Where the Chip Money Went: The Semiconductor Crash and the Software Comeback

The money that left chip stocks didn't vanish into thin air. It just moved.

People are calling it the great rotation. While the crowded, expensive tech winners were crashing, investors quietly shifted capital into everything they'd been ignoring for years: banks, healthcare, industrials, smaller companies, and the software businesses that had been beaten up all year.

What I want to do here is put the two camps side by side. It's hard to find a cleaner demonstration of how markets actually work than two groups moving in opposite directions over the exact same weeks.

Camp A: crowded and expensive semiconductors

The defining feature of this camp was that everyone already owned it.

From their June highs, SanDisk fell 53%, Marvell 45%, Intel 39%, and both Western Digital and Lam Research 38%. AMD dropped 22%, Broadcom 21%. The chip index as a whole is down about 24% from its June peak, putting the entire sector in a bear market.

The critical detail is that this wasn't about deteriorating results. Micron posted its fifth record quarter in a row with revenue above $41 billion. SanDisk nearly doubled revenue, with its data center business up 233%. Their stocks fell 32% and 53% respectively.

There's one reason this camp broke: the price already assumed perfection. Even flawless results only justified the price — there was nothing left to push it higher.

Camp B: the abandoned software names

The second camp started from precisely the opposite condition. Nobody wanted it.

Rewind to earlier this year and software was the group getting obliterated. Wall Street was terrified AI would wipe out SaaS businesses — that new AI tools would undercut their pricing, steal their customers, and make their products obsolete. Out of pure fear, plus a desire to pile into chips, investors dumped these companies as they fell 25%, 30%, even 60% and more.

But the businesses underneath never actually fell apart.

Take ServiceNow. It had tumbled about 27% on AI fears, yet its actual numbers were excellent: subscription revenue growing 25% and over $1 billion of AI-related business already signed. The stock has recovered roughly 33% off its lows — and on July 28, the very day chip stocks were collapsing, it was up almost 5%.

Intuit, the company behind TurboTax and QuickBooks, climbed from around $267 in June to $312 today — a 17% recovery in a month.

Adobe is a name I own, so let me be careful here: never buy something because I or anyone else on the internet holds it. Adobe traded at an all-time high of $700, fell all the way to about $190 recently, and now sits near $260. That's close to a 40% bounce off the lows.

Lined up by recovery from the lows: Datadog +145%, Snowflake +123%, Workday +42%, Adobe +40%, ServiceNow +33%.

The two camps, side by side

	Semiconductors / AI chips	Software (SaaS)
Sentiment at entry	Everyone owned it, extreme optimism	Fear that AI would kill it, forced selling
What the price assumed	Flawless execution continuing	The business model breaking
Actual business results	Record-setting quarters	Steady, solid growth
June-July price action	-16% to -53% from highs	+17% to +145% from lows
Examples	SanDisk -53%, Micron -32%	Datadog +145%, Snowflake +123%

The table makes the conclusion obvious. What separated these two camps wasn't business quality. It was what the price was assuming when you bought in.

What the year-to-date scoreboard says

At the index level, the rotation is even clearer.

The small cap Russell 2000 is up around 19% this year. The Dow is up almost 10%. The S&P 500 is up over 7%. And the tech-heavy Nasdaq is up just 5%. The forgotten, cheaper corners of the market are the ones actually winning.

Through the end of June the gap was starker still: the Russell was up more than 22% while the 50 biggest companies in the market were up just 2%. The little guys trounced the giants.

There's also a version of the S&P 500 where every company counts equally instead of being ruled by a handful of mega caps — and that equal-weight index hit a brand new high while momentum stocks were falling apart.

I want to draw a careful line here, though. None of this means small caps will win for the next ten years. Some of it is just short-term traders shuffling money around. These chip companies could go on to make new highs. I don't know, and I distrust anyone who claims they do.

Still, one thing is unambiguous. The cheap, ignored, unloved stuff held up beautifully while the crowded, expensive, everybody-owns-it stuff got crushed. That's value investing playing out in real time.

The crowd was wrong at both extremes

Overlay the two camps and you see a mirror image. Fear crushed software stocks too low, and when that fear proved overblown they bounced back. Greed pushed chip stocks too high, and when reality couldn't keep up they crashed.

Too scared at the bottom, too greedy at the top. This is what crowds do every single time.

So how do you avoid being the one holding the bag when the music stops? My answer is principles — a handful of standards you set in advance and hold no matter what the crowd is screaming.

One: we are investors, not speculators. We buy pieces of real businesses to own. We don't gamble on where a ticker might bounce next week. Chasing a hot chip stock because it's going up is speculating.

Two: every investment is worth the present value of the cash it can return to its owners over its remaining life. Pay far more than that and you've overpaid, no matter how exciting the story sounds.

Three: if we don't understand it, we don't invest in it. If you can't clearly explain how a business makes money and what it can earn over the next 5, 10, 15 years, you're guessing and hoping. Buffett keeps a "too hard" pile for a reason.

Four: in the short run the market is a voting machine; in the long run it's a weighing machine. Short-term prices swing on hype, emotion, and popularity. Over time they follow the real weight of the business. Peter Lynch's analogy nails it: Coca-Cola's earnings grew roughly 30x over a long stretch and the share price rose about 30x, while Bethlehem Steel's earnings declined significantly and the stock was cut in half.

Five: a great story becomes a bad investment if you pay the wrong price. That sentence explains the last two months entirely. AI is a real story. But saying the story is all that matters is the same as saying no price is too high — and I think we can all agree that's ludicrous.

Hold this momentum crash up against those five and every loss traces straight back to a specific principle that got broken.

Here's the beautiful flip side. If you actually follow them, a crash like this isn't a disaster — it's a gift. Moments of fear are exactly when wonderful companies finally go on sale. Those software stocks everyone was too scared to touch? That fear is what created the bargain.

Related reading: Adobe: value investing versus the 10x story, Intuit and the valuation reset AI fear created, and the sector rotation trap.

FAQ

Q: Should I sell chips and rotate into software now? A: I'd argue the question itself runs opposite to the lesson here. Chasing whatever is going up, late, is what caused the damage in the first place. The decision should hinge on what each individual company's price assumes, not on which sector is currently working.

Q: Will small caps keep beating large caps? A: I don't know. The Russell 2000's roughly 19% gain this year has far outpaced the Nasdaq's 5%, but a good deal of that may be short-term money shuffling. The confirmed fact isn't a directional forecast — it's that the expensive side got hurt far more.

Q: If chip companies are reporting great results, why aren't the stocks recovering? A: Because this is a positioning problem, not an earnings problem. One strategist warned that AI and tech positioning is far from cleared. When everyone who was going to buy has already bought, great results don't generate new buyers.