分析
Micron is set to report earnings later this month.
Based on its historical execution and the recent results from other memory companies, there is a very good chance Micron will deliver another outstanding quarter. But that is no longer the real issue.
The problem is that expectations have already been pushed to an extreme.
Once the quarter is in the books, the next-quarter guide could easily leave investors disappointed. And frankly, that should surprise no one. Long-term contract pricing has already locked in a large part of the margin upside, while much of the incremental capacity is still sitting on blueprints. At some point, exactly how much more acceleration can you squeeze out of revenue and earnings?
Meanwhile, you can almost hear the sell-side narrative machine warming up.
Model efficiency is improving. Training and inference costs are falling. Data-center demand growth could moderate. Stack those arguments together, and suddenly the new story becomes: this memory cycle is peaking.
Short it.
And then watch the market rip your face off on the reversal.
That is the setup heading into Micron's Q4 earnings.
Let us start with the previous quarter.
Micron reported revenue of $41.5 billion. Gross margin expanded by more than 10 percentage points sequentially and 46.5 percentage points year over year. Operating margin improved by nearly 13 percentage points quarter over quarter and 56.1 percentage points year over year. EPS nearly doubled sequentially and increased roughly 14-fold from a year earlier.
Objectively, those were exceptional results. There is no argument there.
But heading into fiscal Q4, the relevant question changes:
How high is the earnings bar Micron must clear for the stock to actually trade higher after the report?
Sell-side consensus currently calls for:
Adjusted EPS: $31.28
Revenue: $50.8 billion
Given the strength of AI-driven memory demand, Micron has a very credible chance of beating those numbers.
But again, that is not the point.
What matters for the stock is the second set of expectations—the expectations sitting inside the heads of buy-side investors.
Without access to proprietary buy-side survey data, one reasonable way to estimate that hidden hurdle is to use last quarter's percentage beat—not the absolute dollar beat—and extrapolate that percentage into Q4.
Last quarter, Micron beat adjusted EPS expectations by $4.42, ultimately reporting $25.11. Revenue beat consensus by $5.64 billion, coming in at $41.5 billion.
That translates into:
Revenue beat: 15.7%
EPS beat: 21.4%
Apply the same percentage upside to current fiscal Q4 consensus, and you get an implied market hurdle of roughly:
Revenue: $58.8 billion
Adjusted EPS: $37.96
In other words, Micron may need to get somewhere near those numbers simply to satisfy the market.
The exact hurdle is obviously debatable. Earnings outcomes are influenced by multiple variables, and expectations are never this mechanically precise.
But Micron is now one of the most closely watched stocks in the global market, and the expectations embedded in the share price are brutal.
Then comes the guidance issue.
Micron does not provide full-year guidance, so investors will focus heavily on its fiscal Q1 outlook.
Over the past three quarters, both revenue growth and EPS growth have continued to accelerate. The following figures represent the year-over-year growth rates implied by company guidance, with fiscal Q4 also based on management's guidance:
FQ1: Revenue +44%, EPS +109%
FQ2: Revenue +132%, EPS +440%
FQ3: Revenue +260%, EPS +903%
FQ4: Revenue +342%, EPS +923%
And those were only the guided numbers.
Actual reported revenue growth came in even stronger:
FQ1: +56.6%
FQ2: +196.3%
FQ3: +345.7%
During this shortage cycle, Micron has materially exceeded its own guidance quarter after quarter.
That is precisely how the base became so demanding.
The trend is obvious: every quarter's outlook has been stronger than the one before it, while earnings have exploded higher throughout the year.
Continuing to accelerate against those increasingly difficult comparisons is becoming mathematically difficult.
The market has become conditioned to an uninterrupted sequence of accelerating positive data. If the fiscal Q1 guide begins to show even a modest deceleration at the margin, the stock could come under pressure.
The absolute level of fiscal Q1 guidance may still look perfectly respectable. But the reality is that the major revenue ramp only began in fiscal Q2.
A slowdown in reported growth rates is therefore becoming almost unavoidable.
Whether some investors want to acknowledge it or not, Micron is still a cyclical company.
The current level of extraordinary profitability is being driven by aggressive data-center channel demand combined with extremely tight inventories.
Absolute inventory fell from $8.355 billion at the end of FY25 to $8.205 billion in FY26 Q1, effectively showing no growth.
Over the same period, quarterly revenue increased from $11.315 billion to $13.643 billion.
As a result, inventory as a percentage of quarterly revenue compressed from 73.8% to 60.1%.
That type of supply-demand imbalance does not last forever.
Periods this strong typically persist for only a handful of quarters.
Now return to the upcoming quarter.
Sequential revenue growth is expected to slow to 20.6%.
The second derivative is already weakening, and that is not a positive setup for the stock.
Even if Micron reaches the implied market expectation of $58.8 billion in revenue, sequential growth would still decelerate.
That would suggest the broader growth trajectory is beginning to cool materially.
So let us make a few assumptions.
Suppose Micron reports $56 billion in revenue.
