Turns out you can't just spend your way to profitability. Who knew?
*Written on 1 November about the October selloff; figures are as reported at the time.*
The tech sector's bill for its artificial intelligence ambitions has come due, and the market is not inclined to extend credit. The Nasdaq 100 is on course to shed more than $2 trillion in value this month alone, a staggering correction that arrives not with drama but with the efficiency of a closing account. This is what happens when an entire asset class built on the assumption that AI spending automatically generates returns collides with the inconvenient fact that it does not.
The October downturn represents something investors have historically struggled with: admitting they were wrong about valuations. For eighteen months, the algorithm-as-perpetual-motion-machine thesis dominated market thinking. Every earnings call that included the words "capital expenditure increase" triggered enthusiasm. Every press release announcing expanded AI infrastructure prompted fresh capital allocation. The market had decided that the only rational response to artificial intelligence was unlimited investment, and questions about whether those investments would ever turn profitable were treated as the concerns of pessimists and luddites.
That consensus has fractured. The Nasdaq 100 has dropped 5.8 percent so far this month, driven by a wave of earnings reports from the hyperscalers who had promised that their AI spending would eventually pay dividends. Alphabet reported negative free cash flow of $9.1 billion in the third quarter while simultaneously doubling capital expenditures to $51.4 billion. Meta and Amazon raised their capex guidance further. These are not companies struggling to survive. These are the most profitable technology firms in existence, armed with fortress balance sheets, announcing that they are burning cash faster than expected and planning to burn it even faster next year.
The market's response has been unforgiving. Intel plunged 29 percent in a single day, its worst drop in decades, a fall that would have seemed impossible to anyone who watched the company dominate semiconductor markets for decades. Reddit, which had benefited from being a destination for human discussion, lost about 30 percent of its value in a single day after warning investors that Google's AI-generated search answers were cannibalizing its traffic. The irony barely needs underlining: the companies investing trillions to build AI are destroying the content ecosystems that AI requires to function.
The Morning Brief
Enjoying this? Get it in your inbox.
The arithmetic that spooked investors is straightforward enough. Morgan Stanley forecast that capital spending on AI infrastructure will exceed one trillion dollars next year alone. These are not investments made from retained earnings. These are projects financed by debt, and rising borrowing costs have made that debt substantially more expensive than it was when the AI boom began. A company that can borrow at three percent can justify spending a dollar to make a dollar-fifty. A company that borrows at seven percent finds the math considerably less attractive.
What makes October's decline particularly damaging to the AI narrative is that it arrived not as a sudden shock but as the logical conclusion of information everyone already possessed. The earnings reports that triggered selling contained no surprises. Alphabet and Meta and Amazon had never hidden their capex plans. They had been explicit about their commitment to AI infrastructure. Yet the market had somehow convinced itself that these massive expenditures would generate returns on a timeline that made the spending appear reasonable. October appears to be the month when enough investors simultaneously realized they had never actually done the math on that timeline.
There are, to be fair, faint signals suggesting that some AI investments are beginning to generate returns. Microsoft and SpaceX both posted results that suggested their capital deployment in artificial intelligence was producing measurable benefits. Yet they are outliers in a landscape populated by companies whose AI spending continues to generate questions rather than answers. Nvidia, the industry bellwether whose results are scheduled for mid-November, will face intense scrutiny over whether the demand for chips used in AI training justifies the astronomical valuations placed on the company.
What we are witnessing is a market correction that should not have required correction in the first place. The companies burning cash faster than expected told investors they were burning cash faster than expected. The market chose to interpret this as bullish. Now investors are recalibrating their expectations, which means recalibrating their valuations downward. Two trillion dollars in value is on course to be erased not because the fundamental business models of these companies have changed, but because their prices reflected assumptions about AI returns that were never tested and, it increasingly appears, were never realistic.
The tech sector will recover. The capital will continue to flow toward artificial intelligence. But the era in which the market automatically rewards AI spending without regard for returns has ended. What comes next is the harder work of actually measuring whether these investments create value. History suggests that phase tends to be considerably less lucrative for shareholders.
Subscriber Only
Subscribe to The Alignment Times and get every article delivered to your inbox.
Photo by Rafael Minguet Delgado via Pexels
Ingrid Holt
Staff writer covering financial markets and corporate strategy. Has strong opinions about spreadsheets.
Numbers Better Than Expected; Feelings Remain Complicated
Apr 5, 2026
Country Famous For Engineering Efficiency Finds Process Difficult To Engineer Away
Apr 3, 2026