Micron’s 6.26x Forward Multiple Exposes a Deep AI Memory Discount

Chloe Harper August 13, 2026

Micron gapped hard on the inflation data. MU traded at $906.61 in Wednesday’s premarket, up $38.09 or 4.39% at 9:02 a.m. ET, and opened the session 4.72% higher after closing Tuesday at $868.52, up $7.52 or 0.87%.

That is a $38 move on a macro print with zero company-specific news attached, and it tells you exactly what has been suppressing this stock. Headline CPI slowed to 3.4% year-over-year in July with core at 2.5%, both matching consensus to the decimal, per the July 2026 CPI release. Nasdaq futures ripped 302.50 points to 29,928.50, a 1.02% advance that outpaced the Dow by 72 basis points.

Micron is the highest-beta expression of that repricing. A stock trading at 6.26x forward earnings with 346% year-over-year revenue growth is a pure duration instrument — every basis point of relief on the discount rate flows straight into the multiple. A Fed that cannot hike in September removes the single largest headwind on a name that has already surrendered 25% from its late-June peak.

The setup going in was compressed. Memory stocks bucked the trend Tuesday and rose while all three major indices closed lower, with recent management commentary at industry technology conferences reinforcing that AI-driven demand for memory hardware is expanding faster than manufacturing capacity can be added.

The thesis here is a valuation gap that has become absurd. Micron trades at 19.74x trailing earnings and 6.26x forward earnings against fiscal Q4 guidance of $50.0 billion in revenue and $31.00 in non-GAAP EPS. Consensus across 46 analysts sits at $1,502, implying 72.94% upside from Tuesday’s close. The company has $100 billion of contracted backlog under 16 strategic customer agreements and cannot fulfill more than two-thirds of customer demand.

The counterweight is equally real. Insiders dumped over $167 million in the trailing three months with no open-market purchases. Citi cut its target to $1,150 from $1,400 on decelerating price growth. Fiscal 2027 EPS estimates have advanced 1.2% over the past month after jumping from $95.80 to $154.70 over three. The stock lost $1 trillion in market cap on July 16 and has not reclaimed it.

Buy the gap. Target $1,150 with the stop at $854.46.

The Three-Day Slide That Built the Base

The reversal came off a genuine flush, and the sequence matters for where the stop belongs.

MU fell 1.89% on Monday August 10, dropping from $877.57 to $861.00. That marked three consecutive days of decline. The intraday range spanned $854.46 to $894.99 — a 4.74% swing inside a single session, which is characteristic of a stock where both sides are fighting over a level rather than trending. The stock declined in six of the prior ten sessions for a cumulative 4.35% loss over that stretch.

Volume decreased on the down day alongside the price, which is the constructive read: selling pressure without participation is distribution exhausting itself rather than institutional exit.

Tuesday reversed it. MU closed at $868.52, up 0.87%, recovering $7.52 while the S&P 500 lost 0.32% to 7,728.20 and the Nasdaq Composite shed 0.6% to 26,445.45. Memory outperformed a red tape by 119 basis points. Monday’s premarket had the stock at $883.96, up 0.73%, before the session gave it back.

Technical scoring going into Wednesday was negative. One systematic model rated MU a Sell candidate with a technical score of -1.08 on a -10 to +10 scale based on data through August 10, a downgrade from Hold, citing small weaknesses in the technical picture and projecting a fair opening price of $870.15 for August 11.

The stock printed $868.52 and then gapped to $906.61.

That divergence — negative systematic scoring immediately preceding a 4.4% gap — is the signature of a stock where the technical picture is lagging the fundamental one. Momentum models measure the last twenty sessions. Micron’s last twenty sessions were a drawdown from a July peak. The fundamentals are a company guiding revenue from $41.46 billion to $50.0 billion sequentially.

The base is $854.46, Monday’s intraday low. That is 5.8% below the premarket print and it is the level that defines whether this gap holds or fills.

The Chart: A $1,200 Peak, a 25% Drawdown, and a Trillion-Dollar Line

Micron’s price structure is a parabolic advance followed by a correction that has not yet resolved.

The stock peaked above $1,200 in late June, immediately after the fiscal Q3 report drove a 15% extended-hours advance. It closed Monday July 20 at $970.82. Market capitalization fell below $1 trillion on July 16 and has remained below that threshold since. The current market cap sits near $991 billion.

From the late-June peak above $1,200 to Monday’s $854.46 low, the drawdown measured 28.8%. From that low to Wednesday’s premarket print at $906.61, the recovery measures 6.1%. The stock has retraced 15.1% of the total decline.

July was the worst month for the stock since 2005. That is the context for the current bounce — a name up over 200% year-to-date that gave back a quarter of its value in four weeks and is now attempting to rebuild.

Year-to-date performance readings depend on the measurement date and the spread is instructive. MU was up 239.2% year-to-date in late July at $970.82. It was up 207% as of the August 7 close. It sits up over 200% at the current level. A stock that can shed 32 percentage points of annual performance in two weeks is a stock where position sizing matters more than direction.

