Micron Technology: A Deep-Dive Value-Investing Analysis

May 20, 2026·~19 min read

Quick ReadPlain-language overview · read this first

Micron is the world's third-largest DRAM maker, with a 25.7% share; HBM4 is in volume production for NVIDIA's Vera Rubin; the current price is $698.74; rating Watch.

In fiscal 2025 CMBU accounted for 36% of revenue with a 45% operating margin; TTM revenue was $58.1 billion, non-GAAP gross margin 74.9%, and net cash $6.5 billion. But the memory industry's strong cyclicality and capital intensity are unchanged in essence: most contracts are short-term fixed-price, FY2023 gross margin fell as low as -9.1% with free cash flow of -$6.1 billion, FY2026 capex exceeds $25 billion, and one customer accounts for 17%; the current P/E of 32.7, EV/EBITDA of 21.2, and P/FCF of 76.7 already price in a long-lasting AI memory supercycle.

Owner Earnings across three scenarios give $120-180 / $230-320 / $400-520; the current price carries an obvious premium to the base range. Ideal buy $160-240, corresponding to a 25%-35% margin of safety; if the industry returns to oversupply and the valuation multiple reverts, a permanent drawdown of 50%-70% is not unimaginable. A good company, not a good price.

Lead
Micron's TTM revenue surged to $58.1 billion on AI/HBM demand, with $23.86 billion in fiscal Q2 alone, and HBM4 is in volume production for NVIDIA's Vera Rubin; but the memory industry's strong cyclicality and capital intensity are unchanged and most contracts are short-term fixed-price, so at roughly $699 the stock sits far above the $120–520 intrinsic-value range with no margin of safety. Rating: Watch.

Full report

Prices in the article are as of publication; see the valuation band above for the live price.

Conclusion First

Preliminary rating: Watch. Does the current price offer a margin of safety: No. Suitable investor type: Better suited to sector/cyclical investors who can tolerate strong cycles and technology-transition risk, rather than ordinary long-term value investors seeking steady compounding. Biggest uncertainties: First, whether the high profitability driven by AI/HBM is merely another upcycle rather than a "permanent improvement"; second, whether the enormous capital expenditure of the next two to three years can convert into sustained free cash flow rather than the next round of oversupply; third, whether the current market capitalization has already priced in the optimistic narrative that "Micron has become a near-permanently high-earning AI memory leader."

Core judgment. Micron is a business I can understand, and one that is higher quality today than in the past: it sells DRAM, NAND, and NOR, along with the HBM, SSDs, Managed NAND, and other products derived from these chips; the long-term direction of demand is sound, especially as AI data centers markedly raise the importance of high-bandwidth, high-capacity memory. The question is not "whether the company is a good company," but "whether a good company has already been bought away by the market at an extremely high price that captures most of the next decade's surprises." At the current quote, a rough calculation on the roughly 1.128 billion shares outstanding at the end of fiscal Q2 2026 puts Micron's equity value at about $788.2 billion; the implied return at this price is not conservative for a memory maker characterized by strong cyclicality, heavy capital intensity, and weak recurring revenue.

One-sentence conclusion. If I view Micron as a business to acquire and hold for the long term, I am willing to track this company for years, but I am unwilling to "bet big" at the current price that it has already escaped the historical cyclical law of the memory industry.

Note on method and definitions. This report tries to distinguish four types of information: facts are drawn mainly from the company's 10-K, 10-Q, proxy statement, quarterly materials, Reuters, TrendForce, and the U.S. Treasury/FRED; assumptions appear mainly in the Owner Earnings and DCF valuation; inferences appear mainly in the assessment of the moat and capital allocation; opinions are expressed in the final rating. Where a given peer valuation metric cannot be fully obtained from official disclosures on the same date and same accounting basis, I state the limitation explicitly rather than fabricating precise figures.

The Business, Industry, and Competitive Landscape

How this company actually makes money. Micron's core is selling memory and storage. Viewed through the four business units after the fiscal 2025 reorganization: CMBU serves large cloud customers and all data-center HBM; CDBU serves mid-size cloud, enterprise/OEM data centers, and all data-center storage; MCBU serves mobile and client; and AEBU serves automotive, industrial, and consumer. At the product level, the company focuses on DRAM, LPDDR, GDDR, HBM, NAND, SSDs, Managed NAND, and NOR. In fiscal 2025, CMBU revenue was $13.524 billion, 36% of company revenue, with a 45% operating margin; this shows the profit pool has already tilted decisively toward cloud data centers and HBM, rather than the traditional PC/phone memory logic.

Who the customers are, and how it charges. Customers are essentially large cloud vendors, enterprise data-center customers, phone makers, PC/OEMs, and automotive and industrial equipment customers. Micron discloses in its 10-Q that revenue is "recognized primarily at the point in time when control transfers," and that "the vast majority of customer contracts are short-term, fixed-price contracts"; this means it is not a subscription or recurring-revenue model, but a typical product-based business whose prices are highly exposed to industry supply and demand. In fiscal 2025 one customer contributed 17% of revenue; in the first six months of fiscal 2026 one customer contributed 13% of revenue, mostly in CMBU. In 2025, roughly 80% of the company's revenue came from shipments to destinations outside the United States, so operations are highly globalized and, by nature, more exposed to geopolitics, export restrictions, and customer concentration.

Is revenue recurring, stable, and predictable. Long-term demand rises steadily, but short-term revenue is not stable. In its fiscal Q2 2026 prepared remarks, Micron emphasized that AI demand is driving a large increase in data-center DRAM and NAND demand, and the company expects the DRAM and NAND supply-demand tightness to persist beyond 2026; TrendForce also disclosed that in Q3 2025 global DRAM industry revenue surged 30.9%, inventories were nearly depleted, and prices rose rapidly. The catch is that this kind of "tightness" usually means precisely that the industry is near a profit peak, not in a stable steady state. In other words, Micron's demand looks good over the long term, but its revenue and gross margin lack the Coca-Cola-style linear predictability.

Cost structure, dependencies, and understandability. The cost core of this business is fabs, equipment, node migration, packaging and test, and R&D; the company explicitly writes that its manufacturing network spans Taiwan, Singapore, Japan, the United States, Malaysia, China, and India, that 300mm fabs typically run 24/7, and that semiconductor manufacturing is highly capital-intensive. The company expects fiscal 2026 capital expenditure to exceed $25 billion, with a further marked increase in 2027, driven mainly by cleanrooms, HBM/DRAM-related investment, and new capacity build-out in the U.S. and Asia. For exactly this reason, Micron is not a "lighter as it sells more" software business, but a "the stronger it gets, the more it must invest" process-manufacturing business. On understandability, I give it 4/5: the products and industry structure are not complex, but the earnings rhythm depends heavily on supply and demand, technology generations, and the capex cycle. If the stock market were closed for five years, I would be willing to hold the Micron business itself, but only if the purchase price were meaningfully below intrinsic value; at the current price, I am not at ease.

