The most important number from Thursday was not Broadcom’s revenue miss. It was the gap between the Dow and the Nasdaq. The blue-chip index closed at a record, up 874.86 points, while the S&P 500 added 0.41 percent and the Nasdaq Composite slipped 0.09 percent. A market does not look like that during a selloff. It looks like that during a rotation.
Broadcom supplied the trigger. The stock fell roughly twelve percent after an update that failed to clear expectations that had been set far too high, and the damage spread through the semiconductor complex with the mechanical efficiency of a sector that had been priced for perfection. Micron fell more than seven percent. The losses were not confined to names with company-specific news, which is the tell: when a single earnings report drags down stocks that did not report, the market is repricing a theme, not a company.
But the money did not leave the building. It moved down the hall. Financials, healthcare, and industrials absorbed what came out of chips, which is why the headline indices rose even as technology was the only sector with a meaningful decline. This is the distinction that matters and the one most easily missed by anyone watching the index level instead of its composition. The tape was green. The leadership was not.
The sharper read came from Neuberger’s Charles Kantor, who narrowed the rotation to a smaller address than the usual complaint about the Magnificent Seven. The exit, in his framing, is not from mega-cap technology broadly. It is from semiconductor capital equipment and hardware specifically — the machines that build chips and the companies that make them. That is a far narrower wound than a flight from big tech, and it has a precise location. His counterpoint deserves equal weight: the demand to build out compute and data-center capacity runs years into the future, and that pipeline is a powerful structural force. Read his way, this is rotation within an intact thesis, investors trimming an expensive pocket rather than abandoning the trade that has defined the cycle.
Friday opened pointing the same direction. S&P 500 futures slipped about 0.61 percent and Nasdaq 100 futures fell 1.19 percent while Dow futures held near flat — the Thursday trade, extended overnight. For anything leveraged to the semiconductor tape, that pre-market reading translates into a gap-down open before a single share changes hands, because daily-reset products amplify the underlying move by two or three times in both directions. The instruments built to express conviction in chips are precisely the instruments sitting at the center of the pocket being sold. Concentration that looks like leverage on the way up looks like exposure on the way down, and the daily reset that powers these funds quietly bleeds value in exactly the kind of choppy, directionless tape that follows a sentiment shock. A leveraged FANG+ product is marginally broader, reaching the full mega-cap basket rather than hardware alone, but it is still all technology and still magnified.
Into this sits the May jobs report, due at 8:30 a.m. Eastern. Economists surveyed by Dow Jones expect 80,000 jobs added, below the roughly 150,000 monthly average of the prior two months, with unemployment holding at 4.3 percent. The temptation is to treat this as the day’s catalyst. It probably is not. A soft number is already the market’s preferred outcome, because cooling labor keeps rate cuts alive, and a cooling labor market is no longer news. If the print lands near consensus, it removes a source of uncertainty and lets the existing flows keep running. The report’s expected value is benign and largely priced. Its risk is not. A meaningful upside surprise revives the higher-for-longer fear that punishes high-multiple semiconductors first, and a sharp downside miss invites a growth scare that would override the dovish read entirely. Some of Friday’s pre-market weakness is the rotation continuing; some is simply nobody wanting to carry risk into a number that can break either way.
The after-hours noise added a coat of red that had nothing to do with any of this. Lululemon fell eleven percent on cut guidance, a consumer-side story that says more about discretionary spending than about compute demand, and it should not be read into the semiconductor picture.
So the fork is clean. If the number prints in line, the rotation remains the story, and the question is how far investors are willing to push the trade out of hardware before the buildout thesis pulls them back. If the number surprises, the number becomes the story, and everything else waits. Either way, the lesson of Thursday holds into Friday: watch what the index is made of, not what it prints. The market is not selling the future of compute. It is selling the stocks that got too expensive betting on it.