Micron Technology reports earnings on Wednesday evening.
The stock is up 293% so far this year. The market cap is sitting just above $1.28 trillion. Wall Street expects roughly $20 in EPS and $34.5 billion in revenue for the quarter. A year ago, the consensus EPS estimate for this exact same quarter was under $2.
None of those numbers are the one that matters Wednesday night.
What Micron Actually Makes
Micron makes memory chips. Specifically, it makes DRAM, NAND flash storage, and a specialized product called HBM (high-bandwidth memory) that sits directly on top of AI processors and feeds them data at extraordinary speed.
Think of it this way. NVIDIA’s Blackwell chip is the engine. HBM is the fuel injection system. You can have the most powerful engine in the world and it still runs exactly as fast as the memory can feed it. No faster.
Every Major AI Company On Earth Is Short On Memory Right Now
At a JPMorgan conference last month, Micron’s Head of Global Operations said something that did not get nearly enough attention.
He said Micron’s key customers can only secure 50% to two-thirds of their bit demand requirements in the medium term.
Read that again. The biggest AI companies in the world are placing orders with Micron and receiving back less than two-thirds of what they asked for. Not because Micron is holding anything back. Because the entire industry genuinely cannot build fast enough to match demand.
The reason is structural. Building a new memory fab takes 3 to 4 years. You pour concrete, install cleanrooms, bring in the tools, qualify the process, ramp the yield. Micron has broken ground on new fabs in Idaho, Taiwan, New York, and Singapore. The earliest any of them produce meaningful output is mid-2027. Most won’t matter until 2028. Meanwhile, every AI data center in the world is ordering more memory than the industry can ship.
The HBM Trade Ratio Is The Thing Nobody Is Talking About
There is a structural wrinkle that makes the supply picture tighter than most people realise.
Making HBM is fundamentally different from making regular DRAM. Micron’s operations chief explained it directly in May: it takes more than three times as many wafers to produce the same number of bits of HBM compared to regular memory. HBM die sizes are larger. The stacking process is more complex. More wafers, more cleanroom time, more of everything. Just to make one unit of HBM.
And this ratio gets worse with each generation. HBM4 requires more wafers per bit than HBM3E. HBM4E will require more than HBM4.
So as AI demand for HBM grows and customers upgrade to newer chip generations, the pressure on the global wafer supply multiplies automatically. The whole industry is running on a treadmill that accelerates every 18 months. Micron told JPMorgan flat out: they expect supply to remain tight well beyond calendar year 2026.
What To Actually Watch Wednesday Night
Analysts are focused on the EPS and revenue numbers. Those will almost certainly beat consensus. In May, Micron told JPMorgan it was tracking for another record free cash flow quarter in Q3. That was stated with a month still left in the quarter.
The number that actually moves the stock is the FY2027 guidance.
Micron’s fiscal year runs to August. Q3 covers roughly March through May. What management says about full-year 2027 expectations is what traders will key on. Analyst consensus for FY2027 EPS is around $97, which is another 69% growth year on top of FY2026’s already staggering 651% EPS growth. If Micron guides in line with or above that, the stock sees new highs. If they introduce any language about demand softening or customer inventory building, expect selling regardless of how clean Wednesday’s numbers look.
Also watch the Strategic Customer Agreements. At the March earnings call, Micron announced its first-ever 5-year SCA with a large customer. A multi-year deal covering both volume and pricing commitments. At the JPMorgan conference in May, management said they had made “meaningful progress” on additional SCAs but did not disclose how many.
Wednesday’s call will answer that. The number of signed SCAs tells you more about the durability of this business than any quarterly figure. A 5-year supply contract at locked-in pricing means Micron knows exactly what it will ship and at what margin for the next half decade. That is not a cyclical business. That is something structurally different from every prior memory upcycle.
The Risk Nobody Wants To Say Out Loud
The stock is up 293% in 2026. The market cap is $1.28 trillion.
Micron has historically been a boom-and-bust company. Memory prices surge, then they crash, then they surge again. The bull case today rests entirely on the argument that this time is genuinely different because AI demand is structural, not cyclical.
That may be true. The evidence from the last 18 months is compelling. Management said in May explicitly: they see no ability for supply to catch up to demand for the foreseeable future. The HBM trade ratio, the greenfield fab timelines, the SCA structure. All of it points toward a supply constraint that cannot be fixed quickly.
But “foreseeable future” is not forever. And the analyst estimate range tells you how uncertain even the experts are. EPS estimates for Wednesday range from $7.53 at the low end to $24.08 at the high end. That is not a normal range for a $1.28 trillion company. It reflects genuine uncertainty about how much HBM pricing held, how much of the SCA upside materialised, and how far margins actually stretched.
One more thing to note. Micron guided for 81% gross margins in Q3 back in March. If they hit that number on Wednesday, it would be among the highest gross margins ever recorded for a semiconductor company of this size. That alone would be a historic data point.
What This Means For The LENS Index
The LENS Index launched on May 22nd running a concentrated portfolio of large-cap companies where the dominant market narrative is provably wrong. The AI memory narrative two months ago was skeptical. The thesis was that supply would catch up, pricing would soften, and the memory cycle would turn.
The data out of Micron’s own calls said something different. And the portfolio has reflected that.
| Metric | Value | Notes | Last Updated |
|---|---|---|---|
| Active Positions | 6 / 30 | Core / Base / Speculative tiers | 18-Jun-26 |
| Cash Deployed | 31% | Full positions visible to subscribers only | 18-Jun-26 |
| Cash Reserve | 69% | Minimum 15% cash reserve maintained at all times | 18-Jun-26 |
| Performance Since Launch | +5.4% | LENS time-weighted index return since May 22 | S&P: +1.0% | 18-Jun-26 |
| Alpha Generated | +4.5% | vs S&P 500 Total Return from inception (May 22, 2026) | 18-Jun-26 |
| Portfolio Beta | 0.58 | vs S&P 500 | 19 daily observations (May 22 – Jun 18) | 18-Jun-26 |
| Max Drawdown | -2.5% | Jun 1 peak (104.81) to Jun 5 trough (102.84) | 18-Jun-26 |
| Sharpe Ratio (Prelim.) | 4.05 | Annualized, 18 obs, rf=3.5%. Full Sharpe from Q4 2026 | 18-Jun-26 |
| Inception Date | May 22, 2026 | DELL pre-mortem published. Warsh sworn in same day. | Fixed |
| Benchmark | S&P 500 Total Return | Outperform on a risk-adjusted basis | Fixed |
| Strategy | Narrative Disruption | Long-only, large-cap, max 30 positions | Fixed |
| Direction | Long Only | Subscribe to access all positions and targets | Fixed |
The LENS return versus the S&P 500 since inception, every active position, and the exact thesis break conditions for each one are above. Wednesday evening at 4 PM Eastern is the single largest near-term catalyst event in the portfolio.
Set the alarm.
Disclaimer: We do not hold any positions in the above stock(s). Read our full disclaimer here.