Industry and competitive landscape. The memory industry is not a declining industry, but a "long-term growth, strong short-to-medium-term cycle" industry. AI, cloud, servers, automotive electronics, and edge computing keep pushing up the memory content per device; but capacity expansion and technology migration on the supply side periodically create oversupply and shortage. On industry structure, DRAM is a highly concentrated oligopoly. TrendForce disclosed that in Q3 2025 SK hynix led with a 33.2% share, Samsung had 32.6%, and Micron 25.7%, making Micron the clear third-largest DRAM maker; NAND is more fragmented, with Samsung at 32.3% in Q3 2025, SK hynix 19.3%, Kioxia 15.3%, and SanDisk 12.4%, and Micron trailing Kioxia, so its position in NAND is weaker than in DRAM. My qualitative judgment is that Micron is more of an "excellent company in a hard industry" than an "asset-light king of a naturally good industry." Industry attractiveness score: 3/5.

Moat and Management

Moat assessment. Broken down along a Buffett-style framework, Micron's brand advantage is weak-to-moderate, because end consumers buy devices rather than bare DRAM dies; its cost advantage and scale advantage are moderate-to-strong, because very few players can stay in the leading DRAM/NAND competitive race over the long run, and the barrier to entry is capital on the scale of tens of billions of dollars plus years of process accumulation; network effects are essentially absent; switching costs are moderate, and higher in segments such as servers, automotive-grade, and enterprise SSDs that require qualification, certification, and long-term supply; patent/process/manufacturing know-how barriers are moderate-to-strong, especially at advanced nodes and advanced packaging such as 1γ DRAM, G8/G9 NAND, and HBM4; channel advantage is average; and corporate culture and operational capability have improved markedly in recent years, especially in the cadence of data-center and HBM products. On balance, I give Micron a moat strength of 3/5: it is not that there is no moat, but that the moat comes mainly from "a handful of survivors plus process/capital barriers," not from strong pricing power.

Is the moat widening, stable, or narrowing. My judgment is: in the HBM and cloud data-center sub-markets, Micron's moat has been widening over the past two years; in traditional commodity DRAM/NAND, the moat is broadly stable but not wide. The evidence is that fiscal 2025 CMBU revenue already reached $13.524 billion, with margins far above the other businesses; meanwhile, in 2026 the company has begun volume production of HBM4 for NVIDIA's Vera Rubin, and expects the mature-yield ramp to be faster than HBM3E. On the other hand, the industry's essence is still supply-demand-driven and price-driven, and Micron itself discloses that the vast majority of contracts are short-term fixed-price, which means that even when the technology is done right, the cycle cannot be fully smoothed away. For competitors to replicate this would require enormous capital, advanced process, packaging capability, and customer qualification, measured in years at least; but Samsung and SK hynix are themselves ready-made super-rivals, so Micron's moat is real, yet not exclusive.

Pricing power, inflation resistance, and recession resistance. Micron has some ability to raise prices during tight cycles and in the high-end HBM phase; in a traditional cycle, its pricing power is very weak, and it is more a matter of the whole industry being buffeted by supply and demand. The fiscal Q2 and Q3 2026 guidance shows the company is indeed currently enjoying a triple tailwind of price, cost, and product mix; but the negative gross margin and enormous losses of fiscal 2023 in the past also remind us that such high profits often contain an obvious cyclical dividend. In other words, it is not that Micron cannot make money, but that it cannot guarantee sustained high profitability during a downturn or a period of renewed supply expansion. This is also why I am unwilling to capitalize the current gross margin and Owner Earnings directly into "permanent earning power."

Is management trustworthy, and is capital allocation rational. Sanjay Mehrotra has been CEO since 2017 and also chairman since 2025; he is a co-founder of SanDisk with deep experience in the flash/storage industry, and the board also includes seasoned members such as former Intel CEO/CFO Robert Swan and a former NetApp CFO. On governance, Micron's board has a lead independent director, and independent directors form the majority. On incentives, management has stock-ownership guidelines, and since 2025 a larger share of performance stock has been tied to HBM3E+ shipments, data-center SSD shipments, and relative shareholder returns, which is more reasonable than simply pursuing revenue scale.

But I want to point out two points that are less "Buffett-comfortable." First, the overall shareholdings of management and directors are not high. According to the 2025 proxy statement, as of October 31, 2025, all directors and executives held about 2.686 million shares combined, less than 1%; this means alignment of interests relies more on compensation design than on the "I'm in the same boat as shareholders" of very large holdings. Second, buybacks have had mediocre effect. Although as of February 26, 2026 the company had cumulatively repurchased about $7.84 billion of stock under its authorization, shares outstanding went from 1.119 billion at the end of fiscal 2021 to about 1.128 billion at the end of fiscal Q2 2026, without forming a clear long-term reduction in share count, indicating that most buybacks were offset by equity incentives and other changes in share capital.

Overall assessment of capital allocation. I still give management and capital allocation 4/5. The reason: in the 2022-2023 trough they took painful but rational actions—in 2022 they explicitly proposed cutting fab equipment capex by nearly 50% year over year, and in 2023 they further cut DRAM and NAND wafer starts to close to a 30% reduction; then, as conditions improved in 2026, they prioritized keeping a strong balance sheet, repurchasing $350 million of stock in fiscal Q2 while repaying $1.6 billion of debt and raising the quarterly dividend by 30% to $0.15 per share. This combination is not perfect, but on the whole it reflects a rational "survive first, then expand the advantage."

Financial Quality and Owner Earnings

The table below lines up Micron's core financial metrics over the past several fiscal years. What matters most is not that it earned a lot in some single year, but that you will immediately notice: both its profitability and cash-flow quality are highly cycle-driven, and its capex intensity has been persistently high over the long run.

Period	Revenue ($B)	Gross Margin	Operating Margin	Net Margin	Operating Cash Flow ($B)	Capex ($B)	Free Cash Flow ($B)	Approx. ROE	Approx. ROA	Approx. ROIC	Debt/Assets	Net Debt ($B)	Shares Outstanding (period-end)
FY2021	27.7	37.6%	22.7%	21.2%	12.5	10.0	2.4	14.1%	10.4%	15.3%	25.3%	-3.6	1.119B
FY2022	30.8	45.2%	31.5%	28.2%	15.2	12.1	3.1	18.5%	13.9%	22.5%	24.7%	-4.1	1.094B
FY2023	15.5	-9.1%	-37.0%	-37.5%	1.6	7.7	-6.1	-12.4%	-8.9%	-12.4%	31.3%	2.9	1.098B
FY2024	25.1	22.4%	5.2%	3.1%	8.5	8.4	0.1	1.7%	1.2%	1.7%	35.0%	4.2	1.109B
FY2025	37.4	39.8%	26.1%	22.8%	17.5	15.9	1.7	17.2%	11.2%	16.3%	34.6%	2.6	1.122B
TTM through Q2 FY2026	58.1	58.4%	48.3%	41.5%	30.7	20.4	10.3	~38.0%	~26.2%	~38.9%	28.6%	-6.5	1.128B

Note: Free cash flow = operating cash flow − capex, before deducting government incentives; ROE/ROA/ROIC are approximations based on disclosed data; net debt is total debt minus cash and short- and long-term marketable securities, with a negative value indicating net cash. TTM through Q2 FY2026 is estimated as FY2025 + FY2026 H1 − FY2025 H1. Sources and definitions come from Micron's FY2021-2025 10-Ks, the fiscal Q2 2026 10-Q, and quarterly materials.

