Global data center spending is expected to reach $1 trillion in 2026 and grow to $2.1 trillion by 2028, according to Bank of America Global Research. The market could expand further to $2.93 trillion by 2030 as companies invest in the computing infrastructure required to run artificial intelligence services, cloud platforms and increasingly data-intensive applications.
The forecast highlights the scale of the AI infrastructure boom. Building modern data centers requires substantial investment in servers, networking equipment, electricity infrastructure, cooling systems and physical facilities. While chipmakers and server manufacturers have already benefited from rising AI demand, the projected expansion could also create opportunities for industrial equipment suppliers, construction companies and businesses involved in power management.
Key Takeaways
- $2.93 trillion by 2030: Bank of America estimates that the global data center market could reach this value.
- $1 trillion in 2026: Global data center spending is expected to reach this milestone as AI infrastructure investment accelerates.
- $2.1 trillion by 2028: Spending could more than double from its projected 2026 level.
- 37% annual growth: Bank of America estimates a compound annual growth rate (CAGR) of 37% between 2025 and 2030.
- Servers dominate costs: Servers are estimated to cost $31.4 million per megawatt of data center capacity.
- AI infrastructure expands opportunities: Power generators, switchgear, networking equipment and cooling systems could benefit from continued construction.
How Big Could the Data Center Market Become?
The latest Bank of America estimates suggest that data center infrastructure is entering a period of unusually rapid expansion. The market was valued at approximately $275 billion in 2022, according to the brokerage, before the acceleration of generative AI investment pushed computing infrastructure higher on corporate spending priorities.
| Year | Estimated global data center spending or market value |
|---|---|
| 2022 | $275 billion |
| 2026 | $1 trillion |
| 2028 | $2.1 trillion |
| 2030 | $2.93 trillion |
<text color=”secondary” size=”sm”>Source: Bank of America Global Research estimates reported by Reuters. These figures are forecasts, not confirmed future spending.</text>
The projected 2030 value represents a substantial expansion from the 2022 market estimate. Bank of America expects the market to grow at a 37% CAGR from 2025 through 2030, reflecting the rising importance of AI services and the infrastructure needed to support them.
The figures should be interpreted as market estimates rather than guaranteed outcomes. Actual investment will depend on demand for AI computing, the economics of data center operations, electricity availability and how efficiently companies use their existing infrastructure.
Why Artificial Intelligence Is Driving Data Center Investment
AI models require significant computing resources for training, inference and serving users. As companies introduce AI assistants, enterprise automation tools and generative applications, infrastructure providers must expand capacity to handle larger workloads.
Unlike many conventional software services, advanced AI workloads can require large clusters of specialized processors working together. These systems also generate substantial heat and need reliable electricity, fast networking and sophisticated cooling.
Several factors are driving the spending cycle:
- Generative AI adoption: Businesses are integrating AI into customer service, software development, search and other workflows.
- Cloud computing demand: Organizations continue to use cloud infrastructure to run applications and store data.
- High-performance computing: AI training and inference workloads require specialized servers and networking systems.
- Capacity expansion: Cloud providers and AI-focused infrastructure companies are building additional capacity to meet anticipated demand.
- Power and cooling requirements: Dense computing clusters require upgraded electrical systems and thermal management.
The growth is not limited to the companies operating the data centers. It also creates demand for the equipment that allows them to function reliably.
Servers and Networking Account for Major Costs
Data center construction is capital-intensive, and the equipment inside a facility can account for a substantial portion of total expenditure.
Bank of America estimates that building a data center costs approximately $49 million per megawatt of capacity, with IT-related expenses accounting for around 78% of total spending.
| Cost component | Estimated cost per megawatt |
|---|---|
| Servers | $31.4 million |
| Networking | $4.8 million |
| Total construction cost | $49 million |
| AI-focused neocloud data center | $57 million |
<text color=”secondary” size=”sm”>Source: Bank of America Global Research estimates reported by Reuters. Figures are estimates and may vary by facility design, equipment configuration and project scope.</text>
Servers represent the largest individual expense in the estimates. Networking is another significant cost because AI systems need to transfer large volumes of data between processors and storage systems.
The remaining investment includes power distribution, switchgear, backup generators, cooling equipment and other facility infrastructure. These systems are essential because interruptions in electricity or cooling can disrupt computing workloads and damage equipment.
The estimates also help explain why the AI infrastructure boom is creating opportunities beyond semiconductor manufacturers. Companies supplying electrical equipment, thermal management systems and construction services may benefit as data center capacity expands.
