The US economy has reached a notable investment milestone as inflation-adjusted spending on information-processing equipment, including computers and equipment used in data centers, has moved above residential investment. The shift highlights the scale of the capital spending boom surrounding artificial intelligence and the infrastructure required to support it.
According to Bureau of Economic Analysis data cited by San Francisco Federal Reserve economist Adam Shapiro, real private residential fixed investment stood at about $748 billion in the second quarter of 2026. Information-processing equipment spending reached roughly $752 billion over the same period.
The difference is small, but the direction of the change is significant. Housing has traditionally represented a major source of US investment, while the rapid expansion of computing infrastructure reflects the growing economic importance of AI, cloud computing and data-intensive applications.
AI Infrastructure Changes the Investment Mix
The latest figures show how quickly the balance has shifted. Real residential investment was about 18% below its early-2021 peak, while spending on information-processing equipment had increased approximately 51% over the same period, according to the data cited by the San Francisco Fed.
Higher borrowing costs have been an important factor affecting the housing market. Mortgage rates have remained elevated, increasing financing costs for both homebuyers and developers and contributing to weaker residential construction.
AI infrastructure spending has behaved differently. Major technology companies have continued committing enormous sums to data centers, advanced processors, networking equipment and other infrastructure needed to train and operate increasingly demanding AI systems.
The scale of those commitments is expected to grow further. S&P Global estimated that capital expenditures by Alphabet, Amazon, Microsoft, Meta, Oracle and SpaceX could exceed $1.3 trillion in 2027, compared with a projected $870 billion in 2026 and $470 billion in 2025.
That spending is creating demand across several parts of the economy, including semiconductor manufacturing, electricity generation, power transmission, cooling systems, construction and data-center equipment. The AI boom is therefore extending beyond software companies and into the physical infrastructure supporting computing.
The AI Buildout Comes With Economic Risks
The investment shift does not necessarily mean data centers themselves are now larger than housing as a category. The comparison involves residential fixed investment and information-processing equipment, a broader category that includes computing equipment used for many purposes. The distinction is important when interpreting the numbers.
Data-center construction alone remains a smaller component of US economic output. The White House’s 2026 Economic Report of the President said data-center structures represented about 0.13% of GDP in 2025, excluding the value of computer equipment installed inside those facilities.
Still, the broader computing investment cycle demonstrates how aggressively companies are preparing for higher demand for AI services. Hyperscalers are spending heavily even as financing costs remain relatively high, with some companies increasingly turning to debt to supplement their cash resources.
Related: U.S. President Trump Predicts AI Could Become 25% of the U.S. Economy
The biggest question is whether the expected economic returns from AI infrastructure will justify the extraordinary level of capital spending. S&P Global has warned that capital expenditure is currently growing faster than revenue for major technology companies, raising the possibility of excess capacity if demand fails to develop as projected.
Housing faces a different set of constraints. The US has experienced limited housing supply growth while high mortgage rates have discouraged transactions and new construction. FRED data shows nominal private fixed investment in residential structures at a seasonally adjusted annual rate of about $1.16 trillion in Q2 2026, underscoring that the exact comparison depends on which investment measure is used.
The broader economic picture is therefore not simply that America has stopped investing in homes. Instead, the data shows that the composition of investment is changing as businesses direct unprecedented amounts of capital toward computing infrastructure.
For technology and financial markets, the shift provides another measure of how deeply AI has entered the US investment cycle. Whether the trend continues will depend on AI demand, corporate returns, electricity availability, financing conditions and the ability of the industry to turn massive infrastructure spending into sustainable revenue.
About The Author
