Cumulative spending globally on data centres alone could top $30tn by 2050, according to a projection by PwC, almost matching the value of outstanding US Treasuries. Of late, however, less-than-stellar earnings results of tech titans and alarming cash burn rates have forced investors to heavily scrutinise whether massive corporate investments in artificial intelligence infrastructure will ever yield real, near-term profitability.
Collectively, 20 of the most valuable global AI and semiconductor stocks shed more than $1tn in value last week. In October, over $2tn in market value evaporated from the tech-heavy Nasdaq-100 index as Big Tech earnings intensified investor scrutiny of whether the billions being poured into AI can generate returns quickly enough.
Several megacap tech giants and semiconductor plays faced eye-watering individual market cap reductions across the 31-day period. Intel cratered 45.3%, fueled by a disastrous 29% crash on October 30 alone following a poor Q4 revenue outlook and PC sales bottlenecks. Google parent Alphabet slumped 18.7% as it recorded its worst single-day drop in history of 12.6% on October 29 after posting a negative free cash flow of $9.07bn.
Meta Platforms dropped 15.2% as investors reacted negatively to plans for accelerated 2027 infrastructure spending.
The semiconductor space took a massive hit, with Nvidia falling 10.4% and AMD sliding nearly 11%. "It may be that we’ve overestimated how strong the growth might be, but they’re not going negative in aggregate,” said John Bilton, global head of strategic research, multi-asset solutions at JPMorgan Asset Management. "We have an insatiable and growing demand for data.”
Microsoft is reportedly planning to raise about $25bn through its first US investment-grade bond sale in more than nine years. Among other hyperscalers, Meta, Amazon and Alphabet have issued hundreds of billions of dollars in bond sales over the past four years, worrying investors about overreliance on the medium for raising capital. JPMorgan expects global AI capex to reach $5.5tn by 2030, and sees $900bn being spent in 2026 alone.
AI companies’ growing reliance on debt funding is emerging as a new risk for investors, challenging the view that the companies at the heart of the boom are insulated from macroeconomic pressures. There are already mounting signs that global growth is cooling. An AI-driven stock market slump could sharply deepen the slowdown and even tip several major economies into recession, according to Jamie McGeever, columnist for Reuters.
The AI boom’s role in the bull run on Wall Street and other stock markets around the world is well documented. It has also propped up the global economy, offsetting one of the biggest energy crises in history, rising global inflation and deepening geopolitical uncertainty.
Any deflating of the AI bubble could have significant economic implications.
In a wider sense, there are two schools of thought. The first dismisses serial scare stories about the power of AI as merely hype to build excitement around upcoming initial public offerings, according to Mike Dolan, a columnist for Reuters. The other is that AI development is now too big to fail strategically, and governments and the Big Tech have no option but to press ahead. For sure, a total AI capex bust and complete market meltdown are both unlikely. The infrastructure build-out is still in its infancy, as is the AI revolution itself. But for the investors, the AI market narrative and sentiment are clearly shifting.