A significant wave of selling hit AI-related technology stocks on Monday and Tuesday, triggering the Nasdaq’s sharpest two-day decline since the post-Iran-war recovery rally began. BBC News reports that the Nasdaq fell roughly 3 percent by close of trade, with semiconductor companies bearing the brunt: Nvidia, Intel, and the global chip index all slid sharply, and SpaceX — which went public on 12 June — dropped below its $150 IPO price before recovering to close around $156. NPR’s analysis shows Micron Technology falling 13 percent in a single session after having surged roughly 800 percent over the previous year on AI-driven memory chip demand, while Alphabet fell 5 percent and Samsung and SK Hynix each shed 12 percent. Separately, BBC Business reported that Oracle has announced the elimination of 21,000 jobs as it restructures around artificial intelligence. The sell-off ended a 90-day rally that had pushed valuations to record highs and follows over $580 billion in corporate AI investment globally in the past year alone, according to figures cited by NPR.
The received wisdom
The mainstream technology-optimist view treats Monday’s sell-off as healthy profit-taking after an unprecedented run — a temporary digestion pause, not a structural reversal. Bank of America’s semiconductor analyst Vivek Arya argued that sticky inflation and strengthening long-term AI demand would ultimately drive sector forecasts higher, and that the industry is transitioning from defending initial return-on-investment to solving physical infrastructure and power constraints, which are solvable engineering problems. The broader optimist case rests on a strong empirical foundation: AI tools are genuinely increasing productivity in software development, content creation, customer service, and drug discovery; the major cloud platforms are reporting strong enterprise uptake; and the upcoming IPOs of OpenAI and Anthropic, potentially among the largest in stock market history, reflect genuine investor appetite. On this reading, those who call bubble are simply underestimating the pace and scope of a genuine technological transformation.
A different read
The bears have the harder-to-dismiss argument, and not simply because any sell-off creates the appearance of vindication for the sceptics.
The structural concern is timing. Technology investment cycles have a consistent pattern: a genuinely transformative technology arrives, capital floods in, early winners achieve staggering valuations on the basis of projected future earnings, and then the market discovers that the technology works but the monetisation timetable was fantastically optimistic. This is not a prediction that AI will fail. It is a prediction, grounded in every previous technology cycle from the railroad bubble of the 1840s through the dot-com crash of 2000–2001, that even correct long-term bets can produce catastrophic short-term losses if the entry price already incorporates a decade of optimistic earnings growth.
The Micron case illustrates this precisely. NPR notes that the stock had risen 800 percent in a year. An 800 percent gain in a commodity semiconductor company means the market was pricing in a sustained, structurally elevated demand environment for memory chips — essentially, that AI training and inference workloads would consume ever-greater quantities of DRAM and NAND, indefinitely, at margins that memory manufacturers have historically never sustained. That is a heroically optimistic assumption. The 13 percent single-day correction, on this reading, is not evidence that the AI trade is over; it is evidence that the price had already incorporated years of upside that the underlying business reality cannot deliver in the near term.
Oracle’s 21,000 job cuts tell a related but slightly different story. The company is not retreating from AI — it is restructuring around it, which in practice means replacing human labour with automated processes. This is precisely what AI boosters have been promising: productivity gains that liberate capital from labour costs. But those gains arrive as job losses for tens of thousands of workers, and they arrive before the new jobs created by the AI economy are visible or countable. The macroeconomic transition cost is real, measurable, and concentrated; the long-run gains are diffuse, delayed, and speculative. This asymmetry is not a reason to oppose AI; it is a reason to be clear-eyed about the human disruption that even a successful AI transition entails.
Gil Luria of D.A. Davidson described the market’s oscillation with unusual precision: “The market just continues to oscillate between ‘AI is going to be great and increase productivity and all these companies are going to win’ and ‘AI is a big waste of time and it’s not worth the return on investment at all and this is all one big bubble.’” What Luria is describing is not irrational; it is the correct response to genuine uncertainty about a timetable, not about the technology’s ultimate value. The question is not whether AI will eventually deliver transformative productivity. It almost certainly will. The question is whether the current stock prices already fully discount that future — and whether the investors holding those prices at these levels can survive the gap between now and then without being forced to sell.
The right conservative instinct here is different from both the tech-booster enthusiasm and the progressive suspicion of big tech. It is the instinct of the prudent investor: acknowledge the long-run secular trend, resist the temptation to extrapolate it into equity prices, and remember that markets have a well-documented tendency to overshoot in both directions. The honest version of that view was said by Warren Buffett about a different era’s enthusiasm: “Only when the tide goes out do you discover who’s been swimming naked.”
What to watch
Micron’s earnings report this week is the immediate bellwether — if AI data-centre demand is as robust as bulls claim, Micron’s numbers should confirm it; a miss would validate the sceptics’ timing argument. Watch whether the OpenAI and Anthropic IPO timetables shift in response to market conditions; delayed or repriced listings would signal that institutional investors are re-examining AI valuations more broadly. Watch Oracle’s workforce reduction timeline and whether other major enterprise software companies follow with similar AI-driven restructurings. And watch whether enterprise customers in the next quarterly earnings cycle report measurable productivity gains from AI tools — that data, more than any market movement, will determine whether this sell-off is a correction or a turning point.
— J