How to read the financial quality. First the good side: the company is extremely good at making money at the top of the cycle. In fiscal Q2 2026, single-quarter revenue was $23.86 billion, non-GAAP gross margin 74.9%, operating cash flow $11.9 billion, and adjusted free cash flow $6.9 billion, showing that Micron's earnings elasticity in the current supply-demand structure is extremely strong. The balance sheet is not fragile either: at the end of fiscal Q2 2026 the company had cash and investments of about $16.7 billion and net cash of about $6.5 billion, and it continues to pay down debt.

Now the bad side, which is also the more important side. Micron's free cash flow is not stable. Operating cash flow in each of the five full fiscal years 2021-2025 was decent, but because capex has been persistently high, on the plain "operating cash flow − capex" basis the five-year average free cash flow is nearly zero; in 2023 it was even sharply negative. The company will of course emphasize "adjusted free cash flow net of government incentives," which is quite helpful when analyzing real cash pressure; but as a long-term business owner, what I care about more is this: to keep its technology lead, must this company keep absorbing enormous capex. The answer is "yes." This means it is not the kind of asset-light super-cash-machine whose profits are naturally distributable.

Are the profits real cash profits, or accounting profits. For now I lean toward defining Micron's problem as "cycle and capital intensity" rather than "poor accounting quality." Over the past few years, net income and operating cash flow have not diverged over the long run; in fact, during downturns operating cash flow was better than profit, mainly because of depreciation and amortization and swings in inventory and working capital; the sharp increase in receivables in the first half of 2026 can also be seen directly in the cash flow statement, and matches the revenue surge. Based on the disclosures reviewed, I have not seen obvious red flags of aggressive accounting or profit manipulation; but I will not equate this with "no risk," because the most common misjudgment in the semiconductor industry has never been financial fraud, but rather mistaking peak-cycle profits for the norm.

Owner Earnings estimate. On a Buffett-style "owner earnings" basis, Micron's real earning power near the peak can roughly be estimated as follows: TTM net income about $24.1 billion; add back depreciation and amortization of about $8.7 billion; add back stock-based compensation of about $1.1 billion; then deduct what I consider a fairly conservative maintenance capex of about $12 billion to $14 billion, plus about $1 billion to $2 billion of normalized working-capital use, giving conservative Owner Earnings near the current peak of about $18 billion to $20 billion. What is most sensitive here is not the accounting add-backs, but the judgment on maintenance capex: Micron's fiscal 2026 capex is expected to exceed $25 billion, a substantial part of which is growth cleanrooms and capacity build-out; but even excluding the obvious growth items, node migration for advanced memory requires maintenance capex far higher than for most industrial companies.

But in valuation I will not directly capitalize this $18 billion to $20 billion. Because the current margins and ASPs clearly benefit from an extremely tight supply-demand environment. What is more meaningful for valuation is mid-cycle, sustainable Owner Earnings. In my valuation I use three standardized starting points: $10 billion in the conservative case, $14 billion in the base case, and $18 billion in the optimistic case. In other words, I acknowledge that Micron is no longer the Micron of 2023, but I also do not accept the lazy assumption that "current peak-cycle profit ≈ average profit over the next decade."

Valuation, Margin of Safety, and Opportunity Comparison

Current price and the trading picture. As of now, MU's most recent trading price is about $698.74 per share.

Translating the current price into an "acquirer's quote." A rough calculation on the roughly 1.128 billion shares outstanding at the end of fiscal Q2 2026 puts Micron's equity market cap at about $788.2 billion; using period-end total debt of about $10.1 billion minus cash and investments of about $16.6 billion, enterprise value is about $781.7 billion. Against TTM through fiscal Q2 2026, this corresponds roughly to: P/E about 32.7x, P/B about 10.9x, EV/EBITDA about 21.2x, P/FCF about 76.7x. Even if you use adjusted free cash flow net of government incentives, these multiples remain high. For a memory maker that is strongly cyclical, capital-intensive, and fast-moving in technology, this valuation level is no longer a price with "room to spare," but a price that "believes the good times will last a long while."

Method 1: Owner Earnings discounting. The three scenarios below are all my model assumptions, not company guidance. The starting Owner Earnings uses a "mid-cycle standardized basis," not a direct extrapolation of current peak profit.

Scenario	Standardized Starting Owner Earnings	10-Year Growth Assumption	Discount Rate	Terminal Growth	Estimated Intrinsic Value per Share
Conservative	$10 billion	4%	10%	2.5%	about $135
Base	$14 billion	6%	9%	3.0%	about $268
Optimistic	$18 billion	8%	8.5%	3.5%	about $471

Based on model error, I write the actionable ranges a bit wider: conservative intrinsic value range $120-180; fair intrinsic value range $230-320; optimistic intrinsic value range $400-520. At the current price of about $699, Micron trades at a premium to all three ranges, with the premium to the base range especially pronounced.

Method 2: Relative valuation. Strictly speaking, Micron's most appropriate peers are Samsung and SK hynix; but obtaining fully comparable EV/EBITDA, P/FCF, and ROIC for all three at the same point in time and on the same basis is not easy. Even without "false precision," the relative-valuation conclusion is still stark: Reuters reported that Samsung's total market cap had topped $1.03 trillion in early May 2026, and SK hynix was also approaching $1 trillion by mid-May; while Micron, roughly calculated from its share price and share count, is already near $790 billion. At the same time, TrendForce disclosed that Micron is still third in DRAM, and the combined share of Samsung/SK hynix is clearly higher than Micron's. Against this backdrop, the market is in effect already pricing Micron as "close to the world's most core AI memory winner," rather than as a traditional third-largest memory maker. This pricing is not impossible to justify, but it is clearly aggressive.

Method 3: Asset/liquidation value. As of fiscal Q2 2026, Micron had total assets of about $101.5 billion, shareholders' equity of about $72.5 billion, net cash of about $6.5 billion, and a large amount of plant and equipment on the books. The problem is that memory fabs and related equipment are highly specialized, and liquidation discounts are usually large, so book assets signify "replacement cost and barrier to entry" more than reliable downside protection for investors. The current equity value is roughly more than 10x book equity, which means what you buy today is not an asset margin of safety, but a prepayment for the sustainability of future high profits.

Is the margin of safety sufficient. My answer is clear: insufficient, indeed one could say nonexistent. The most fragile assumption in the valuation is that "the current high gross margin, high ASPs, and AI-driven tight supply can be sustained for a very long time." If growth falls short of expectations, margins retreat, or the valuation multiple reverts toward that of a typical cyclical stock, buyers at the current price could still suffer significant permanent capital loss. Micron right now looks very much like a classic "good company, but bad price." I therefore prefer to wait rather than chase the price.