The trillion-dollar market cap line is the psychological level above the price. Reclaiming it requires roughly $875 per share on current share count, which the stock has already cleared in premarket. Holding above it through a full session is the confirmation the technical picture needs.

Systematic August projections bracket the range: one model targets $904.50 by August 31 for 4.18% upside, another projects $714.63 for a 17.69% decline, averaging $809. A separate monthly framework places August’s range at $702 to $987 with an average of $832 and a month-end close near $817.

That dispersion — $714 to $987 — is a 38% spread on a four-week horizon. Micron is not a stock to trade on point estimates.

Fiscal Q3 Delivered an 84.9% Gross Margin and That Is Not a Typo

The June quarter is the most consequential report in the company’s history and it is the entire foundation of the bull case.

Revenue reached $41.46 billion for the third quarter of fiscal 2026, ended May 28, against $23.86 billion in the prior quarter and $9.30 billion a year earlier — up 345.7% year-over-year and 74% sequentially, the fifth consecutive quarterly revenue record. That topped the $35.82 billion consensus by $5.64 billion.

Non-GAAP EPS printed $25.11 against a $20.71 consensus, a 21% upside surprise. GAAP diluted EPS came in at $24.67. GAAP net income reached $28.24 billion against $1.89 billion, or $1.68 per share, a year ago. Non-GAAP net income hit $28.86 billion.

The margin is the number that redefines the company. Non-GAAP gross margin reached 84.9%, a company record, up from 74.9% in the prior quarter and 39.0% a year earlier. GAAP gross margin printed 84.6%. That is a 4,590 basis point expansion in twelve months on a business that historically fought for 20-30% margins.

Cash generation followed. Operating cash flow reached $25.39 billion against $11.90 billion the prior quarter and $4.61 billion a year ago. Capital expenditures ran $7.1 billion net, leaving adjusted free cash flow of $18.3 billion. Cash and investments totaled $30.2 billion. The board declared a $0.15 quarterly dividend.

Data center revenue exceeded $25 billion in the quarter — an annualized run rate above $100 billion from a single segment. Data center SSD revenue exceeded $5 billion, more than doubling sequentially.

Put the margin and the cash flow together and the picture is unambiguous. Micron converted $41.46 billion of revenue into $25.39 billion of operating cash flow, a 61.2% conversion rate. That is a software company’s cash conversion on a semiconductor manufacturer’s asset base, and it exists because the supply deficit lets the company price without competitive constraint.

Every key financial metric beat the high end of management’s own guidance and crushed consensus.

DRAM at $31.3 Billion Is 76% of the Company

The revenue mix concentration is both the engine and the risk.

Fiscal Q3 DRAM revenue reached $31.3 billion, a company record representing 76% of total revenue, up 343% year-over-year and 67% sequentially. Bit shipments rose only a low-single-digit percentage. Average selling prices climbed in the low-60s percentage range sequentially.

Read those two figures together. Bit shipments up low-single-digits with ASPs up low-60s percent means the entire revenue expansion is price, not volume. Micron shipped roughly the same quantity of memory and charged 60% more for it. That is what a structural supply deficit looks like in a P&L.

The prior quarter established the pattern. Fiscal Q2 DRAM contributed $18.8 billion at 79% of revenue, up 207% year-over-year, with prices rising mid-60s percent quarter-on-quarter. Two consecutive quarters of 60%-plus sequential price increases on flat volume.

Industry pricing data corroborates it. Conventional DRAM contract prices rose 93-98% quarter-on-quarter in the first quarter of 2026, followed by 58-63% in the second. Third-quarter forecasts project another 58-63% increase in DRAM contract prices.

DRAM accounts for almost 80% of total revenue and quarterly DRAM sales have exploded roughly fivefold from a year ago, fueled by demand for high-bandwidth memory used in AI servers.

The mechanism that sustains it: HBM uses far more wafer capacity than conventional DRAM, so rising HBM production mechanically limits the supply of ordinary memory chips. The pivot toward manufacturing high-bandwidth memory for AI accelerators reduces conventional memory chip yields, tightening the overall market. Micron is shipping HBM4 in high volumes for its lead customer, and HBM capacity for 2026 is fully sold out under binding contracts.

That creates a self-reinforcing loop. More HBM demand means less conventional DRAM supply means higher conventional DRAM prices means higher blended margins. The loop breaks only when total wafer capacity expands, and that does not happen before fiscal 2028.

NAND Doubled Sequentially on Mid-80s Percent Price Increases

The flash business has gone from commodity to constraint, and the numbers are more extreme than DRAM.

Fiscal Q3 NAND revenue hit a record $9.9 billion, up 361% year-over-year and 99% sequentially, representing 24% of total revenue. Bit shipments increased in the mid-single-digit percentage range while prices increased in the mid-80s percentage range.

