Record AI Spending Fuels Debate on U.S. Economic Sustainability

Artificial intelligence is on track to become the largest investment boom in U.S. history, with spending expected to surpass that of railroads, highways, and the internet. A small group of large technology firms, including Alphabet and Amazon, is leading this expansion, investing hundreds of billions of dollars into data centers, hardware, and software to build their AI models.
While this investment has significantly boosted U.S. stock market performance—effectively doubling portfolio values since 2022 and contributing to consumer spending—investors are increasingly cautious. Despite the record expenditures, firms have yet to realize meaningful returns. Rising debt levels and softening revenues have spurred skepticism in financial markets, causing stocks to fluctuate in November. Rebecca Patterson, a macroeconomic researcher and senior fellow, noted that while the industry has adopted a “whatever it takes” approach to spending to maintain its position in the AI race, financial markets are growing increasingly wary of that strategy, particularly as the cost of debt rises.
Data from the Brookings Institution projects that U.S. investment in AI and related equipment could exceed $10 trillion between 2025 and 2032, averaging 3.6 percent of GDP annually. Roger W. Ferguson Jr., a distinguished fellow for international economics, described the scale as unprecedented, potentially one-and-a-half to three times larger than previous technology booms as a percentage of GDP. However, the benefits remain uneven. While sectors like finance and software development have seen productivity gains, many other industries have reported no measurable improvement. Furthermore, a McKinsey study suggests that AI-driven automation could displace between six million and 16 million American workers over the next decade as companies transition toward technical roles.
Concerns regarding an AI bubble are mounting. Critics point to the “winner-take-all” nature of the market, which can induce overinvestment, and the fact that several major tech companies reported negative free cash flow in their third-quarter earnings reports. Market volatility intensified in late October following disappointing earnings and further declined after the U.S. midterms, as investors reacted to potential government regulation and restrictions on the sector. Adding to the unease are reports of circular financing, where firms trade AI infrastructure and lend each other capital to secure future purchase commitments. Sona, an asset manager, identified 176 such deals totaling $3.6 trillion over the past three-and-a-half years. The ecosystem also faces “key man risk,” highlighted by the market value drop of Nvidia following the departure of CEO Jensen Huang amid an alleged Chinese espionage operation.
Despite these risks, experts suggest the current environment differs from past bubbles. Unlike previous start-up-driven booms, the companies currently leading AI development are established and profitable. Additionally, the technology is inherently multipurpose, and adoption has occurred at a faster pace than that of the personal computer or the internet. Experts maintain that even if financial setbacks occur, the fundamental technology is unlikely to vanish.
Meanwhile, significant constraints threaten to slow the build-out, including shortages in electricity, land, and water for data centers, as well as backlogs of two to ten years for critical imports like memory chips and power grid equipment. Although the Trump administration has moved to deregulate land use and permitting, experts like Vinh Nguyen suggest that most control remains at the state and local levels, where communities often oppose large-scale data center projects due to concerns about utility costs and quality of life.
Foreign policy also complicates the sector’s trajectory. Dependence on foreign imports for 90 percent of critical AI inputs leaves the U.S. vulnerable to supply chain disruptions, a risk further complicated by tariffs on materials like copper and steel. While the U.S. government views rapid AI deployment as essential to outcompeting China, Chinese AI models are rapidly closing the performance gap. Ultimately, Nguyen noted, the U.S. advantage may lie in AI safety and security, as the long-term success of the technology will depend on user trust.