Comparison with the index, bonds, and alternative opportunities. In mid-May 2026, the U.S. 10-year Treasury yield was roughly in the 4.61%-4.69% range. For a balanced investor, buying Micron should at minimum deliver a long-term expected return meaningfully above this risk-free yield and enough to compensate for industry-cycle and multiple-compression risk; and on my base valuation, buying Micron today does not offer such compensation. By comparison, buying a broad index like SPY, while not necessarily cheap, is at least far more diversified; buying the 10-year Treasury, while capped in upside, is priced more clearly. So my comparison conclusion is: Micron is not clearly superior to the index at present, and not necessarily superior to the risk-reward of high-grade bonds.

Price-range recommendation. Under my valuation framework, the ideal buy price range is roughly $160-240, offering some margin of safety against base-case value; the acceptable holding price range is about $240-400, better suited to those who already hold, face high taxes, and believe the industry structure has improved; and the clearly overvalued range I would put at above $500. By this standard, the current price of about $699 is already clearly expensive. This conclusion does not mean Micron will necessarily fall right away, but rather that, from a "long-term business owner" perspective, the odds you get on a new purchase are not good.

Risks, Checklist, and Final Verdict

The most important risk. Micron's biggest risk is not short-term volatility, but that you mistake "temporary excess profit" at the top of the cycle for "structural long-term cash flow." Competitively, both Samsung and SK hynix are strong rivals capable of going head-to-head with Micron on process, capital, and customer relationships; technically, once HBM, advanced packaging, or node migration slips, high profits will quickly flow back to the stronger players; financially, although Micron is currently in net cash, FY2026 capex exceeds $25 billion and FY2027 will rise further, and if this ultimately creates the next round of oversupply, cash flow will come under obvious pressure; on customers, one customer contributed 17% of revenue in FY2025 and 13% in the first half of FY2026, a concentration that cannot be ignored; on regulation and geopolitics, China's CAC decision in 2023 already created significant resistance to the company's recovery, and U.S. export restrictions may continue to reshape its customer structure; and legally, the company has disclosed multiple patent lawsuits such as Netlist and YMTC, as well as securities class actions and shareholder derivative suits related to disclosure, with uncertain outcomes.

The strongest bear case. Those bearish on Micron might say: the market is pricing it as a "quasi-platform key asset of the AI era," but Micron is still, in essence, a company that sells memory bits. Today's margins are not direct evidence of a "perpetual moat," but the result of "tight supply + a high-end HBM mix + customers scrambling for goods." Once Samsung/SK hynix supply catches up, advanced-packaging expansion takes effect, and the industry once again misjudges demand, Micron will, as it has countless times in history, turn from "shortage hero" back into an "oversupplied cyclical stock." I consider this a very powerful bear case.

Which facts would overturn my current cautious conclusion. If the following facts emerge over the next two to three years, I will admit I may have been too conservative: first, that Micron can maintain free cash flow and ROIC meaningfully above its historical average through a complete downcycle, rather than looking good only in tight-supply years; second, that HBM and data-center businesses shift the profit pool durably toward CMBU/CDBU, and customer relationships gradually move from short orders and spot prices toward longer-term reservations/commitments; third, that the enormous capex does not bring industry oversupply, but instead lifts standardized mid-cycle Owner Earnings to around $30 billion per year. Conversely, if HBM yield, customer design-in, ASPs, the shift from net cash to net debt, the capex/sales ratio, single-customer dependence, and other metrics deteriorate, I will be more convinced of the current overvaluation judgment.

Investment Checklist

Checklist Item	Conclusion
Can I understand this business?	Pass
Does it have long-term stable demand?	Pass
Does it have a durable moat?	Uncertain
Does it have pricing power?	Fail
Can it generate stable free cash flow?	Fail
Is its return on capital excellent?	Uncertain
Is management trustworthy?	Pass
Is capital allocation rational?	Pass
Is the balance sheet sound?	Pass
Is the valuation below intrinsic value?	Fail
Is the margin of safety sufficient?	Fail
Does holding it long term let me rest easy?	Uncertain
Which key facts would make me sell?	Declining HBM share/yield, industry expansion causing oversupply, weakening FCF, disappearance of net cash, material deterioration in regulation/litigation
Am I only tempted to buy because the stock has risen or because of market sentiment?	This is the point to be most wary of right now

The Checklist above is an overall judgment based on the preceding financial, industry, management, and valuation analysis, and is not a prediction of the stock's short-term movement.

Final rating: Watch. One-sentence investment thesis: Micron is a memory leader that has become stronger, but the current stock price looks more like buying the optimistic storyline that "the AI memory supercycle will persist for a long time" than like buying business ownership with a margin of safety.

Core bull points.

Data centers and HBM have become the center of Micron's profit pool; in fiscal 2025 CMBU accounted for 36% of revenue, with margins far above the other businesses.
The DRAM industry has a clear oligopoly structure, Micron is the world's third-largest DRAM maker, and it has real scale and process barriers to entry.
The current balance sheet is sound, Micron was in net cash in fiscal Q2 2026, and it actively paid down debt during the upcycle.
Management cut wafer starts and equipment capex during the downcycle, showing a degree of discipline.
AI servers, traditional server refresh, and rising automotive memory density provide real support for long-term demand.

Core bear points.

The current valuation is too high: roughly calculated at the current price, P/E is about 32.7x, P/B about 10.9x, and EV/EBITDA about 21.2x.
Revenue and contract structure are not recurring; the vast majority of contracts are short-term, fixed-price, and pricing power comes more from supply and demand than from brand.
Free cash flow has fluctuated widely over the long run, growth requires enormous capital investment, and shareholders do not naturally receive stable distributable cash.
Both Samsung and SK hynix are stronger or no-weaker rivals, and Micron has no monopoly-style moat.
The deep losses of 2023 prove this is still a cyclical business that can make investors pay a high price at the wrong moment.

Key assumptions. For the Micron investment to work, at least these conditions must hold: the HBM/data-center share of the business keeps rising; capex ultimately lifts rather than destroys mid-cycle Owner Earnings; supply discipline is better than the historical average; key customer relationships are stable; the regulatory environment in China and the U.S. does not deteriorate materially further; and litigation risk does not evolve into major operational disruption.

Fair buy price. The fair buy price I give is $160-240, based on reserving roughly a 25%-35% margin of safety against the $230-320 base-case intrinsic value range. If an investor is willing to adopt more aggressive assumptions, the acceptable range can shift slightly upward, but I will not classify the current price near $700 into any "value-investing-friendly" range.

Target holding horizon. If the price enters a clearly discounted zone in the future, Micron can be a candidate for five years or more, even across-cycle holding; but at the current price, it is more like a watch-list name that requires patiently waiting for the right odds than a target to allocate heavily for ten-plus years right away.

Expected annualized return. Based on a rough scenario estimate at the current price, I offer a very restrained expectation: conservative scenario about -12%/year; base scenario about -5% to -6%/year; optimistic scenario about 0% to 3%/year. These figures come from my valuation model of medium-to-long-term intrinsic value and mean reversion, and are not a prediction of the stock price over the next year. There is only one core message: at the current buy-in, the odds are not attractive.

