EquitiesIn plain sightTuesday 20 October 2026, 10:38
Is the Nasdaq cheap? Tech's P/E ratio fell from 25.6 in June to 20.4, below consumer staples. Prices barely moved: chip profit forecasts soared, and can fall.
A video doing the rounds on Instagram this week makes a bold claim: “This is the cheapest the NASDAQ has been since the invention of AI.” The post by The Trading Fraternity shows a Bloomberg chart that puts the Nasdaq-100’s P/E ratio, its price against the next year’s forecast profits, at 20.48.
The figures we can check point the same way, with one correction. Technology companies in the S&P 500 now trade at 20.4 times their forecast earnings for the next 12 months, FactSet’s weekly Earnings Insight shows. That is well below their five-year average of 25.5 and their ten-year average of 23.1. It is also below consumer staples, at 22.4: the makers of soap, food and drink, which investors usually buy for safety, not growth.
The surprise is not that tech looks cheap. It is how it got there. The price has barely fallen. The forecasts have soared. And a P/E built on forecasts is only as solid as the forecasts themselves.
Almost, but not quite. We read the sector table in 36 of FactSet’s weekly reports since October 2022. OpenAI released ChatGPT on 30 November 2022. That week, tech traded at 21.9 times forecast earnings. Five weeks later, on 6 January 2023, it was down to 19.6, below today’s 20.4, and below consumer staples too, FactSet’s report of that week shows. So the accurate version of the claim is: the cheapest since January 2023.
In the reports we read, tech’s P/E peaked at about 30 twice: in December 2024 and again in August 2025. It was still 25.6 on 5 June this year, FactSet’s report of that week shows. Since then it has fallen almost every month, to 21.8 at the end of July and 20.4 now.
Tech has also dipped below consumer staples before this year, on 27 March, 31 July and 9 October. What is new is the size of the gap. At 2.0 points, it is the widest in all the reports we read.
A P/E ratio can fall two ways: the price falls, or the forecast profits rise. This time it is mostly the second.
Prices did fall, but not by much. The Nasdaq-100 closed at 29,529 on Monday, the index’s daily figures show, 5.4% below its record of 31,225 on 6 October. On 5 June, when tech’s P/E was 25.6, the index closed at 28,958, lower than today. It is still almost two and a half times its level on the day ChatGPT came out.
The forecasts moved much more. Tech companies in the S&P 500 are on course for earnings growth of 65.1% in the third quarter from a year earlier, FactSet’s latest figures show. Their net profit margin is 32.3%, up from 27.5% a year ago. The chipmakers lead: earnings at semiconductor and chip-equipment firms are expected to rise 130%. Without them, tech’s growth would be 24.4%.
Since 30 September, tech is the only sector in the S&P 500 whose price has fallen while its forecast earnings rose, on FactSet’s chart of the two. When profits climb faster than prices, the P/E drops even if no one sells.
Much of that profit comes from one product: memory chips for AI data centres. Micron, the largest American maker, reported revenue of $54.23bn for the three months to 3 September, against $11.32bn a year earlier, its results show. Its gross margin was 87.0%, and it expects revenue of about $61.5bn this quarter. Micron and Nvidia are the two biggest contributors to the S&P 500’s earnings growth this quarter. Without them, the index’s growth would be 20.0% instead of 29.0%, FactSet calculates.
The boom is about price more than volume. China’s exports of integrated circuits rose 108.1% by value in the first nine months of the year but only 3.8% by number, China’s customs figures show. The value per chip roughly doubled.
Memory has long been a business of boom and bust. Prices rise when supply is short, makers build new factories, and prices fall when the factories open. Micron expects supply and demand to be “much tighter in fiscal 2027 and 2028” than this year, its prepared remarks say; we covered its results on 2 October. If that view proves wrong, the earnings forecasts fall, and tech’s P/E rises again without its share prices moving at all. The chip stocks have already shown this month how quickly the mood can turn, as we reported on 15 October.
The same report gives the S&P 500 two very different P/E ratios. On next year’s expected profits, the index trades at 19.0, just below its ten-year average of 19.1. On the profits companies actually made over the past 12 months, it trades at 26.0, above its ten-year average of 23.6.
The gap between the two is the growth analysts expect: 32.8% for the index’s earnings this year and 16.0% next year. Those forecasts slow quickly. For the second quarter of 2027, analysts expect growth of just 2.8%. And for the fourth quarter of this year, they expect the “Magnificent 7” to grow their earnings more slowly than the other 493 companies, 23.6% against 30.6%, for the first time since 2022.
There is another way to judge a P/E: turn it upside down. A P/E of 20.4 means forecast earnings of 4.90% of the price, the earnings yield. A 10-year Treasury, which carries no business risk, paid 5.09% on 16 October, Federal Reserve figures show. On that measure, tech still earns less than a government bond.
| Week | Tech P/E | Tech earnings yield | 10-year Treasury | Gap |
|---|---|---|---|---|
| 2 Dec 2022 | 21.9 | 4.57% | 3.51% | +1.06 points |
| 6 Jan 2023 | 19.6 | 5.10% | 3.55% | +1.55 points |
| 6 Dec 2024 | 29.9 | 3.34% | 4.15% | −0.81 points |
| 1 Aug 2025 | 30.0 | 3.33% | 4.23% | −0.90 points |
| 16 Oct 2026 | 20.4 | 4.90% | 5.09% | −0.19 points |
So tech is far cheaper against bonds than at its peaks in 2024 and 2025. But in the weeks after ChatGPT, when its P/E was about the same as today, bonds paid little, and tech’s earnings yield was well above them. Today the bond yield is higher than tech’s. The P/E has come back to where it was in early 2023; the cost of money has not.
Results will test the forecasts. Only 9% of S&P 500 companies had reported by Friday. Meta and Tesla, the first of the largest tech firms, report on 21 October, and FactSet updates its figures each Friday. Two things matter. If profits beat forecasts and forecasts rise again, tech’s P/E falls further and the cheap case gets stronger. If the memory boom cools and forecasts come down, today’s low P/E will turn out to have been a high one.
This is analysis and opinion, not investment advice.
WeeklyThe Notebook by email
Once a week: the stories we published, and the releases and decisions due in the week ahead. Free. One click stops it.
We email you a link to confirm. By signing up, you agree to the Terms and the Privacy policy.
Check your email. We sent a link to . Open it within 15 minutes to confirm.
You are signed in as . You can change this on your Saved page.
You get the weekly email. To stop it, use the link at the end of any of them, or your Saved page.
Our research · 52 min read
The lek has not bought so many euros since at least 1995, when Eurostat's records begin. In September 2026 a euro cost 91.85 lek on average, the lowest monthly figure in those records. In 2019 it cost 123.01. Put the other way round, the lek now buys 33.9% more euros than it did in 2019, and 37.7% more dollars.
Continue reading the story