An 85% sequential price increase on mid-single-digit volume growth is the tightest pricing environment NAND has produced in the modern era. The prior quarter delivered $5.0 billion, up 169% year-over-year, with prices up high-70s percent sequentially.

Industry contract data tracks it. NAND flash contract prices climbed 85-90% in the first quarter of 2026 and 55-60% in the second. Third-quarter forecasts project a 70-75% increase — an acceleration from Q2 rather than a deceleration.

The demand driver has shifted. NAND flash is moving from commodity status to an AI-critical asset, with enterprise SSDs and data center storage absorbing capacity that previously served consumer devices. Micron’s data center SSD revenue alone exceeded $5 billion in the quarter, more than doubling sequentially.

Supplier behavior confirms the tightness. Micron paused its NAND and DRAM quotes, signaling sharper increases ahead, after a competitor announced a 10% hike for NAND products. Spot prices continue climbing, with mainstream DDR4 chips advancing 3.31% in a single week from $4.896 to $5.058.

The friction point is buyer resistance. Transaction volumes are shrinking as buyers resist steep increases, showing limited willingness to absorb higher costs, and DDR5 spot trading remains subdued despite representing the newest technology.

That resistance is the leading indicator to watch. Price increases that customers refuse to transact at are not price increases — they are quotes. When spot volumes contract while prices rise, the market is approaching the level where demand destruction begins, and that is the mechanism through which the cycle eventually turns.

The $100 Billion RPO Is the Structural Change

The single most underpriced element of the Micron story is the contract structure that now sits underneath the revenue.

Management has signed 16 strategic customer agreements. Remaining performance obligations at the end of fiscal Q3 exceeded $5 billion, and including agreements executed after the quarter closed, RPO totals approximately $100 billion. That figure is determined based on minimum committed volumes.

Read that against the company’s revenue base. Fiscal Q4 is guided to $50.0 billion. A $100 billion RPO represents two quarters of contracted minimum volume at the current run rate, locked under agreements with defined pricing — either fixed or subject to floor and ceiling terms.

That is a fundamentally different business than the DRAM producer of 2019. Historically, memory companies sold into spot and short-cycle contract markets with no forward visibility, which is why the sector traded at trough multiples through every upcycle. A $100 billion book of minimum-volume commitments with price floors converts a cyclical into something closer to a contract manufacturer with take-or-pay economics.

Management framed it directly: visibility on demand, committed volume the company can be confident about when making investment decisions.

The disclosure itself is new — the company began reporting RPO in the May quarter specifically because the SCA structure required it under revenue accounting standards. That means the market has had one quarter to price a structural change, during which the stock fell 28.8%.

The qualification is that RPO is not revenue. Minimum committed volumes at floor pricing convert to revenue at the floor, not the ceiling, and floors sit below current spot. If memory prices peak and roll over, the SCAs protect volume and cap downside pricing rather than preserving current margins.

That is still the most valuable thing on the balance sheet. A price floor across $100 billion of committed volume is exactly what the equity has never had, and it is the reason a 6.26x forward multiple is defensible rather than a value trap.

Fiscal Q4 Guidance Is $50 Billion at an 86% Gross Margin

The forward quarter is guided to a level that would have been unthinkable eighteen months ago.

Management guided fiscal Q4 2026 revenue to $50.0 billion plus or minus $1.0 billion, with gross margin near 86% and non-GAAP EPS of $31.00 plus or minus $1. That implies 20.6% sequential revenue growth off a quarter that already grew 74% sequentially, and another 110 basis points of margin expansion off a record 84.9%.

Consensus for the September 22 report sits at $31.24 to $31.29 in EPS on revenue of $50.72 billion to $50.82 billion, against $3.03 per share on $11.31 billion in the year-ago period. That is a 931% EPS increase and a 349% revenue increase year-over-year.

The margin trajectory is the tell on pricing power. Gross margin ran 39.0% in the year-ago quarter, 74.9% in fiscal Q2, 84.9% in fiscal Q3, and is guided to 86% in fiscal Q4. Management had guided 81% for Q3 and delivered 84.9% — a 390 basis point beat driven by pricing power and node cost reductions.

New fab startup costs of $100 to $200 million per quarter remain manageable at current revenue scale. Against $50 billion of quarterly revenue, $200 million of startup drag is 40 basis points.

The forward-looking commentary is where the durability case sits. DRAM and NAND industry demand continues to significantly exceed industry supply, and management expects tight conditions to persist beyond calendar 2027 as a result of AI-driven demand across all segments coupled with structural supply constraints.

The severity of the constraint was quantified earlier in the year: Micron can fulfill only 50% to two-thirds of customer demand in the medium term. A company rationing output at that ratio does not have a pricing problem. It has an allocation problem, and allocation problems produce margins like these.

The company acknowledged no high-confidence view of when supply catches demand, with new fabs not delivering meaningful output until fiscal 2028.