AI-Focused Neoclouds Could Require More Investment
AI-focused data centers, sometimes called neoclouds, are designed primarily to support high-performance computing and AI workloads rather than a broad mix of conventional cloud applications.
Bank of America estimates that these facilities cost approximately $57 million per megawatt, compared with around $49 million for data centers in its broader estimate. Their higher costs reflect the specialized infrastructure needed to support demanding AI workloads.
These facilities may use high-density server racks, specialized networking and advanced cooling technologies. Although their physical footprint can be smaller than that of some traditional facilities, their equipment and power requirements can make them more expensive to build per unit of capacity.
The distinction matters because a megawatt of capacity does not necessarily represent the same computing capability across different facilities. AI-oriented data centers can pack substantially more specialized computing equipment into a given area, increasing both investment requirements and potential computing output.
Which Industries Could Benefit From the Spending Boom?
The projected expansion could create business opportunities across the data center supply chain.
| Industry | Potential benefit |
|---|---|
| Semiconductor manufacturers | Demand for AI accelerators and other computing chips |
| Server manufacturers | Higher orders for AI servers and related systems |
| Networking equipment providers | Demand for high-speed switches and connectivity |
| Electrical equipment manufacturers | Investment in generators, switchgear and power distribution |
| Cooling technology companies | Demand for systems that manage high heat loads |
| Construction and engineering firms | More projects involving specialized facilities |
| Cloud and AI infrastructure providers | Opportunities to sell computing capacity and hosted services |
Chipmakers and server developers have been among the most visible beneficiaries of the AI infrastructure buildout, particularly in the United States. But power and cooling requirements are creating additional demand for industrial suppliers.
The eventual winners will depend on project execution, pricing, competition and the ability of suppliers to meet delivery schedules. Rising market spending does not automatically translate into higher profits for every company involved.
What Could Slow Data Center Expansion?
Despite the optimistic forecast, the sector faces practical constraints.
Electricity availability: Large facilities require reliable power, and connecting new projects to electricity grids can take time.
Construction delays: Permitting, land acquisition, equipment shortages and specialized labor requirements can slow projects.
Rising costs: Higher prices for equipment, construction materials and financing could make some projects less attractive.
AI investment returns: Companies may reassess spending if revenue from AI services fails to justify the cost of building and operating new infrastructure.
Utilization risk: Data centers built ahead of demand could leave operators with expensive capacity that generates insufficient revenue.
These factors mean the path to $2.93 trillion is not guaranteed. Developers and investors will need to balance expected demand against capital costs, operating expenses and the availability of power.
The Bigger Picture
The projected growth in data center spending shows how AI is reshaping investment across the technology industry. The expansion is no longer just about buying more chips or developing larger models. It requires physical infrastructure capable of delivering electricity, cooling, networking and computing capacity at an enormous scale.
For investors and businesses, this broadens the AI opportunity beyond software and semiconductor companies. Industrial equipment suppliers, construction firms and energy infrastructure providers could become increasingly important to the development of AI services. However, the market’s eventual size will depend on whether real demand and economic returns keep pace with the current investment cycle.
FAQs
1. How much could global data center spending reach by 2030?
Bank of America Global Research estimates that the global data center market could reach $2.93 trillion by 2030. This is a forecast, not a guaranteed spending outcome.
2. How much data center spending is expected in 2026?
Bank of America estimates that global data center spending will reach $1 trillion in 2026, driven primarily by growing demand for AI services and infrastructure.
3. Why are AI companies building more data centers?
AI training and inference require significant computing capacity, specialized servers, fast networking and reliable electricity. More AI applications and users are increasing the need for infrastructure.
4. How much does it cost to build a data center?
Bank of America estimates construction costs of around $49 million per megawatt for data centers generally and approximately $57 million per megawatt for AI-focused neocloud facilities. Actual costs vary by design and equipment.
5. Which companies could benefit from data center growth?
Potential beneficiaries include semiconductor manufacturers, server and networking suppliers, power equipment makers, cooling technology providers and construction companies. Their financial performance will depend on orders, margins and execution.
Looking Ahead
The next phase of data center investment will depend on how quickly AI demand translates into revenue and whether infrastructure providers can overcome electricity, cooling and construction constraints. Spending is likely to remain a major focus for technology companies, but the market’s growth path could change if projects are delayed or AI returns fall short of expectations.
For businesses and investors, the key will be identifying where infrastructure demand creates sustainable value. The $2.93 trillion projection signals a potentially enormous market, but long-term success will depend on efficient construction, reliable operations and enough demand to justify the capital being deployed.
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.
