The day after Microsoft Corp. reported earnings, its stock climbed by more than the entire market value of most companies. Shares rose enough to add about $450 billion to Microsoft’s market capitalization, the biggest single-day gain in nearly 18 years, according to Dow Jones Market Data.
The move rippled through global markets before the New York open. South Korea’s KOSPI posted its largest single-day gain on record, and Taiwan’s benchmark index closed at a record, up 3,186 points, its biggest point gain ever. Chip stocks that had been sold off through July rebounded across the board, from Nvidia and AMD in the U.S. to SK hynix and TSMC in Asia.
The trigger was Microsoft’s cloud results. Azure and related services grew faster than analysts expected, and the company laid out a computing investment path that investors read as disciplined rather than reckless. The spending numbers were large; the difference, fund managers said, was that Microsoft showed the revenue attached to them. Bookings for Azure contracts grew at a double-digit clip, and the company’s comments on capacity utilization suggested the data centers being built are filling up.
The July selloff had been driven by a simple fear: that the biggest AI spenders were pouring hundreds of billions into data centers with nothing to show. Microsoft’s quarter answered with cloud growth and, critically, bookings and revenue that are starting to move with the spending. The panic that defined the month, in which chip stocks fell sharply and several AI-linked funds took heavy losses, retreated in a single session.
Cathie Wood’s ARK Investment Management bought $15.6 million of Nvidia stock across five of its exchange-traded funds on the same day, a small sum for a fund complex but a pointed gesture from a firm known for trading on conviction. The purchases came as Nvidia shares recovered from their July lows, and the timing was widely read in trading circles as an attempt to catch the bottom of the AI selloff.
Analysts described the session as a reset of the AI trade’s terms. Investors no longer pay for promises, they pay for evidence. The companies that can show quarterly returns on capital, Microsoft on Wednesday evening and Amazon the next day, are being rewarded. Those that cannot, such as Meta, whose shares have now fallen for eleven straight sessions, are being punished. The divergence has become the defining pattern of the market’s second half.
The scale of the move deserves context. Microsoft’s $450 billion single-day gain is roughly the size of the entire market value of Oracle, one of its biggest rivals in the cloud race. It also erases a large part of the losses the stock suffered during July, when the AI trade came under its most serious attack since the ChatGPT boom began in late 2022.
What made the rally durable, in the view of traders, was where it started. The gains began in Asia before the U.S. open, spread across chipmakers and cloud names, and held into the close. A market that had been selling AI exposure for a month bought it back in one day, and the breadth of the move, from Seoul to Taipei to New York, gave it the look of a genuine turn rather than a technical bounce.
The underlying debate has not disappeared. Microsoft’s report does not settle the question of whether AI spending will ever earn a return for the industry as a whole; it settles, for this quarter, the question of whether the largest spender can show results. The companies reporting in the weeks ahead, including the chip suppliers whose order books drive the buildout, will face the same standard, and the ones whose numbers fall short will find no shelter in Microsoft’s rally.
Microsoft’s report also gave the supply chain a lift. Nvidia, whose shares had fallen sharply from their June highs during the July selloff, joined the rally, and analysts said the strength in Azure validates the order books of the chipmakers building the infrastructure. The logic runs in a circle: cloud growth justifies more data centers, more data centers mean more chips, and more chips mean more revenue for the suppliers that had been sold off the hardest. That circularity has been the crux of the debate all summer, with bears arguing it will eventually break and bulls arguing it is the whole point of the trade.
For the AI trade, the message was plain: the correction is over for companies with earnings, and the scrutiny moves to those without. The companies that can point to quarterly evidence, in cloud revenue, in chip orders, in data center utilization, will get the benefit of the doubt. The ones that cannot will keep sliding until they produce it, and the gap between the two groups, visible in a single week of trading, is likely to define the market for months.
