From our debut pick to a $1.2 trillion company: the earnings avalanche, the sold-out HBM order book, and the next leg.
Micron Technology (MU) was the first name in the first edition of StockResearch Intelligence Weekly, published January 20, 2026. We rated it STRONG BUY on the six-month and one-year horizons and built the thesis on three pillars: the HBM memory bottleneck, accelerating earnings momentum, and a margin inflection as the business shifted from commodity DRAM to high-value memory. All three played out. The stock has since gone from roughly $365 to about $1,064 — a gain of around 192% — and crossed a $1.2 trillion market capitalization, joining a club that previously had no memory company in it.
This report does three things: documents the call, isolates the catalysts that actually drove the move, and lays out where the risk-reward sits now that the easy money has been made.
01CHAPTER 01The Call
The setup nobody was crowding into was, in hindsight, hiding in plain sight. In mid-January 2026, memory was still widely treated as a late-cycle commodity trade. The consensus framing was that DRAM pricing had run, that the AI build-out would mostly benefit logic and GPUs, and that Micron was a cyclical that had already moved.
We disagreed, and said so in print. In Edition #1, Micron led the Top 3 with a clear directional view: the binding constraint on AI compute was no longer just the GPU — it was the memory feeding it. High-bandwidth memory was the bottleneck, and Micron's position in HBM3E made it a supplier hyperscalers could not route around. We paired that with two corroborating signals: a sharp upward revision cycle in earnings estimates, and a margin mix shifting decisively toward high-value product.
The stock was trading around $365 the week the edition published. Our one-year target was $360-plus. With the benefit of the move since, that target reads as far too conservative — it was effectively achieved within the first week. The thesis was right; the ceiling was higher than we, or anyone on the Street, dared to model.
02CHAPTER 02The Scorecard
The horizon ratings we published — HOLD for one week, BUY at one month, STRONG BUY at six months and one year — all resolved correctly, and quickly. The price targets, drafted on a high-$200s basis from late-December data, were overtaken almost immediately as the stock gapped through them.
We flag this openly because it is the more useful lesson than a clean victory lap would be: the alpha was in the thesis, not the target. Calling the direction and the magnitude of the structural shift — memory becoming a strategic asset — was worth far more than the specific number we attached to it. A correct framework that you then under-target still compounds; a precise target on a wrong framework does not.
03CHAPTER 03Catalyst #1 — The Earnings Avalanche
The run was not a sentiment melt-up. It was a sequence of financial results that repeatedly broke the model.
The fiscal Q1 report that underpinned our original call already showed the inflection: $13.64 billion in revenue, up 57% year-on-year, with non-GAAP EPS of $4.78 beating consensus by roughly 27%. Strong, but the market was still pricing Micron as a cyclical.
Then came fiscal Q2, reported March 18. Revenue landed at $23.86 billion — up 196% year-on-year and 75% sequentially, and well above Micron's own guidance of around $18.7 billion. Non-GAAP EPS of $12.20 beat consensus by nearly 39%. GAAP gross margin reached 74.4%, up from 36.8% a year earlier. Operating margin hit 69%. Free cash flow set a quarterly record at $6.9 billion, and the board raised the dividend 30%.
The guidance was the part that re-rated the stock. Management guided fiscal Q3 to a record $33.5 billion in revenue at approximately 81% gross margin and $19.15 in EPS — figures that would have been dismissed as fantasy for a memory company two years ago. Each report did not just beat; it reset the ceiling on what the business could earn.
04CHAPTER 04Catalyst #2 — Memory Stops Being a Commodity
The single most important structural development is that memory began behaving like a scarce asset rather than a commodity.
The proof is in the composition of the growth. Contract DRAM prices rose roughly 15–20% quarter-on-quarter, with enterprise SSD pricing up over 25% in some segments, while bit shipments were flat to slightly down. When revenue climbs almost entirely on price with volume static, you are looking at a supply-constrained market — the opposite of the oversupply gluts that historically crushed memory margins.
Two things make this more than a cyclical price spike. First, HBM capacity is sold out through all of 2026 under multi-year contracts, including next-generation HBM4. Volume production of HBM4 for Nvidia's Vera Rubin platform has already begun, with HBM4E set to ramp in 2027 and custom HBM slated for Nvidia's Feynman generation in 2028. Second, customers are signing three-to-five-year supply agreements — Micron's first five-year Strategic Customer Agreement landed this cycle.
05CHAPTER 05Catalyst #3 — The Trillion-Dollar Re-Rating
As the earnings reset, the Street's price targets chased the stock upward in a way that is rare to watch in real time. In late 2025 and January 2026, the bullish targets clustered around $450–500 — Barclays lifted to $450, Rosenblatt set a then-street-high of $500. Those numbers now look quaint.
By late May, Micron joined the trillion-dollar market-cap club alongside SK Hynix. UBS raised its target from $535 to $1,625, a move that helped trigger a roughly 19% single-day surge. Susquehanna went to $1,750, and Morgan Stanley to $1,050. The re-rating was not just price; it was the market reclassifying what kind of company Micron is — from a memory cyclical to a scarce AI-infrastructure resource.