Maximum loss risk. For a new buyer today, the worst case is not "a scary 20% drop," but rather that if margins revert over the next two to three years, the industry returns to oversupply, and the market re-prices Micron from "AI super-winner" back to "typical cyclical stock," a permanent capital loss exceeding 50%-70% is not unimaginable. The current price is too far from both book value and base-case valuation, so the asset itself offers little downside cushion.

Tracking indicators. Going forward I will continue to track ten indicators: CMBU revenue share; HBM volume production/yield and customer design-in; data-center SSD design wins and gross margin; DRAM/NAND industry prices and inventory; Micron's capex/sales ratio; operating cash flow and free cash flow; changes in net cash/net debt; single-customer revenue share; China/U.S. regulation and export restrictions; and the expansion cadence of Samsung and SK hynix.

Signals that trigger reassessment. The investment logic must be re-examined if any of the following occur: Micron proves mid-cycle Owner Earnings to be substantially higher than my current assumption; the stock falls sharply below base-case value; HBM4/HBM4E progress substantially beats expectations; or capex falls rapidly while free cash flow does not decline. Conversely, if HBM progress stalls, key customers lose share, net cash disappears, or litigation or regulatory risk escalates materially, an immediate reassessment is also required.

Open questions and limitations. I did not obtain, at a single point in time, a complete EV/EBITDA and ROIC comparison table for Samsung, SK hynix, and Micron on a fully consistent basis, currency, and accounting standard; so the relative-valuation section emphasizes "market position + market cap + industry structure + Micron's own implied multiples" rather than a seemingly precise but inconsistently-defined peer table. To move to a live investment decision, I would recommend adding an updated same-basis comparison model after the latest quarterly reports from Samsung and SK hynix.

Final recommendation. Coolly stated, Micron now looks more like a good business worthy of respect than a good opportunity worth bidding for today. The easiest mistake in long-term value investing is not missing a company, but giving it a price that leaves no room for error at its most glorious moment. On MU, my stance now is: keep researching, wait patiently, don't chase.

This report is based on public information and does not constitute investment advice. Markets carry risk; invest with caution.

Baillie Framework · Ten Questions for Growth Investing

Hunting ten-year five-baggers among great growth stocks — pressing the upside question: "Can it get much bigger?"

How high is its market ceiling? Is it growing an existing pie, or creating an entirely new market?6/10

The ceiling is high, but in essence this is "growing an existing installed-base pie," not conjuring an entirely new market out of thin air. This is the boundary to clarify first when assessing Micron's growth.

Memory (DRAM/NAND) is a mature category that has existed for decades. Micron is not inventing demand; rather, amid the wave in which AI raises the entire curve of "how much memory a single device/single server needs," it is competing for a market that already exists. So on the Baillie Gifford-style question of "creating an entirely new market," Micron barely qualifies; where it does qualify is on "an existing large market being enlarged again by AI."

That said, the magnitude by which this pie is being enlarged is indeed considerable. On its earnings call, Micron framed the HBM segment as a long runway: management expects the HBM market to grow at roughly a 40% CAGR and ultimately approach $100 billion in size, exceeding the entire 2024 DRAM market. The report also notes that the profit pool is tilting toward cloud data centers: in FY2025 the cloud- and HBM-facing CMBU business unit had revenue of $13.524 billion, 36% of company revenue, with a 45% operating margin; by fiscal Q2 2026, CMBU single-quarter revenue reached a record $7.7 billion, 32% of company revenue, up 47% quarter over quarter. This shows AI data centers' pull on high-bandwidth, high-capacity memory is real, and the ceiling has been raised markedly.

But two boundaries should be stated honestly. First, this is "installed-base expansion with both volume and price rising," not a network-effect-style brand-new market: memory is still a commodity priced by supply and demand, and the report explicitly notes that for Micron "the vast majority of customer contracts are short-term, fixed-price contracts," not subscription-style recurring revenue. Second, demand trends up over the long run, but the supply side periodically creates oversupply, and the report repeatedly stresses that memory is a "long-term growth, strong short-to-medium-term cycle" industry. In other words, the height of the ceiling is beyond doubt, but how much Micron can steadily capture, and at what price, is constrained by the industry cycle and the joint expansion cadence of the three giants (Samsung, SK hynix, Micron), rather than decided by Micron alone. This differs fundamentally from the kind of target Baillie Gifford prefers, one that "defines and monopolizes a new market on its own."

Jun 10, 2026
Over the next five years, can its revenue at least double? Is growth driven mainly by volume, price, or new businesses?4/10

The probability of revenue doubling over the next five years is not low, but this "doubling" depends heavily on the price cycle and a rising HBM share, rather than steady volume growth. So whether it can "at least double and hold that level" is far harder to answer than whether it "doubles in some single year."

First, the facts: Micron's current revenue elasticity is extremely strong. Fiscal Q2 2026 single-quarter revenue was $23.86 billion, up about 196% year over year and about 75% quarter over quarter; the company's fiscal Q3 2026 guidance is for revenue of about $33.5 billion (±$0.75 billion) and a gross margin of about 81%. Compared with full-year FY2025 revenue of $37.38 billion, and annualizing at the current quarterly run rate, FY2026 revenue has already more than doubled versus FY2024's $25.1 billion. So to the plain question of "can it double within five years," the answer is almost certainly yes: it has already done so.

But this is exactly what exposes the core feature of Micron's growth: the driver is mainly "price," and only after that a combination of "volume" and "new business (HBM)," while "price" is a cyclical variable, not a structural one. TrendForce data show this round is a classic price-driven boom: in Q4 2025 global DRAM industry revenue surged 29.4% quarter over quarter, driven mainly by price increases rather than a surge in shipments, and Samsung's ASP rose about 40% quarter over quarter that quarter. The report also warns that "this kind of 'tightness' usually means precisely that the industry is near a profit peak, not a stable steady state."

My judgment: the volume part (AI server memory density, HBM penetration) is real long-term support; the price part may reverse at some point and pull the current high revenue base back down. The report offers a specific contrast: full-year FY2023 revenue was only $15.5 billion with a gross margin of -9.1%, the scar left by the last round of oversupply. So the more rigorous answer is: on a "mid-cycle, standardized" basis, Micron doubling its five-year revenue versus a normal year (rather than a boom peak) is possible, mainly through a rising HBM share and growth in memory content per unit; but doubling again "using the FY2026 boom peak as the base" would require yet another supercycle and should not be treated as the base case. Growth is driven by volume, price, and new business together, and the unsustainability of price is the biggest variable.

Jun 10, 2026
Five years from now, what will take over as the next growth engine? Does this "second curve" exist today?5/10

The "second curve" already exists today, and it is HBM and advanced packaging. But it is more like a "high-end upgrade of the same memory business" than a genuinely cycle-independent new growth pole.

If the current main curve is understood as "traditional commodity DRAM/NAND ramping along with the AI cycle," then the second curve clearly taking over is HBM (high-bandwidth memory) and the advanced packaging capability behind it. This curve is not a slide deck today, but has already generated hard cash revenue: in fiscal Q2 2026, CMBU revenue facing cloud and data-center HBM reached a record $7.7 billion, up 47% quarter over quarter. The report also corroborates that the company began volume production of HBM4 for the NVIDIA Vera Rubin platform in 2026, and expects the yield ramp to be faster than HBM3E.

