Deutsche Bank and TD Cowen Both Raise Micron Price Target to $150
0xBroomberg
Ahead of Micron's fiscal Q3 report on June 24, Deutsche Bank, TD Cowen, and Cantor Fitzgerald all anchored their target at $1,500 — implying roughly 47% upside from Tuesday's close. The core shift: Wall Street is repricing Micron from a traditional inventory cycle to an AI-capex-driven story.
Three firms land on the same $1,500 — why now?
Deutsche Bank raised its target from $1,000 to $1,500 (up 50%); TD Cowen raised from $660 (up 127%); Cantor Fitzgerald set $1,500 simultaneously.
All three imply roughly 47% upside from Tuesday's close — this is not one outlier call but a coordinated repricing.
This means → the Street's disagreement on Micron has narrowed sharply, and "AI makes memory worth more" is becoming consensus.
How does Deutsche Bank get there?
Deutsche Bank lifted its fiscal Q3 (May quarter) revenue estimate to $35.1 billion, above the top end of Micron's own guidance.
It projects CY2027 EPS of $160 and gross margins staying above 80% for the foreseeable future.
The core logic: AI-driven, memory-intensive workloads — large-model training and inference that demand massive memory bandwidth — are turning both conventional DRAM and low-power DRAM into growth engines.
In plain terms = DRAM used to live off phones and PCs. Now a single AI server rack swallows what an entire production line used to supply, and capacity expansions still can't keep up.
Why did TD Cowen reverse its own call?
Analyst Krish Sankar had previously forecast a digestion period in H1 CY2027 — buyers slowing purchases, prices softening. He has now withdrawn that call.
New logic: CPU demand is leading buyers to expect strong pricing through H2 CY2027; server DRAM prices are projected to peak around Q3 CY2026.
TD Cowen's CY2027 EPS estimate is $150, anchored on one key assumption: even after SOCAMM — a new memory-module form factor — adjustments, DRAM content per gigawatt of power consumption keeps rising.
This means → AI infrastructure build-out is not a one-time stockpile. It structurally raises the memory content per server, and the old "what goes up must come down" cycle may not apply this time.
How far have other firms followed?
RBC Capital raised to $1,200: the DRAM upcycle has now run for 12 consecutive quarters, with both pricing and shipment volumes strengthening.
Wolfe Research raised to $1,250: expects DRAM and NAND pricing to keep climbing through 2026–2027.
Aletheia Capital went further — $1,600 — and explicitly switched to a CY2027 forward P/E valuation framework.
This reflects a collective shift in Wall Street's valuation anchor — pricing Micron not on current earnings but on how long AI capex can sustain the cycle.
What does the market most want to hear on June 24?
Deutsche Bank flagged that the earnings call's central focus will be Micron's previously announced Strategic Customer Agreements (SCAs) — long-term supply contracts with major buyers.
Three things the market wants to know: the post-settlement pricing mechanism, whether customers are co-investing in capex, and what share of Micron's total business these agreements cover.
Analysts expect management to disclose some incremental detail, but with contract negotiations still underway, the scope of disclosure may be limited.
In plain terms = investors want confirmation of how much volume and how many years Micron has locked in — that is what turns "AI demand is not a passing wind" from narrative into numbers.
Where does this memory cycle end?
Multiple firms now extend the DRAM supply–demand imbalance through 2028 — far beyond the typical memory-cycle length.
Even though the entire DRAM ecosystem has announced accelerated capacity expansion over the past 180 days, the supply–demand tightness shows no sign of easing.
This means → the pricing logic of this upcycle has changed. It is no longer the old playbook of "inventory bottoms → restocking → prices recover." Instead: sustained AI capex → structurally higher memory demand → a higher price floor.
The June 24 earnings report and call will be the first key checkpoint for whether this new logic holds up in actual numbers.
Content is for reference only, not financial advice.