The same chart shows the risk. The bull–bear spread is now extraordinarily wide. Against the cluster above $1,000 sits a bearish model at roughly $476 with a SELL rating, arguing that a memory cyclical trading at a high trailing multiple after a near-tenfold one-year run does not reconcile mathematically. That dispersion is itself information: conviction is high on both sides, which tends to mean volatility ahead.
06CHAPTER 06The Bigger Picture — Scarcity as a Feature
The most counterintuitive part of the story is that Micron's capacity expansion does almost nothing for 2026 — and that is bullish, not bearish, for the near term.
The new fabs are real but distant. Meaningful Idaho output arrives around mid-2027. The $1.8 billion Tongluo acquisition in Taiwan contributes no wafers until the second half of 2027. Singapore NAND first wafer-outs are a 2028 event. The New York mega-complex is a 2030 story. For the entirety of 2026 and into early 2027, Micron is operating within fixed capacity that cannot be meaningfully expanded.
That is the structural reason 80%-handle gross margins and locked pricing look durable through the year: there is no new supply to break the constraint. The flip side is the bear's strongest point — that wall of capacity does eventually arrive, and in a commodity industry, new supply has historically been the thing that ends every up-cycle. Management frames the HBM total addressable market expanding from roughly $35 billion in 2025 toward $100 billion by 2028 — about a 40% annual growth rate — which is the demand curve that would have to hold for the supply additions to be absorbed rather than to trigger a glut.
The quarter that reset the ceiling
Fiscal Q3 guidance ($33.5B revenue, ~81% margin, $19.15 EPS) is management guidance, not a reported result.
07CHAPTER 07Valuation & Technical Structure
The valuation debate is unusually clean because it reduces to a single question: which earnings number do you capitalize?
On trailing earnings, MU trades around 51x — expensive, and the bear's anchor. On next-twelve-month earnings, it trades around 11x, and on FY2027 estimates it falls below 10x — cheap for a company consensus expects to earn around $58 per share in FY2026 on revenue that could exceed $120 billion, against roughly $37 billion in FY2025. The entire bull-bear gap is whether you believe the forward number is real and repeatable, or a cyclical peak that will revert.
Technically, the stock closed June 2 at an all-time high of $1,064.10 and printed an intraday high of $1,089.29 on June 3, with a $1.2 trillion market cap and a beta above 2. There is no overhead resistance — by definition, at all-time highs — but also no technical support nearby to lean on. For a name this extended into a binary catalyst, position sizing matters more than entry precision.
08CHAPTER 08What Comes Next
The next hard catalyst is fiscal Q3 earnings on June 24, covering the period ending around May 31. Management has guided to $33.5 billion in revenue, ~81% gross margin, and $19.15 EPS. Given the pattern of beat-and-raise, the bar is now “beat the guide and raise again.” A clean beat-and-raise could kick off another leg; an in-line print that fails to raise would likely be treated as a disappointment at this multiple.
We frame the twelve-month outlook in three probability-weighted scenarios, set with the humility that our last target was too low.
HBM4 ramps on schedule, pricing holds through 2027, contracts prove sticky, and June 24 delivers another beat-and-raise. The UBS/Susquehanna zone.
Pricing plateaus at a high level, FY2027 stays strong but decelerates, and the multiple settles between cyclical and secular.
The cycle peaks in late 2026 or 2027, CXMT and incremental supply pressure pricing, AI capex digestion slows hyperscaler memory orders, and the multiple compresses back toward cyclical norms — a 35–55% drawdown from current levels.
Key dates: June 24 (fiscal Q3 results), late September (fiscal Q4 and the additional work-week in this 53-week fiscal year), and the HBM4E ramp through 2027.
09CHAPTER 09Risks
- Cyclicality has not been abolished. The bull case is explicitly that AI changes the structure of the memory cycle. If that is wrong, the reversion is severe and the trailing multiple offers no protection.
- Chinese competition. CXMT and peers are advancing; incremental Chinese supply is the most credible path to breaking the current pricing constraint.
- Capex consuming cash. Funding fabs across three continents while at peak earnings leaves trailing free-cash-flow margins thin.
- AI capex digestion. The entire demand stack depends on hyperscalers continuing to spend. Any pause on ROI concerns would hit memory demand first.
- Valuation and concentration. At ~51x trailing earnings and a $1.2 trillion cap, expectations are high and the shareholder base is crowded with momentum capital.
- Single-catalyst exposure. Near-term direction is heavily tied to one print on June 24 in a name with a beta above 2.
The thesis is intact but the entry is no longer asymmetric. A position can be justified on the forward multiple and the sold-out order book, but sizing should respect that you are buying after a near-triple, into a binary earnings print on June 24, with a beta above 2. Scaling in rather than chasing a single price is the disciplined approach. Existing holders are sitting on a rare gain — consider trimming toward a target weight while retaining a core position for the HBM4/HBM4E ramp into 2027.