Chip-stock revival plus strong earnings.
The VanEck Semiconductor ETF (SMH) jumped roughly 4%, with Micron ~+10%, Intel ~+7%, Marvell ~+7% and Astera Labs ~+6%
A rebound after the sector's recent wobble and ahead of a heavy week of tech results.
On the earnings side, FactSet noted ~88% of the 66 S&P 500 companies that had reported so far beat profit estimates, which kept sentiment firm.
RIP Canada
Trump announced 50% tariffs on a range of Canadian goods (beer, dairy, chemicals; oil excluded) set to take effect in ~30 days.
Notable movers
- 3M: ~+9% beat on EPS ($2.40 adj.) and sales, and raised full-year profit guidance. A Dow standout.
- General Motors posted record first-half results and beat estimates ($3.57 adj. EPS on $48.0B revenue).
- Danaher: ~−13.6% beat, but flagged softer-than-expected bioprocessing revenue.
- Equifax: ~−7.2% narrow beat undercut by weak forward guidance.
Too Late to Get Into Micron (MU)?
Micron has gone from a boom-and-bust commodity memory maker to the single most direct beneficiary of AI memory scarcity and the numbers (84% gross margins, revenue up 346%) are so far beyond anything memory has ever produced.
What the company actually does
Micron makes memory and storage chips — the two kinds that matter here:
1) DRAM (the fast, working memory).
This is the chip that holds data a processor is actively using right now.
Real-world way to picture it: if a GPU is a chef cooking as fast as possible, DRAM is the countertop space next to the stove , the bigger and faster it is, the more ingredients the chef can reach without running back to the pantry.
AI accelerators (Nvidia-class GPUs and custom chips) are so fast that ordinary DRAM can't feed them quickly enough, which created the product that's driving this whole story:
There’s a semiconductor company that supplies both Apple and SpaceX … that’s absolutely ON FIRE here in 2026.
The Wall Street Journal just reported in June that this company DOUBLED its data center revenue forecast to $1 BILLION (up from its prior forecast of $500M). “
Surging demand for the building blocks of AI has led to a memory chip shortage and fueled a global rally in semiconductor stocks
2) HBM (High-Bandwidth Memory) the crown jewel.
HBM is DRAM stacked vertically into a tower and wired directly next to the AI chip, so data moves in a firehose instead of a garden hose.
Every high-end AI accelerator needs a stack of it, and there are only three companies on earth who can make it well: Micron, SK Hynix, and Samsung.
Micron's release says HBM4 (its newest generation) is already "in high-volume shipments for our lead customer's platform" — i.e., it's designed into the flagship AI systems shipping now.
Because HBM is hard to make and demand is exploding, it sells at far higher margins than commodity memory, which is the mechanical reason gross margins have gone vertical.
3) NAND flash (storage).
The chip that remembers data when the power is off (SSDs).
Smaller part of the AI story, but AI data centers need enormous storage too.
Micron has always made these products.
What changed is that HBM turned memory from an interchangeable commodity, where the only question was price into a scarce, spec'd-in component that hyperscalers are fighting to lock up.
That's why a company that earned a few billion a year is now printing $28B of net income in a single quarter!
Memory is the most violently cyclical corner of semiconductors.
Every prior up-cycle (2017–18, 2021) ended with a vicious glut, collapsing prices, and losses.
Bulls argue "HBM + AI is different it's spec'd in, capacity is constrained, this is durable."
Bears argue "it's always different until it isn't, and 84% margins are a flashing red 'peak' sign that pulls in competitor capacity."
You cannot own this name without picking a side.
At ~$972, Micron trades at ~19.5x trailing earnings but only ~6x forward earnings.
The "Super Chip" Powering 99% of ______ It's not NVIDIA. And it's not Intel.
But this little-known stock's designs already power 99% of a critical technology. Now it's expanding into AI chips for data centers, self-driving cars, and robotics.
Big Tech is racing to license its architecture. Governments are pushing to secure it.
And Wall Street still hasn't caught on. Click here to see the full report >>>
Analysts just cannot agree... one analyst target sits at $361, ~63% below today's price, while another has it at $2,200.
FWIW, I bought some in June at $1080 ish.. and I'm looking to add some more... sure it's high risk, but I don't think I can fully participate in the AI movement without owning some.
Editor’s Note: Ignore the headlines about an “AI bubble.”
According to Silicon Valley insider Jeff Brown — the man who called NVIDIA before it rocketed 28,000% — we’re only at the foothills of the next big AI boom.
But this time, Jeff says the biggest winner won’t be a chipmaker…
It’ll be a company producing something he calls “AI Fuel.”
Click to see what it is — and get the ticker.
Best
-Greg
Disclaimer:
This content is for informational and educational purposes only and does not constitute financial, investment, or trading advice. All opinions expressed are based on publicly available data believed to be reliable but are not guaranteed for accuracy or completeness. Investing and trading involve risk, including the possible loss of principal. Always perform your own due diligence and consult with a qualified financial advisor before making investment decisions. Past performance is not indicative of future results. Paid ads from 3rd parties are clearly identified. We do NOT specifically endorse these products nor are we responsible for the content of these ads. You should assume an affiliate payment will be made to us if you click a link or buy a product.