Semiconductors: did our stocks crack in July?

2026-07-06 · By Lubin Danilo, founder of Lubin Investment

Analyze a stock on Lubin Investment

On July 1, 2026, Meta announced it would rent out its excess computing capacity, casting doubt on the long-feared AI chip shortage and sending the whole sector lower: Micron fell 13% in a single session, Intel 9%, AMD 7%. Taiwan Semiconductor and ASML, two stocks our method rates highly, reacted very differently.

• On July 1, 2026, Meta announced Meta Compute, a service renting out its excess AI computing capacity to outside clients.
• The announcement flipped, in a single day, the market's long-held belief that AI computing power would stay scarce for years, triggering a broad semiconductor selloff.
• Micron lost 13% in a single session (roughly $138 billion in market value erased), Intel 9%, AMD 7%, and ASML about 4.7%.
• Taiwan Semiconductor, rated 9 out of 10 by our method, held up far better than ASML, rated 10 out of 10: the valuation gap between the two explains much of the difference.
• Nothing in this correction changes Taiwan Semiconductor's or ASML's industrial monopoly: what moved is the price the market is willing to pay for their future growth, not their actual business.

July 1, 2026: the announcement that flipped the table

On July 1, 2026, Meta announced Meta Compute, a service renting out to outside companies the excess computing capacity it isn't yet using for its own AI needs. On paper, it looks like a simple diversification move. In practice, it instantly broke the story the market had been telling itself for two years: a durable shortage of AI computing power, where demand would keep outrunning supply for years. If Meta, one of the world's biggest chip buyers, suddenly has spare compute to sell, maybe supply is catching up with demand faster than expected. That single idea was enough to send an entire stock sector tumbling within hours.

Why the correction hit some stocks harder than others

The selloff didn't hit every stock in the sector equally. Micron, the memory maker, lost 13% in a single session, roughly $138 billion in market value erased in a few hours. Intel dropped 9%, AMD 7%. In Asia, SK Hynix and Samsung, the two other global giants of HBM memory (the high-speed memory used in AI chips), fell 9% and 7% respectively, even triggering a temporary trading halt on the Seoul exchange. Memory was the hardest-hit segment, which makes sense: it's the link most directly exposed to fears of an oversupply. ASML, the Dutch maker of the machines that etch the world's most advanced chips, fell about 4.7%. Taiwan Semiconductor (TSM), the world's largest chip foundry, held up much better.

Taiwan Semiconductor: the stock that held

TSM manufactures roughly 85% of the world's most advanced chips, the ones found in Nvidia, Apple, and AMD processors. None of its customers can switch to a rival: building a competing fab costs between $15 and $30 billion and takes years, while still trailing TSM's technology by two to three years. That industrial monopoly has nothing to do with Meta's announcement, and the market seemed to understand that: the stock was still up roughly 46% year to date in 2026 after the correction, driven by orders from Nvidia, AMD, Apple, and Broadcom. It trades around 24 times expected earnings, cheaper than Nvidia or Broadcom. Our method rates TSM 9 out of 10, and crucially, the stock trades at only 2.4 times its annual free cash flow (the cash the business actually generates once every bill is paid). A multiple that low leaves very little room for disappointment: there was hardly any optimism left to correct out of the price.

ASML: same quality, a more fragile price

ASML holds a near monopoly on EUV (extreme ultraviolet) lithography, the etching technology required to manufacture every advanced semiconductor. Without its machines, neither TSM, Samsung, nor Intel can produce the most powerful chips. Our method rates it the maximum, 10 out of 10. But ASML trades at 51.6 times its annual free cash flow, against 2.4 times for TSM. A multiple that high already prices in an enormous amount of future growth. When an event like Meta's announcement makes the market doubt the AI growth trajectory, it's precisely this kind of stock, expensive relative to expectations, that has the most ground to lose, even though nothing changed in its actual business. That's exactly what happened: ASML fell more than an already cheap stock like TSM.

Not the first time this kind of fear has hit the sector

This isn't the first correction of its kind since the AI boom began. Every time doubt surfaces about the strength of demand, the whole semiconductor sector absorbs the shock together, before sorting itself out by actual company quality in the following weeks. What separates a passing correction from a real reversal is what shows up next in the numbers companies themselves report: orders, margins, guidance. A price drop only becomes a real signal once confirmed by several quarters of results, not by a single announcement, however dramatic.

The lesson behind this market move

What happened over the past few days illustrates exactly why I always separate two distinct questions when I look at a stock. One: is this a good business? Two, entirely separate: what price is the market asking me to pay for it? Meta's announcement changed nothing about TSM's or ASML's quality: their industrial monopolies are intact, their customers have nowhere else to go. What moved is the price investors are willing to pay for future growth, and that price depends directly on whichever narrative dominates at a given moment. A stock already trading at a low multiple, like TSM, has less optimism to lose. A stock trading at a high multiple, like ASML, has more, even though its underlying quality remains unchanged.

What I'm watching now

I don't claim to know whether Meta Compute really signals a coming glut of AI computing capacity, or whether the market simply overreacted to an isolated piece of news. Nobody knows that for certain at this stage. What I do instead is apply the same grid before and after the news: business quality first, price second. On that basis, TSM remains a very high quality stock at a price that leaves a real margin of safety. ASML remains just as solid industrially, but at a price that assumes AI growth keeps running smoothly for a long time yet. You can check every criterion on the <a href='/analyse/TSM'>TSM analysis page</a> and the <a href='/analyse/ASML.AS'>ASML analysis page</a>, or read our full write-ups on <a href='/blog/taiwan-semiconductor-tsm-analyse-fondamentale'>Taiwan Semiconductor</a> and <a href='/blog/asml-holding-analyse-fondamentale-monopole-euv'>ASML</a>.

FAQ

What is Meta Compute?

It's a service Meta announced on July 1, 2026, to rent out to outside companies the excess computing capacity it isn't using for its own AI projects. The announcement made the market fear that the long-feared AI computing shortage might be less severe than expected.

Why did all semiconductor stocks fall at once?

Because the entire sector was priced on the assumption that AI demand would keep outrunning supply for years. Once that assumption was called into question, investors repriced the whole sector at once, especially the memory makers most directly exposed.

Are TSM and ASML still good stocks after this correction?

On business quality, yes for both: their industrial monopolies are intact. The difference is in price. TSM trades at 2.4 times its annual cash flow, a level that leaves a margin of safety. ASML trades at 51.6 times, a level that assumes nothing disrupts its growth trajectory.

What is P/FCF again?

P/FCF (price to free cash flow) compares the stock price to the cash the business actually generates each year, after every bill is paid. A P/FCF of 10 means you're paying today for ten years of that cash. The lower it is, the cheaper the stock.

Should you buy the dip in semiconductors?

I don't give individual buy recommendations. This episode mostly shows the value of separating a business's quality from the price you pay for it: a one-off piece of news rarely changes the real quality of an industrial monopoly, but it can move the price disproportionately. Do your own research before any decision.

Analyze a stock on Lubin Investment

About the author

Written by Lubin Danilo, founder of Lubin Investment. A self-taught individual investor, I find fundamental analysis fascinating, and it has delivered excellent results. For three years now, my performance has beaten the S&P 500. But analyzing every stock took too much time: sites with incomplete data, calculation methods and criteria never aligned with mine. And spotting the best stocks was just as time-consuming, even with my own well-defined checklist. So I put my software development background to work to build this software, base my investment strategy on its results, and share it with people who share the same passion as me. It judges a company's quality and its price separately, using criteria drawn from the financial literature (Warren Buffett, Michael Mauboussin, Aswath Damodaran).