The rapid expansion of artificial intelligence and data center infrastructure has reached a notable milestone in the United States: investment in information-processing equipment, including data centers and computer hardware, has now surpassed residential investment.
As recently reported by Fortune (citing data from the US Bureau of Economic Analysis), private investment in the residential sector in the second quarter of 2026 was approximately $748 billion, down 18% from its early 2021 peak. Over the same period, the investment in information-processing equipment increased by 51%, reaching approximately $752 billion.
This crossover, further supported by the statistics we discuss in this blog post, provides another measure of the extraordinary amount of capital flowing into AI and digital infrastructure and highlights the evolution of data centers from a specialized real estate asset class to an important driver in the US economy.
The increasing scale of data center investment, along with the convergence of real estate, power, finance, technology, and infrastructure reflected in the statistics above, underscores the importance of considering these issues together as projects move from site selection and financing through construction and operation.
Data Center Construction Continues to Accelerate
The investment shift is particularly visible in construction spending.
According to a recent KPMG analysis of US construction spending, data center construction spending reached a record annualized rate of approximately $75.2 billion in July 2026, approximately 57% higher than the previous year. By comparison, overall US construction spending declined 3.8% year over year during the same period.
Despite record construction, new supply continues to be absorbed rapidly. According to CBRE’s North America Data Center Trends H1 2026, approximately 7,481 MW of data center capacity was under construction across the eight primary North American markets during the first half of 2026, a 24.8% increase from the prior year and a record level. More than 80% of that capacity was already pre-leased. This leaves less than 1,500 MW available—approximately six months of supply at the current pace of demand.
CBRE also reports that primary market supply increased 33.7% year over year to approximately 10,903 MW, while vacancy nevertheless declined to a record low of 1.4%.
Northern Virginia illustrates the supply-demand dynamic particularly well. According to CBRE, the market reached approximately 4,496.5 MW of inventory during the first half of 2026, yet vacancy fell to just 0.2%, with approximately 2,420 MW of additional capacity under construction.
AI Investment Extends Beyond the Building
Those figures capture only part of the investment required to bring new capacity online. The physical building represents one component of a data center project. Servers and accelerators, networking infrastructure, cooling systems, electrical equipment, and the generation and transmission infrastructure necessary to support increasingly dense computing environments can add substantially to overall capital requirements.
Dell’Oro Group reported in September that worldwide data center capital expenditures increased 92% year over year during the second quarter of 2026. Dell’Oro attributed the acceleration to continued AI infrastructure investment across compute, storage, networking, and physical infrastructure as well as increases in server costs associated with higher memory and storage prices.
The scale of these investments is also influencing how data center projects are structured and financed. As projects become larger and more capital intensive, developers, technology companies, lenders, and investors are evaluating combinations of traditional debt and equity financing, leases, joint ventures, special-purpose vehicles, guarantees, and other structures.
These arrangements can raise complex questions regarding the allocation of construction, financing, technology, equipment, credit, power, and residual-value risks. As a result, the growth of AI infrastructure is increasingly connecting areas that historically may have been viewed separately as real estate, technology, energy, or finance matters.
Power Remains a Critical Constraint
Capital and construction capacity alone will not determine how quickly the next generation of data centers can be delivered. Access to sufficient power, and the ability to obtain it within project development timelines, continues to be a central consideration.
A September analysis from Axios based on data from Cleanview reports that an average new data center operating at approximately two-thirds capacity can use enough electricity to power approximately 54,000 homes. The analysis also estimates that the next generation of US data centers is on track, on average, to consume approximately 2.5 times as much power as facilities opening in 2026.
These trends reinforce issues we have previously discussed in our Data Center Bytes blog. Data center developers and operators are considering an expanding range of energy strategies, including power purchase agreements, specialized utility tariffs, negotiated service agreements, behind-the-meter generation, microgrids, dedicated generation, nuclear power, and “bring your own power” approaches.
Grid interconnection delays, regulatory approvals (including moratoria), cost allocation, transmission availability, and project development timelines can materially affect the economics and feasibility of proposed facilities.
From Real Estate Asset Class to Economic Driver
The comparison between computing investment and housing therefore represents more than an interesting statistical crossover.
A rapid increase in data center construction during an overall construction spending decline, coupled with the increased power requirements associated with the next generation of facilities, demonstrates that data centers have evolved beyond a specialized real estate asset class. They increasingly sit at the intersection of real estate, energy, technology, finance, construction, telecommunications, and infrastructure.
For developers, operators, hyperscalers, investors, lenders, utilities, and other participants in the data center ecosystem, that convergence means that decisions made in one area of a project can have significant consequences elsewhere. For example:
- Site selection may depend on available power.
- Power strategy may affect regulatory approvals and project financing.
- Technology and equipment requirements may influence construction design and procurement.
- Financing structures may depend on long-term capacity commitments, leases, and the allocation of technology and residual-value risk.
The milestone identified by Fortune provides a useful measure of how far the sector has already come. The next phase of development may increasingly depend on whether the physical, financial, energy, and regulatory infrastructure necessary to deploy this capital can keep pace.
Morgan Lewis Data Center Initiative
Our data center practice brings together lawyers across disciplines to address the increasingly interconnected issues involved in developing, financing, investing in, and operating data centers and digital infrastructure.
The firm’s multidisciplinary team advises participants across the data center lifecycle, including on-site acquisition and development, construction and leasing, power procurement and energy infrastructure, project and structured finance, investments and joint ventures, network connectivity, equipment and technology procurement, supply chain matters, tax and economic incentives, regulatory compliance, data center operations, and compute capacity arrangements.
For additional developments and analysis, visit our data center practice and Data Center Bytes blog.