Looking ahead, this curve's continuity has external support: Micron states that its 2026 HBM capacity is essentially sold out, demand extends into 2027, and it expects the HBM market to grow at roughly a 40% CAGR and ultimately approach $100 billion in size; the next-generation HBM4E is expected to ramp in 2027. The report also mentions that the new fabs in Boise and Japan, along with the HBM advanced-packaging plant that broke ground in January 2025, will ramp in stages around 2027. So on the question of "what takes over five years out," Micron's answer is clearer than that of most pure cyclical stocks.

But three limitations should be pointed out honestly, to avoid inflating this into a Baillie Gifford-style "independent second curve." First, HBM is in essence still a high-end form of DRAM; it upgrades the profit pool but has not escaped the gravity of memory supply and demand: the report repeatedly stresses that even when the technology is done right, "the cycle cannot be fully smoothed away." Second, it is not a track Micron monopolizes: Samsung and SK hynix are ready-made super-rivals, and once HBM's high profits draw all three into synchronized expansion, oversupply could recur. Third, the report notes that Micron trails Kioxia in NAND, with a moat narrower than in DRAM, so the strength of the second curve is uneven across product lines. On balance, the second curve genuinely exists and already contributes profit, which is a plus for Micron relative to an ordinary cyclical stock; but it is a "high-end extension of the main curve," not a brand-new engine that can carry the company across the cycle, and that is where it falls short of a truly great growth stock.

Jun 10, 2026
What is its core competitive advantage? Will this moat widen or narrow over the next three to five years?5/10

The core competitive advantage is "a handful of survivors plus a capital/process barrier on the scale of tens of billions of dollars"; the moat is real but not wide. Over the next three to five years it will widen in the HBM/data-center sub-market and stay broadly stable in traditional commodity memory. Overall it is not a monopoly-style moat.

The report gives the moat a qualitative score of 3/5, and I agree with this order of magnitude, because the source of its moat is unusual. Micron has no strong brand (end consumers buy devices rather than bare DRAM dies) and no network effect; its barriers lie mainly in three places: first, scale and cost—only three players worldwide can stay in the leading DRAM/NAND competitive race over the long run, and the barrier to entry is tens of billions of dollars of capital plus years of process accumulation; second, process/manufacturing know-how, especially at advanced nodes and advanced packaging such as 1γ DRAM, G8/G9 NAND, and HBM4; and third, moderate switching costs, higher in scenarios such as servers, automotive-grade, and enterprise SSDs that require certification and long-term supply.

Whether it "widens or narrows over the next three to five years" must be viewed market by market. In the HBM and cloud data-center sub-market, the moat has clearly widened over the past two years: in fiscal Q2 2026 CMBU revenue was a record $7.7 billion, 32% of company revenue, up 47% quarter over quarter, with margins far above the other businesses; the company also states that its 2026 HBM capacity is sold out, and it has begun volume production of HBM4 for NVIDIA's Vera Rubin, with advanced packaging a real technical and capital barrier. But in traditional commodity DRAM/NAND, the moat is broadly stable yet not wide, because price is still driven by supply and demand, and the report notes the company itself discloses that "the vast majority of contracts are short-term fixed-price."

The counter-evidence should be given honestly: the moat is real, but not exclusive. The competitors are not new players starting from scratch, but Samsung and SK hynix, two ready-made super-rivals. The latest data show Micron's DRAM share is actually being overtaken: TrendForce disclosed Q4 2025 DRAM shares, with Samsung returning to first at 36%, SK hynix second at 32.1%, and Micron still third at 22.4%, and Micron's share fell 3.3 percentage points quarter over quarter that quarter. In other words, when the industry booms all three giants expand together, and Micron has not clearly pulled ahead. Conclusion: the moat will keep widening at the high end (HBM/packaging), but whether it "can widen the relative gap versus the other two super-rivals" is in doubt, which is exactly why the report rates "durable moat" as "uncertain."

Jun 10, 2026
If its core business were disrupted, does it have the DNA to reinvent itself? How does it handle mistakes and bad news?5/10

Micron has truly weathered a life-or-death ordeal of "its core business crushed by the cycle," and survived through disciplined production cuts, cost reduction, and protecting the balance sheet. Its "self-reinvention DNA" leans defensive (survive, then expand the advantage) rather than offensive (jumping into a brand-new track when disrupted). Its way of handling mistakes and bad news is on the whole pragmatic and does not cover up faults.

First, "the response when disrupted/badly hit." The memory industry is no hypothetical enemy for Micron, and the report gives the hardest evidence: full-year FY2023 revenue was only $15.5 billion, gross margin -9.1%, net margin -37.5%, and free cash flow about -$6.1 billion (basis in the report's financial table). Facing such a deep downturn, management's actions were disciplined, painful but rational: the report records that in 2022 the company explicitly proposed cutting fab equipment capex by nearly 50% year over year, and in 2023 further cut DRAM and NAND wafer starts by close to 30%. This shows that when the core business is "punched through" by the cycle, Micron's instinct is to contract supply, hold its position in the technology race, and protect cash, rather than blindly tough it out. This DNA is crucial in a heavy-capital cyclical business, and on this basis the report scores management and capital allocation 4/5.

But two kinds of "reinvention" must be distinguished honestly. What Micron demonstrates is the resilience of "toughing out the winter within the same business, then expanding the advantage in spring": by the fiscal 2026 upcycle, the company was buying back stock and cutting debt while upgrading the profit pool to HBM. This is not quite the same as what Baillie Gifford really asks—"if the core business is entirely disrupted by a technology paradigm, does the company have the DNA to jump into a brand-new business." The memory business's paradigm has been relatively stable for decades (always storing bits), and Micron has not truly faced the test of "its main business completely replaced by another technology, forcing it to reinvent itself." So this DNA is more "anti-cyclical resilience" than "the ability to pivot against paradigm disruption."

Now "how it handles mistakes and bad news." The evidence leans positive: the report discloses that the company truthfully disclosed large-customer concentration in its 10-Q (one customer contributed 17% of revenue in FY2025, and one customer contributed 13% in the first half of FY2026), multiple patent lawsuits such as Netlist/YMTC, and unfavorable information such as securities class actions/shareholder derivative suits, without evasion. After reading through the disclosures, the report's author also states that "no obvious red flags of aggressive accounting or profit manipulation were seen." Overall judgment: Micron faces bad news head-on, discloses transparently, and makes rational decisions in downturns, which is a respectable side; but its "reinvention" is the resilience of toughing out cycles, not the explosive power of starting anew after being disrupted, and investors should not misread the former as the latter.

Jun 10, 2026
Does management (especially the founder) have a long-term vision and interests deeply tied to the company? Is it willing to sacrifice current profit for five to ten years out?4/10

Management has a long-term vision, deep industry credentials, and incentive design more tied to strategic products than peers, and it has indeed shown the discipline to sacrifice the short term for the long term; but there is one clear weak spot: management is not the founder, and overall shareholdings are extremely low (less than 1%), so "interests deeply tied to the company" relies more on compensation design than on being in the same boat with real money.