That would represent a very large beat versus sell-side consensus, but it would still fall short of the market's more euphoric implied hurdle.
To reach that higher expectation, DRAM and NAND pricing would likely need to rise by an unusually large amount sequentially.
That is a difficult ask.
Now assume gross margin reaches 86.5%, or 50 basis points above company guidance.
Margins are already operating at exceptionally high levels. The room for another enormous upside surprise is becoming increasingly limited, and there is only so much additional cost efficiency management can extract.
Gross margin began rising sharply from last year's Q4, when it increased 965 basis points year over year to 44.97%.
Within roughly one year, the margin profile has nearly doubled.
The operating performance has already been extraordinary.
On operating expenses, the key swing factor is R&D.
R&D has consistently grown more slowly than revenue, creating enormous operating leverage.
Management is guiding to approximately $1.65 billion in operating expenses, including R&D, this quarter.
The absolute expense number is not changing materially, while operating expenses as a percentage of revenue should fall modestly to below 3%.
But that leverage works both ways.
When the cycle turns, R&D as a percentage of revenue can rise very quickly, turning it into a meaningful pressure point on profitability.
SG&A expense is expected to increase approximately 15% sequentially. Other expenses are assumed to remain flat versus Q3 at $15 million.
Based on management's guidance, I therefore assume total operating expenses of $1.65 billion.
Excluding nonrecurring items and applying a 15% tax rate, I estimate adjusted EPS of approximately $34.73 for the quarter.
That would be 11.0% above sell-side consensus of $31.28.
On the surface, that is an extremely strong quarter.
But under today's expectation regime, it may still not be enough to push the stock higher.
If Micron wants to support the share price, management may need to become substantially more aggressive with capital returns.
Consider SK Hynix's recently announced shareholder-return framework.
For the 2025-2027 period, the company plans to return 50% of cumulative free cash flow to shareholders.
That is an exceptionally shareholder-friendly policy.
Micron looks far less aggressive by comparison.
Annual dividends remain below $1 billion. With a quarterly dividend of $0.15 per share, annualized dividend payments are approximately $690 million.
Share repurchases have effectively stalled.
During the first three quarters of FY26, Micron repurchased only about $686 million of stock, including just approximately $420,000 in fiscal Q3.
Against roughly $26.1 billion of free cash flow generated year to date, that level of shareholder return is difficult to defend.
Investors have every right to demand more aggressive dividends and buybacks.
There is another path: Micron could follow SK Hynix and deploy more capital into capex and R&D.
But that creates another problem.
If the company materially increases investment next year just as the memory cycle begins weakening, gross margins could erode rapidly over the following several quarters.
The current bull case argues that Micron deserves a structurally higher valuation multiple because ROIC is improving and AI-driven memory demand is stronger than in previous cycles.
There is some truth to that.
AI is clearly creating a larger incremental revenue opportunity than prior technology cycles.
But the market is ignoring one critical fact:
The memory industry's current supply shortage will eventually be resolved.
Consider peer valuations as of early September 2026, using trailing-twelve-month metrics:
Samsung Electronics:
P/E of approximately 11.2x
EV/EBITDA of approximately 6.3x
SK Hynix:
P/E of approximately 7.4x-8.2x
EV/EBITDA of approximately 8.2x
SanDisk:
P/E of approximately 21x
EV/EBITDA of approximately 17x-18x
Micron:
P/E of approximately 21.6x
EV/EBITDA of approximately 15.5x
Micron's valuation premium relative to SK Hynix is difficult to justify.
That premium becomes even harder to defend when cash flow has improved dramatically while the company's capital-allocation framework remains relatively unconvincing.
Once the earnings-upgrade cycle ends and additional industry supply gradually enters the market, the valuation equation changes dramatically.
Consensus currently implies FY2028 EPS of $18.56, down roughly 75% from approximately $73 in FY26E EPS.
Apply an 11x normalized cyclical P/E multiple, and the resulting target price is approximately $204 per share.
Against a current share price of $958, that implies approximately 79% downside.
There are important risks to this thesis.
If token costs stop declining—or actually begin rising again—ASP support across AI infrastructure products could prove stronger than expected.
More broadly, the biggest risk to everything discussed above is the duration of the cycle.
AI technology and application use cases are evolving at extraordinary speed.
If Micron's current upcycle lasts longer than I expect, the stock could remain strong for considerably longer before ultimately peaking.
My base-case assumption is that the memory cycle tops sometime within the next 12 months.
If the cycle extends materially beyond that window, Micron's share-price performance could be substantially stronger than my current forecast.
The market's disagreement over Micron has become increasingly obvious over the past month.
You can see it directly in the stock's volatility.
During July 2026 alone, Micron traded between $737.88 and $1,096.78, a range of approximately 48%.
Investors are increasingly debating the underlying variables that matter most: AI demand, token economics, training efficiency, inference costs, and the durability of the current infrastructure spending cycle.
Micron's upcoming earnings report could very well be spectacular.
But a spectacular quarter is not necessarily a bullish catalyst for the stock.
That distinction is the entire trade.