For much of the past year, the surest bet on the stock market was anything connected to artificial intelligence: the chipmakers, the data-center operators, the power and cooling suppliers feeding an unprecedented building spree. In June, that confidence cracked — and the questions that emerged are not going away, as CBS News reported.
A wobble in the AI trade
A sharp sell-off in June hit AI-linked stocks, with technology-heavy indexes falling and marquee names such as Nvidia and Broadcom dropping in a single session, according to market reporting. Much of the loss was recouped within days — a sign of nervousness rather than collapse — but it marked a shift in mood. Investors who had cheered every announcement of new data centers began to scrutinize whether the spending makes sense.
The spending at the heart of it
The numbers are staggering. The biggest US technology companies — Microsoft, Alphabet, Amazon and Meta — are on course to spend several hundred billion dollars between them this year, much of it on AI data centers, chips and the power to run them, with industry-wide commitments running into the trillions over the next few years. That outlay is now large enough to weigh on the companies' cash flows even as their revenues grow.
The doubts
The core worry is return on investment. Surveys of corporate AI projects have found that the great majority have so far produced little measurable financial payoff, fueling fears that demand and profits may not justify the build-out — at least not yet. Skeptics point to echoes of past technology booms, when heavy over-investment in infrastructure ended in painful corrections, and note the added risk that some of the spending is financed with debt. Power is another constraint: connecting vast new data centers to electricity grids has become a bottleneck, raising costs and slowing projects.
The bull case
Supporters argue the opposite reading. The very scale of the commitments, they say, reflects conviction that AI demand is real and durable, not a bubble mentality — and that data centers, chips and power systems are tangible infrastructure underpinning workloads that are growing fast. The major players are not retreating; they are pressing ahead, betting that whoever builds the capacity now will dominate later. By this logic, the June jitters are a buying opportunity, not a warning.
What to watch
The argument will be settled, in part, by the numbers companies report in the coming weeks. Investors will pore over capital-spending guidance, cash-flow trends and any evidence that AI is translating into durable revenue. A confident message — or a decision to rein in spending — from the hyperscalers could swing sentiment sharply for the chipmakers, data-center landlords and power suppliers whose fortunes are now tied to the AI build-out. None of this is investment advice, and the "bubble" question remains a matter of debate, not settled fact.
