Micron Just Crossed $1 Trillion. Here's Why I Think This Is a Waypoint, Not a Ceiling
Today's +19% move in Micron is not the story. The story is what's underneath it, and I think most people are still framing this wrong.
Let me start with what actually happened. UBS analyst Timothy Arcuri raised his price target from USD 535 to USD 1,625 — a 204% increase in a single revision — and argued that "AI has permanently changed how Micron deserves to be valued." He made the direct comparison to NVIDIA's P/E, suggesting there's no structural reason MU should trade at a discount. That's an extraordinary claim from a sell-side desk that doesn't make calls like this lightly.
The data backs it up.
The Numbers I Can't Stop Looking At
Micron reported Q2 FY2026 revenue of USD 23.9 billion — a 196% increase year-over-year and 28% above their own guidance of USD 18.7 billion. Beating your own guidance by 28% is not a rounding error. It means demand is outrunning even the company's internal forecasts. EPS came in at USD 12.07, up 756% year-over-year.
Q3 guidance calls for USD 33.5 billion in revenue at approximately 81% gross margins. Twelve months ago, Micron's gross margins were in the mid-30s.
When I look at this data on Qualtrim, the free cash flow trajectory has shifted in a way I haven't seen in this company's history. This is what margin expansion driven by genuine pricing power looks like. Not financial engineering. Not one-time items. Structural demand that outpaces supply.
The Supply Constraint No One Can Fix Quickly
Here's what I keep coming back to. Micron's CEO Sanjay Mehrotra confirmed that the company's entire 2026 HBM4 supply is sold out under long-term, fixed-price contracts. They are currently fulfilling only 50 to 65% of their key customers' medium-term demand. Mizuho's analyst said there is "no clear line of sight" on when that demand-supply gap closes.
HBM (high-bandwidth memory) is not standard DRAM. You cannot repurpose a conventional memory fab to produce it. The manufacturing process involves stacking memory dies with precision tolerances that took years to achieve. Only three companies can do it at volume: Micron, SK Hynix, and Samsung. Samsung has been losing AI customer share due to yield issues. That leaves Micron and SK Hynix splitting demand that is outpacing their combined capacity.
This is the part of the thesis I find most compelling from a compounder perspective. True pricing power exists when customers need what you make, cannot get enough of it, and have no viable substitute. That describes HBM4 right now.
What This Means for My Thinking
I'm not suggesting buying at any price without doing your own analysis. At USD 1 trillion in market cap and a USD 1,625 sell-side target (implying roughly 80% upside from today's level, even after the +19% surge), the market is clearly pricing in a sustained AI capex cycle. If that cycle slows faster than expected, the margin story changes.
But the earnings are real. The supply constraint is real. The demand being left unfulfilled is real. When a business has more customers willing to pay than product it can sell, that is the most durable form of pricing power there is.
Micron at USD 1 trillion is a milestone. Based on the fundamentals I'm seeing, I think it's a waypoint.
What's your read — does the HBM supply constraint hold through 2027, or does Samsung solve its yield issues and change the picture?
The copyright of this article belongs to the original author/organization.
The views expressed herein are solely those of the author and do not reflect the stance of the platform. The content is intended for investment reference purposes only and shall not be considered as investment advice. Please contact us if you have any questions or suggestions regarding the content services provided by the platform.
