AMD rose 11.04%, Intel rose 7.71%, Micron rose 5.12%, and the Dow fell. A single session after the worst tech selloff since 2020 revealed that the AI trade is no longer one trade at all.
Friday's close was the mirror image of the two sessions before it. After a steep selloff driven by post-midterm anxiety about what a Democratic-controlled Congress would mean for data-center permitting and AI regulation, semiconductor names rebounded hard. AMD added 11.04%, Intel 7.71%, Micron 5.12%, Nvidia 2.78%. Rocket Lab rose 15.39%, Coherent 13.61%, Airbnb 11.67%, Vistra 10.51%, AppLovin 9.86%. The Nasdaq Composite gained 264.43 points to 24,579.42. The S&P 500 added 0.31%. The Dow Jones Industrial Average fell 237.33 points, or 0.46%.
The dispersion is the story. A flat-to-up index produced by an AI-related rally on one side and a defensive blue-chip pullback on the other is not a market repricing risk. It is a rotation inside the AI complex, and it happened against a backdrop of a genuinely strong macro print: October payrolls came in at 165,000 against expectations closer to 60,000, which supported risk appetite while pushing yields higher.
That combination is unusual and worth examining. Strong employment data is normally read two ways, and markets have historically resolved the ambiguity by selling duration: higher yields hurt the long-duration growth equities that trade on distant cash flows. What happened here was the opposite. The strong jobs print pushed yields up, and the stocks that rallied were the ones with the most near-term earnings leverage to AI infrastructure spending, not the ones with the most distant optionality.
The 2020 comparison is the part that deserves scrutiny. When the market sells off sharply and then recovers within days with the same leadership intact, it usually means the selloff was a positioning event rather than a thesis change. Leverage unwound, and when it was done, the same buyers returned. A genuine change in the AI investment case does not typically recover in forty-eight hours.
The real risk in the current setup is not the midterms. It is that AI infrastructure has become correlated with the business cycle in a way it did not before. Data centers are now capital projects with construction schedules, power contracts, and equipment backlogs, and the companies selling into them have revenue that converts to cash on a normal industrial timeline. Vistra and Coherent are power and component suppliers, not software companies, and their earnings respond to order flow like any capital goods business.
That correlation cuts both ways. It makes the AI complex more responsive to economic strength than the software analogy ever implied, which is part of why Friday's rally happened alongside a strong payrolls number rather than in spite of one. It also means the AI trade no longer hedges the macro cycle. Investors who bought the sector as a hedge against economic weakness bought something considerably more cyclical than they intended.
The breadth remained negative overall, which is the detail most likely to matter on Monday. A Friday rally led by a narrow set of high-beta semiconductor names, in a session where the Dow fell and overall breadth stayed negative, is a rotation within a crowded trade rather than a broad risk-on move. When the index recovers on the back of a handful of names while the median stock does not participate, the recovery says more about positioning than about conviction.
For now the market has made its position legible: it is not pricing a break in the AI investment case, and it is not pricing regulatory retreat either. It is pricing a stronger economy alongside continued infrastructure spending, and it is choosing the segment of the theme with the shortest distance between order and cash flow. That is a coherent thesis. It is also a more cyclical one than the sector's reputation suggests, and the next payroll print will test it.
Image source: i.ibb.co