers, are more expensive at approximately $57 million per megawatt, although they require less physical space than traditional facilities. Their specialized design reflects the growing importance of high-density computing environments.

Servers and Networking Dominate Infrastructure Costs

Servers represent the largest individual component of data center spending, with BofA estimating costs of approximately $31.4 million per megawatt. Networking follows at $4.8 million per megawatt, underscoring the importance of moving increasingly large volumes of data between computing resources.

Other infrastructure components, including switchgear, power distribution systems, generators, backup power, chillers and cooling equipment, generally cost between $100,000 and $1.1 million per megawatt.

These cost structures indicate that the AI infrastructure boom is creating a layered supply chain in which technology and industrial companies can benefit at different stages of construction and operation.

AI Infrastructure Spending Creates a Long-Term Investment Cycle

The projected increase in data center spending represents a fundamental expansion of the physical infrastructure supporting the digital economy. AI demand is transforming data centers into one of the largest areas of global technology investment, while simultaneously increasing requirements for electricity, cooling and specialized equipment.

Going forward, investors will monitor AI service demand, data center construction rates and the ability of infrastructure suppliers to meet rapidly expanding requirements. The pace of capital deployment will also be important, as the projected scale of spending raises questions about power availability, construction capacity and whether AI-related revenues can ultimately justify the enormous infrastructure investment required.