First the positives. CEO Sanjay Mehrotra has led the company since 2017 and also served as chairman since 2025; he is a co-founder of SanDisk with very deep credentials in the flash/storage industry; the report also mentions that the board includes seasoned members such as former Intel CEO/CFO Robert Swan and a former NetApp CFO, and has a lead independent director with an independent-director majority. On incentives, the report notes that since 2025 a larger share of performance stock is tied to HBM3E+ shipments, data-center SSD shipments, and relative shareholder returns—more reasonable than simply pursuing revenue scale, showing the board is steering management to focus on strategic value rather than volume.

On "whether it is willing to sacrifice current profit for five to ten years out," Micron has provided proof. The report records that in the FY2022-2023 trough, management proactively cut equipment capex by nearly 50% year over year and lowered wafer starts by close to 30%, preferring to sacrifice short-term output to protect the balance sheet and long-term competitive position; and by the fiscal 2026 upcycle, it repaid debt while raising the quarterly dividend by 30% to $0.15 per share, all while sustaining enormous capex of more than $25 billion per year to build out HBM and new capacity. This "contract in the downturn, invest in the upturn" combination reflects long-termism rather than short-term window-dressing.

But there are two hard weak spots that do not fit Baillie Gifford's ideal of "founder deeply tied in," and they must be stated honestly. First, Mehrotra is a professional-manager-type founder (of SanDisk, not Micron), and he has no founder-style ownership attachment to Micron itself. Second, and more critically, according to the 2025 proxy statement cited by the report, as of October 31, 2025, all directors and executives held about 2.686 million shares combined, less than 1%. This means "alignment of interests" relies heavily on the design of compensation contracts, rather than on the very large personal holdings of "I'm in the same boat with you, and if the boat sinks I take huge losses too." I agree with the report's conclusion: capital allocation scores 4/5, but the binding is institutionalized and lacks founder-style skin in the game. Against a Baillie Gifford standard demanding "deeply tied to the company, able to sacrifice the present for ten years out," Micron passes on "vision and discipline" but clearly falls short of founder-controlled companies on "depth of personal-interest binding."

Jun 10, 2026
If it disappeared tomorrow, how much would customers miss it? Is its growth sustainable and not reliant on harming society or regulation?5/10

If Micron disappeared tomorrow, customers would "miss it greatly in the short term, but it is replaceable in the long term": it is currently near-indispensable in HBM/advanced memory, but in standard DRAM/NAND it is a replaceable "two-of-three" supplier; its growth model itself does not harm society, but it is highly exposed to the external sustainability risk of U.S.-China export controls and geopolitics.

First, the "indispensability" dimension. In the short term, the level of missing it is high: AI memory is currently in extreme shortage, and Micron's 2026 HBM capacity is essentially sold out, with demand extending into 2027; when supply is tight, losing one advanced-memory maker would immediately hit customers (cloud vendors, the NVIDIA platform, etc.). The report also notes that switching costs are not low in scenarios requiring long-term certification, such as servers, automotive-grade, and enterprise SSDs. But in the long term, the indispensability is discounted: memory is a commodity, in Q4 2025 the DRAM market had Samsung at 36%, SK hynix at 32.1%, and Micron at 22.4%—customers always have two other super-suppliers, and the industry generally follows a multi-supplier strategy to avoid being locked in to a single vendor. In other words, Micron is "one of the key suppliers," not "the sole chokepoint." This falls short of the kind of target Baillie Gifford prefers, one that "is irreplaceable to customers if it disappears."

Now the "social/regulatory sustainability" dimension, where Micron is more fragile than one might think. The growth model itself is legitimate: selling memory to support AI, cloud, and automotive electronics, with no society-harming business model. But the external sustainability risk is significant: the report explicitly discloses that about 80% of the company's revenue comes from shipments to destinations outside the United States, so operations are highly globalized and by nature more exposed to geopolitics, export restrictions, and customer concentration. The report also names two specific risks: China's CAC (Cyberspace Administration) decision in 2023 has already created significant resistance to the recovery of the company's business in China; and U.S. export restrictions on China may continue to reshape its customer structure. Combined with customer concentration (one customer contributed 17% of revenue in FY2025, and one customer contributed 13% in the first half of FY2026) and patent lawsuits such as Netlist and YMTC, the uncertainty of the external environment cannot be ignored.

Overall judgment: Micron does not rely on harming society or regulatory arbitrage to grow, which is clean; but its "indispensability" is cyclical and relative, not structural and absolute, and its globalized operations place it on the front line of the U.S.-China tech contest, so regulatory/geopolitical sustainability is a real long-term variable. To Baillie Gifford's twin test of "how much it would be missed if it disappeared + sustainability," Micron's answer is "it would be greatly missed in the short term, but there are substitutes, and the sword of export controls hangs overhead."

Jun 10, 2026
What are the unit economics of this business (gross margin, incremental returns)? Do they improve or worsen as it scales? Where does the money it earns go?5/10

The current unit economics are extremely attractive (record gross margin, very high incremental returns), but this is the "peak economics" of a cyclical top, not a sustainable norm; scaling itself can spread costs and improve unit economics, but once the cycle reverses, unit economics will violently give back gains. This is exactly the fundamental difference between Micron and the asset-light businesses that "make more money the larger they get, and steadily so." The money it earns goes mainly into enormous capacity and advanced-packaging capex.

First, the current unit economics are indeed strong: in fiscal Q2 2026, GAAP gross margin was about 74.4% and non-GAAP gross margin about 75%, both company records, and the company guides for Q3 gross margin to rise further to about 81%. On incremental returns, the report estimates approximate ROIC of about 38.9% and ROE of about 38% for TTM through fiscal Q2 2026, an extremely excellent return on capital at a cyclical top. Scaling does have a positive effect in this business: a larger capacity base can spread fixed fab depreciation, node-migration, and R&D costs, and a rising HBM share can also improve the product mix—this is the "larger scale → better unit economics" side.

But the yardstick of the cycle must be laid down honestly. The report's conclusion after lining up the entire financial history is: both profitability and cash-flow quality are highly cycle-driven. The same company had a gross margin of -9.1% and net margin of -37.5% in FY2023 (basis in the report's financial table). So today's 75% gross margin is not "structural unit economics brought by scale," but the result of a triple tailwind of "tight supply + a high-end HBM mix + customers scrambling for goods." Scaling cannot stop the next round of oversupply from knocking unit economics back into negative territory—this is why the report rated both "can it generate stable free cash flow" and "does it have pricing power" as "fail."

