Pulse24 Original
Tech Stocks Cost 17 Times Earnings Now. That's Cheaper Than the S&P 500 for the First Time Since 2017.
November 9, 2026
Tech's forward earnings multiple has roughly been cut in half since its October 2025 peak, and for the first time in nine years the sector trades below the broader market. Truist's chief investment officer calls it a buyers' strike, not a verdict on the business.
Seventeen times forward earnings. That's the going rate for the technology sector inside the S&P 500 right now, according to a note from Truist Wealth chief investment officer Keith Lerner. The broader index costs 18 times. For the first time since 2017, tech is the cheaper half of that pair.
A year ago the gap ran the other direction, and by a lot. At its October 2025 peak, the sector traded at 32 times expected earnings, roughly double where it sits today. Lerner's framing is blunt: at that level, the market had "baked in a lot of good news." At almost half that multiple, he argues, it's now pricing in a lot of bad news instead.
[[IMG1]]
What Changed
The repricing didn't happen gradually. The PHLX Semiconductor Index has fallen for five straight weeks, shedding close to 30% of its value over that stretch, an extension of the plunge that wiped $1.4 trillion off chip stocks in a single session earlier in the month. Comparisons that steep only turn up a handful of times in modern market history: the dot-com unwind in 2001, the 2002 bear market, the 2008 financial crisis, and the pandemic crash of March 2020.
Inside individual names, the damage has been uneven but severe. In the week ending November 6, Marvell and Sandisk each lost roughly 28%. Micron and Dell fell about 22%. Nvidia, the stock most synonymous with the AI trade, was comparatively calm, down nearly 9%. Amazon moved the other way entirely, climbing around 9% on the same tape.
That split matters more than the headline index numbers suggest. The S&P 500 and the Nasdaq each fell close to 3% that week, a fourth consecutive losing week for both. But the equal-weight version of the S&P 500, which strips out the outsized influence of a handful of megacap names, closed at a record on Wednesday, one more sign that stripping AI-exposed megacaps out of the index tells a very different story than leaving them in. The broad market isn't struggling. A specific corner of it is.
Why It Matters
Valuation multiples are a crude tool on their own. A stock trading at 17 times earnings isn't automatically a bargain, any more than one at 32 times was automatically overpriced. What the multiple actually tells you is how much optimism, or pessimism, is baked into the price relative to the earnings analysts expect.
Lerner's view is that the pendulum has swung far enough to change the risk calculus. "We are moving toward a situation where a little good news could go a long way," he wrote, adding that a slower pace of data center construction could ease near-term costs for AI infrastructure spenders and stretch out the investment cycle rather than end it.
Not everyone reads the chart the same way. Michael Martin, founder of TraderMindset.com, argues the price action itself is the signal, regardless of what the multiple says. "Lower prices mean lower market cap," he said. "It doesn't mean better value." His own approach for now is to raise cash rather than buy the dip, and he'd only consider call options if the S&P 500 holds its 200-day moving average.
What to Watch Next
History offers a reason for cautious optimism, with caveats attached. The S&P 500 has logged 33 streaks of four or more losing weeks since 1990, by Yahoo Finance's count. Three months after those streaks ended, the index was higher about 80% of the time, with a median gain of 6%, better than the roughly 3% median return that follows any random week.
The caveats are the ones that matter most. The exceptions were the dot-com bust and the 2008 financial crisis, both periods when the index kept falling for another year rather than recovering. Which outcome this looks more like probably depends on whether AI capital spending is slowing because demand is genuinely cooling, or simply because the hardest-hit suppliers are digesting an unusually fast buildout.
Truist's own answer, for now, is to stay overweight US large caps while adding small caps as a partial hedge against sharp rotations away from megacap tech, pointing to a pattern in which stocks have risen in the twelve months after every US midterm election since 1946. Whether that seasonal tailwind offsets a slowing AI buildout is the question the next few months of earnings and spending data should start to answer.
The Pulse24 Take
A cheaper multiple is not the same thing as a floor. What changed this month is that the market stopped assuming AI infrastructure spending only accelerates, and started pricing in the possibility that it merely continues. Those are different stories for the same group of stocks, and the valuation reset reflects genuine uncertainty about which one is true, not a verdict either way.
The more useful habit for investors watching this unfold is less about guessing where the multiple bottoms and more about separating the companies whose earnings depend on AI capex from the ones that simply got dragged down alongside them. Marvell and Micron selling off on memory and networking demand questions is a different story than Nvidia drifting lower mostly on sentiment. The next few weeks of earnings and spending commentary should start to separate those threads.
How we read the data
Curious how we get from raw data to a take like this? Our Trader's Toolkit walks through the tools we lean on.
Explore the Toolkit