Photo by Alex Luna on Pexels · Pexels License
On October 29, 2026, Nasdaq futures fell after Alphabet and Samsung both posted earnings that failed to answer the question hanging over the entire market: when does the AI spending start paying for itself?
The timing matters. Alphabet and Samsung sit at opposite ends of the AI supply chain, one buying computing power by the gigawatt and the other selling the memory chips that make it possible. When both disappoint on the same day, investors hear a message about the whole ecosystem, and they heard it. Treasuries gained for a second straight day as money moved toward safety, with the Federal Reserve having left interest rates unchanged.
The scale of what is at stake is hard to overstate. The Wall Street Journal has called the AI buildout the biggest economic bet in U.S. history. By 2026 the financing shifted in a telling way: early AI investment ran on operating cash flow, but the bond market now dominates, and roughly $3 trillion in off-balance-sheet commitments has been layered on top. That is a polite way of saying the industry stopped spending money it had and started spending money it borrowed, then moved some of the obligations where casual observers would not see them.
Then there is the circularity. Leading AI firms invest in one another's products and capacity, a flow of money that inflates revenue on both sides of the ledger. Graphics mapping these loops went viral in October 2025, and the Journal's warning still hangs over the market: the structure can be virtuous on the upside and vicious on the downside. Nvidia rode it to a $4 trillion valuation in July 2025, then past $5 trillion months later, a figure exceeding the GDP of every country except the United States and China. In 2025, AI-related companies produced roughly 80% of the gains in the American stock market.
The warning signs have been visible for a while. In late June and July 2026, South Korea's KOSPI index crashed 44% in 40 days, erasing $2.18 trillion in value, driven by a selloff in Samsung and SK Hynix after Big Tech showed weak short-term returns on AI infrastructure and fears spread about demand for high-bandwidth memory chips. A National Bureau of Economic Research study published in February 2026 found that 90% of firms reported no measurable impact from AI on productivity, even as executives projected AI would lift productivity by 1.4%. That gap between executive projection and workplace reality is the entire bubble debate in miniature.
Sam Altman, whose company OpenAI created ChatGPT, said in August 2025 that he believes an AI bubble exists. Jamie Dimon of JPMorgan Chase has said AI is real but that some of the money being invested will be wasted, and he sees a higher chance of a meaningful stock drop than the market was pricing in. Ray Dalio has compared current investment levels directly to the dot-com era. The valuation math supports the skeptics: the S&P 500 has traded around 23 times forward earnings, and the Shiller price-to-earnings ratio exceeded 40 for the first time since the dot-com crash. Microsoft offered a preview of the mood in late 2025, when it posted 18% revenue growth and 12% net income growth and still saw shares drop 4% after hours, purely because investors flinched at nearly $35 billion in quarterly AI infrastructure spending.
Here is the part that deserves the most attention. The people who made these bets will be fine no matter how they turn out. The exposure has been socialized. Five companies, Nvidia, Microsoft, Apple, Alphabet, and Amazon, held up 30% of the S&P 500 and 20% of the MSCI World index in late 2025, the greatest concentration in half a century. That means every worker with a target-date retirement fund, every pension, every college savings plan is riding on the assumption that AI revenue will eventually justify AI spending. Demand for computing power was still exceeding supply as of October 2026, and prices for capacity kept rising, which bulls read as proof of health and skeptics read as proof of a gold rush where the pickaxes cost more than the gold.
One bad earnings day proves nothing. But when the financing has moved to debt, the valuations sit at dot-com levels, the academic evidence shows productivity gains that exist mainly in executive presentations, and an entire national stock market has already crashed once this year on exactly this fear, the October 29 selloff is a stress test the whole market should take seriously.
Free Game Takeaway
If you have a 401(k) or index fund, check your actual exposure to the five largest tech stocks rather than assuming you are diversified. With those companies holding up 30% of the S&P 500, a broad index fund is a concentrated AI bet in disguise. Watch three signals in coming months: whether Alphabet and its peers show AI revenue growth matching capital spending in the next earnings cycle, whether bond issuance for data centers keeps accelerating, and whether more executives follow Altman and Dimon in publicly questioning valuations. The Korea KOSPI crash showed how fast a 44% drop can happen when the AI return question goes unanswered. Rebalancing decisions are personal, but understanding what you own is the free game here.