The question of "where the money goes" is especially critical for Micron, and the answer is: the vast majority flows back into heavy-capital expansion rather than being distributed to shareholders. The report notes that FY2026 capex is expected to exceed $25 billion and to keep rising markedly in 2027; on the plain basis (operating cash flow − capex), the five-year average free cash flow for FY2021-2025 is nearly zero. In other words, this business "must invest more the stronger it gets," and the money it earns is largely consumed by capacity, cleanrooms, and advanced packaging, so shareholders do not naturally receive stable distributable cash. There is a small amount of return: about $650 million of stock was repurchased in the first half of FY2026, and the quarterly dividend was raised to $0.15 per share, but these are trivial relative to the scale of capex, and the report notes that long-term buybacks are mostly offset by equity incentives, forming no clear share-count reduction. Conclusion: unit economics look beautiful at the peak, but this is a peak amplified by the cycle, not a norm locked in by scale; the money earned goes mainly to reinvestment rather than shareholder returns, an unavoidable reality when judging its long-term compounding ability.

Jun 10, 2026
For it to rise fivefold in ten years, which conditions must hold simultaneously? Are these conditions realistic? What expectations does today's stock price imply?3/10

For Micron to rise fivefold in ten years from today's price of about $920 (to about $4,600, corresponding to a market cap on the scale of about $5 trillion), a string of demanding conditions must hold simultaneously, and the realistic probability is not high; and today's stock price already implies the optimistic expectation that "the memory supercycle will not recede for a long time and Micron will permanently maintain near-peak profitability"—not only is there no margin of safety, it has deteriorated further versus when the report was written.

First, a key factual update (which must be stated honestly): the report body was written anchored on MU at about $698.74 and equity market cap of about $788.2 billion, but that was a snapshot around May 20, 2026. As of June 9-10, 2026, MU's stock price had risen to about $920, market cap about $1.04 trillion, and P/E about 44x. In other words, the stock rose another roughly 30% in one month, and market cap topped $1 trillion for the first time. The report at the time judged $699 to be "already clearly expensive, with an especially pronounced premium to the base intrinsic-value range"; at the current price of about $920, this premium is larger, and the report's cautious conclusion only holds more firmly, not less.

"Which conditions must hold simultaneously for a fivefold rise in ten years"—applying the report's logic to the current price, at least: first, HBM/data-center businesses shift the profit pool durably and stably to the high end, lifting standardized mid-cycle Owner Earnings from the report's assumed $10 billion to $18 billion to about $30 billion per year (the report explicitly calls this the threshold at which it "would admit it was too conservative"); second, the enormous capex of more than $25 billion per year no longer creates the next round of oversupply, but instead keeps converting into free cash flow; third, industry supply discipline stays better than the historical average over the long run, with Samsung and SK hynix not launching a price war; fourth, the valuation multiple stays at today's high level ten years out rather than reverting toward that of a typical cyclical stock. For these four to hold at once amounts to requiring that this decades-old strongly cyclical memory business be "truly different this time"—the report's bear case puts it bluntly: once the other two catch up on supply and the industry misjudges demand again, Micron will turn from "shortage hero" back into an "oversupplied cyclical stock."

"What expectations does today's stock price imply"—by the report's three-scenario DCF (conservative intrinsic value $120-180, base $230-320, optimistic $400-520), the current price of about $920 is already nearly 80% above the top of the optimistic range ($520). This means the market is not just pricing Micron as the "world's third-largest memory maker," but as "close to the world's most core AI memory winner that can permanently maintain peak profitability"—whereas the fact is that in Q4 2025 Micron was still third in DRAM, with a 22.4% share, down 3.3 percentage points quarter over quarter. The report rated both "is the valuation below intrinsic value" and "is the margin of safety sufficient" as "fail," and estimated a base-case expected annualized return of about -5% to -6%/year based on the then-price of $699; plugging in the current price of about $920, the implied return only gets worse. Conclusion: the conditions required for a fivefold rise in ten years are, stacked together, low in realism, and the current price has instead already prepaid the "most optimistic storyline" as a done deal—this is a classic "good company, bad price," and the price is worse than when the report was written.

Jun 10, 2026
Why has the market not yet realized all this? Is it that it cannot understand it, looks down on it, or cannot see far enough? What will become the "narrative inflection point"?3/10

This question must be asked in reverse for Micron: the issue is not "why the market has not yet realized how good it is," but "whether the market has already over-realized it and re-priced it from a cyclical stock into a quasi-platform asset." The current situation looks more like "seeing it too clearly, even overshooting" than "cannot understand it / looks down on it / cannot see far." So the narrative inflection point truly worth watching is "whether the supercycle will end," not "whether the value will be discovered."

First, calibrate the premise. Baillie Gifford's "cannot understand, looks down on, cannot see far" applies to a growth stock that is "undervalued and waiting to be discovered"; but Micron's reality is the reverse—far from being ignored, it is enthusiastically chased. Over the past year MU's stock price rose about 900%, and its market cap climbed further from about $790 billion when the report was written to about $1.04 trillion in early June 2026, with a P/E of about 44x. Over the same period, SK hynix's market cap was about $955 billion and Samsung's about $1.02 trillion, with all three memory makers synchronously approaching or topping $1 trillion. The market has clearly fully (even excessively) realized the value of AI memory. So the correct way to ask is the one the report gives: is the market pricing Micron as "close to the world's most core AI memory winner" while ignoring that it is in essence still a cyclical maker that "sells memory bits, ranks third in share, and is buffeted by supply and demand"?

If one insists on applying the "cannot see far" tier, the only tenable interpretation is: the market may "see the near but not the far"—it fully prices in the high profitability of the HBM shortage over the next 1-2 years (2026 capacity is sold out), but underestimates the possibility that supply catches up and the cycle reverses in 3-5 years. This aligns with the report's bear case: today's gross margin of about 75% is not evidence of a perpetual moat, but the result of tight supply; the -9.1% gross margin of FY2023 (basis in the report's financial table) is a reminder that this business reverses with the cycle. In other words, rather than the market "not realizing how good Micron is," it is more that the market "has not fully realized it is still a cyclical stock."

"What will become the narrative inflection point"—for a target already priced optimistically, the inflection point is almost certainly a downward de-glamorizing event, and the report's tracking list is in fact the inflection-point list: first, Samsung/SK hynix advanced-packaging expansion takes effect, HBM supply catches up, the shortage eases, and ASPs peak and pull back; second, some quarter's company or TrendForce data show for the first time that DRAM/NAND inventories are rebuilding and contract prices are weakening (the present is precisely a price-increase-driven peak, with 4Q25 industry revenue up 29.4% quarter over quarter driven mainly by price increases); third, HBM4 yield/customer design-in falls short of expectations, or a single large customer (contributing 13% of revenue in the first half of FY2026) loses share; fourth, U.S.-China export controls escalate materially. A reverse, upward inflection point that could keep supporting the high valuation also exists—if Micron can still maintain free cash flow and ROIC far above its historical average through a complete downcycle, proving "this time is different," the report's author also says he would admit he was too conservative. Overall judgment: the current perception gap on Micron is more likely "priced too full" than "undervalued," and the narrative inflection point will most likely come from the real data of cycle and supply breaking the expectation that "the supercycle never recedes," rather than from a belated discovery of value.

Jun 10, 2026

Ask about this report

Members can ask about this report; once answered it appears under "Reader Q&A" on this page. You can also highlight a passage in the text to ask about it directly.