AI's $9 Trillion Investment Boom Faces Growing Profitability Doubts

Traders work on the floor of the New York Stock Exchange (NYSE) as concerns mount over whether massive investments in artificial intelligence will deliver sustainable returns. Astronomical sums are being poured into the artificial intelligence (AI) industry, but warnings are mounting that history could repeat itself, with massive capital investment failing to deliver expected returns, as seen in the railroad and internet industries. James Mackintosh, a columnist at The Wall Street Journal (WSJ), argued in a column published on Oct. 18 (local time) that there is little evidence to suggest the current AI investment boom will avoid the pitfalls of past investment cycles. The scale of investment flowing into AI infrastructure is already staggering. Stijn Van Nieuwerburgh, a professor at Columbia Business School, estimated that total AI-related investment in the United States could reach $9 trillion by 2032. That would be equivalent to 3.2% of annual U.S. gross domestic product (GDP), potentially making it the largest capital investment undertaking in American history. However, even if enormous sums are spent building data centers, a significant portion of that investment may prove difficult to recover if AI technology fails to generate sufficient economic value. Mackintosh cited the British railway investment bubble of the 1840s and the fiber-optic network buildout during the dot-com bubble around 2000 as representative historical examples. While both ventures delivered enormous benefits to society, it took years for returns to materialize. As competition intensified, much of the economic value flowed to users rather than the companies that owned the underlying technology, leaving investors with disappointing returns. The AI industry could follow a similar path if the benefits of the technology take too long to materialize or if profits become increasingly concentrated among a small number of companies. Markets are already showing signs of concern that current high earnings levels may not be sustainable over the long term. Nvidia's forward price-to-earnings ratio (PER), based on projected earnings over the next 12 months, has fallen from 23 times to 15 times even as its stock price rose 15%. Micron's stock price has tripled, yet its forward PER remains below 5 times. This suggests that investors are assigning lower valuations to future earnings even as corporate profits continue to grow. The concentration of profits is also pronounced. Major AI-related companies, including Amazon, Alphabet, Meta and Nvidia, are projected to post a combined earnings increase of $700 billion this year, compared with less than $500 billion for the remaining companies in the S&P 500. The data center construction boom has benefited semiconductor and cooling equipment companies, but capital spending in related industries is widely expected to fall sharply once the current wave of construction subsides. Investor attention is now turning to the third-quarter earnings reports of major technology companies scheduled for release this week. In particular, results from Meta and Tesla, both members of the Magnificent Seven (M7), are seen as pivotal. Meta is expected to generate more than $63 billion in revenue, but key questions remain over slowing growth and its ability to sustain massive AI investment costs. For Tesla, investors will focus on profitability, cash flow and the scale of its investments in robotaxis and Optimus. Kevin Shea, chief analyst at BNY Wealth, said a key question this earnings season is whether hyperscalers can grow their profits faster than their capital expenditures. Capital spending by the five largest hyperscalers, including Meta, is expected to exceed $1.2 trillion next year. The industry now faces a difficult dilemma: If AI investment proceeds as planned but fails to generate sufficient returns, vast amounts of capital could go to waste. If investment slows, semiconductor manufacturers and other industries tied to the AI infrastructure boom could suffer significant setbacks.