Micron printed the best quarter in its history, and the 26% haircut since is the market asking what comes after perfect.
A $41.5 billion quarter at an 84.6% gross margin, with a $50 billion guide on top, would normally end the argument. Instead the stock has round-tripped from a $1,255 intraday high through $739 and back to $924, because the debate was never about this quarter: it is about whether floor-priced contracts really repealed the memory cycle.
Q3 FY2026 earnings, decoded
The June 24 print beat every line of its own guidance, and the guide that came with it was bigger than the beat.
Where the revenue comes from
Micron reports four business units since its April 2025 reorg, and every one of them at least tripled year over year in Q3. Tap a card.
Who Micron buys from & sells to
A trillion-dollar memory maker drags its whole neighborhood with it: four toolmakers upstream, the AI complex downstream. Click any node.
Flow weight is a subjective 1-10 read of how tightly each name's results move with Micron's memory cycle, not a revenue rank.
Bull, base & bear case
Bull and bear agree on the facts: record results, sold-out capacity, floor-price contracts. They disagree on whether that makes the cycle dead or just late. Tap a case.
- Sold out through 2027 HBM3E and HBM4 capacity is fully booked, and industry experts see the memory bottleneck persisting into 2028.
- $100B at floor prices Fourteen of Micron's 16 strategic customer agreements carry minimum-price commitments worth roughly $100B, backed by $22B of committed deposits and letters of credit (~$18B cash, ~$4B LCs).
- Everyone needs more Nvidia prepays for allocation, hyperscalers plan roughly $700B of 2026 AI capex, and Musk is complaining about memory prices in public.
- Cheap on the guide At 7.5x annualized Q4 guidance EPS, the multiple already prices in a downturn the order book says isn't coming.
Price-target calculator
Pick your FY27 EPS and the multiple you would pay for it. The entire Micron debate fits inside these two sliders.
Implied value = EPS × P/E multiple
Price target = forward EPS × P/E. The defaults ($115 of FY27 EPS at 8x) roughly reproduce today's $924 share price. Street FY27 EPS estimates span roughly $55 to $155 depending on where you think memory pricing settles, and the bear and bull cases above live near the ends of these ranges.
What Wall Street says
Forty-six analysts cover Micron and not one of them says sell; July's 41% drawdown produced zero downgrades and one added buy rating.
Technical picture
One year, one 11.5x melt-up, one air pocket: the entire memory argument is drawn on this chart.
An 11.5x run from $109 to a $1,255 intraday high by late June, a 41% five-week air pocket to $739 as profit-taking, executive sales and China's CXMT debut hit at once, and a sharp dip-buyer bounce back above $900: the tape shows total conviction in the earnings and much less in the multiple.
Past performance does not guarantee future results. Prices are daily closes from Yahoo Finance.
Final assessment
The pullback did the underwriting for you: the same order book, 26% cheaper.
Build the position on weakness rather than chasing strength; the crash repriced the multiple, not the business.
The quarter removed the fundamental debate. Revenue of $41.5B at an 84.6% gross margin, a $50B guide, and roughly $100B of contracted floor-priced revenue make these earnings as close to underwritten as memory has ever been. What July repriced was the multiple, not the business: the 41% air pocket traced to profit-taking, executive share sales and hyperscaler capex nerves, not to a single order cancellation.
At 7.5x annualized guidance EPS the stock pays you to hold cycle risk that management has partially contracted away. The honest caveat: every memory cycle in history looked repealed at the top, and the 2027-2028 capacity wave is the exam this thesis has not yet sat. Accumulate here, add on weakness, and size the position for that exam rather than for the order book.
- Q4 confirms the guide A clean $50B quarter at ~86% gross margin in late September would validate the contract-floor thesis at even bigger scale.
- HBM4 share gains Micron holds roughly 21% of HBM against SK Hynix's ~57%; every point of HBM4 share won on the lead platform is upside the base case doesn't count.
- Visibility extends into 2028 New long-term agreements or deposits that push booked capacity into 2028 would break the peak-2027 bear case outright.
- Hyperscaler capex flinches July was the rehearsal: any guided slowdown in AI infrastructure spend hits memory expectations first and hardest.
- Spot prices roll over If spot DRAM turns down while contract prices hold, the market will treat it as the cycle reasserting itself and de-rate first, ask later.
- Supply lands early Accelerated 2027 capacity from Samsung and SK Hynix would put the floor-price contracts to their first real stress test.
- Insiders keep selling Executive share sales helped spark the July slide; continued distribution into strength deserves respect.