AI-led memory demand and constrained supply could sustain elevated pricing and earnings longer than past cycles, while customer commitments improve visibility.
The bull case depends on scarcity economics persisting; supply growth, customer flexibility, and rising costs could instead compress margins and valuation.
The stock showed strength, contributing to the semiconductor sector's rise.
SNDK/MU +4% on the back of Hynix’s solid print that came above buyside bogeys on both #s and capital return (Kioxia +14% /Samsung +4% overnight). Software -1% with NOW -2% leading the way lower ahead of their print tonight (Bernstein put a damper on bulls h...
Memory remains a critical AI bottleneck and a key driver of token/$ and power efficiency, with recent de-specs reflecting constrained supply rather than diminished strategic value.
New 3–5 year take-or-pay SCAs improve demand visibility and capex discipline.
The company maintains leadership across HBM/DRAM/NAND, with HBM scaling through 16-hi and G11 NAND targeting 400+ layers.
Management reiterated stronger capital returns beginning after the Dec. 9 CHIPS restriction deadline.
The company's memory products are performing strongly, with DRAM outperforming NAND and spot prices hitting new highs, indicating robust demand.
Management sees the memory market as tighter in CY27/28 compared to CY26, and de-specifications are attributed to compute, land, and power constraints rather than weak memory demand.
A significant portion of the company's sales are locked at price floors above prior cycle highs, with a target to increase this percentage, providing revenue stability.
New capacity from Samsung and SK Hynix may not significantly lift memory supply until 2028, indicating a lag between investment and increased production.
Potential price normalization for memory is flagged from late 2027 or early 2028 as production ramps up.
The company's stock was mixed, with sentiment suggesting that if Elon Musk can add significant capacity, the market might not be as tight as perceived.
JPMorgan states that memory shares are trading near historical downcycle trough valuations despite spot DRAM pricing holding flat-to-up, HBM ASP negotiations biased higher, and order discussions extending into 2028/29.
Micron's Taoyuan union plans to strike on October 22 over employee bonuses, escalating a dispute and potentially impacting production from a major DRAM and HBM manufacturing hub.
The company's new Taiwan plant is already sold out for its planned 2027 capacity, indicating strong demand.
COO Manish Bhatia expects HBM to outgrow the broader DRAM industry through 2028.
Rosenblatt reiterates a Buy rating, arguing that the current P/E multiple reflects an outdated memory-cycle framework and that AI has structurally improved DRAM visibility.
Rosenblatt believes that CXMT’s planned 4F² DDR5 RDIMM does not establish near-term competitiveness due to its larger cell size and unproven yields and economics.
AI has structurally improved DRAM visibility, with hyperscalers planning years ahead and new fabs requiring significant time and investment, reducing the risk of oversupply.
AMAT's new $5 billion EPIC Center aims to deepen customer integration by bringing Micron and other ecosystem partners into parallel R&D, which should shorten AI-driven development cycles and enable system-level co-optimization.
HDDs and NAND bounced hardest (STX/WDC opened -8%/-7%, SNDK -5.5%), MU much less so. Near the bottom today, WDC/STX were trading at ~5x/7x ‘28 buyside numbers (assuming pricing reaches $30/TB) so definitely investors sharpening their r/r numbers here despit...
GS favors SNDK and SK Hynix for HBF leadership, with Samsung also well positioned; MU is a less direct beneficiary.
Counterpoint's checks indicate that most Nvidia Rubin products are converging on 12-layer HBM, countering speculation of a move to eight layers, and found no evidence of rumored four-layer HBM configurations for custom AI chips.
HBM requires more than three times the production capacity of conventional DRAM, which will maintain pressure on broader memory supply.
Investors are underappreciating the multi-year memory shortage, as AI models require more memory for performance, speed, and context, leading to demand outstripping supply through 2027–28.
Long-term agreements now cover a significant portion of supply, and demand is increasingly concentrated among hyperscalers, reducing contract breakage risk.
Chinese capacity additions are expected to be absorbed by strong domestic demand, not alleviating the global shortage.
The analyst sees a significant multi-year growth runway.
The company settled a patent dispute with Netlist for $600 million over five years, which includes a license for HBM technology.
Micron's Taoyuan union secured authorization to strike over profit sharing, which could disrupt production in a key manufacturing hub.
Memory lagged: MU -1%; SNDK -92bps; SKHY flat
Bernstein's positive note on pricing indicates a significant rise in DRAM contract prices and an increase in enterprise SSD/mobile NAND prices, with TrendForce also projecting strong NAND growth.
The company settled a dispute with Netlist for $600 million, removing a potential overhang.
Nanya and Winbond reported record Q3s, with Nanya only filling 50-60% of demand, and JPM's MU IR call projects significant bit deficits in CY27/28.
Memory lagged again: SKHY -4.2%; SNDK -1.7%; MU +21bps. Korean chatter is similar to last quarter where mgmt making the sounds on the sell-side saying SKHY’s Q3 OP will land below the Street on an FX hit (MSD-HSD off ASPs) and that 2028’s wave of new clean...
Memory lagged: MU -1%; SNDK -92bps; SKHY flat
The company exceeded revenue and EPS expectations for FQ4, and provided strong FQ1 revenue and EPS guidance, indicating better-than-expected financial performance.
NAND revenue significantly contributed to the upside, accounting for two-thirds of the revenue beat, driven by strong pricing and bit growth.
Management anticipates tighter supply conditions in CY27/CY28 compared to CY26, suggesting a prolonged period of favorable market dynamics and extended earnings potential.
The company's strategic customer agreements (SCAs) have increased, covering a larger portion of future revenue and providing better visibility and stability, with new agreements being negotiated at higher prices.
The vast majority of CY27 HBM bit supply is contracted at significantly higher year-over-year prices, which is narrowing the profitability gap with conventional DRAM.
HBM revenue is growing faster than total company revenue, the customer base has expanded, and HBM4 execution is on track, including a co-development with NVDA for custom HBM4E.
NAND pricing increased significantly quarter-over-quarter, and bit shipments also saw strong growth, indicating robust demand and effective pricing strategies.
Data center SSD revenue saw substantial year-over-year growth, driven by KV-cache offload and HDD displacement, expanding the company's opportunity in the full AI memory hierarchy.
The company achieved strong adjusted free cash flow, and management expects to reach its target cash balance soon, enabling increased capital returns primarily through share repurchases.
Management expects sequential revenue growth each quarter and gross margins to remain above the November floor, indicating a positive outlook for profitability.
The company is becoming more strategic in the datacenter, with gross margins expected to remain higher for longer due to increasing long-term agreements, which should lead to a higher valuation.
The company is becoming less dependent on consumer device units and more exposed to AI-driven memory demand, making it less cyclical.
The company's leading production nodes, custom-HBM collaboration, and expanding customer relationships provide a strong company-specific execution story.
Strategic customer agreements offer a bridge from strong current demand to a higher through-cycle earnings floor, with committed volumes improving investment planning and pricing frameworks protecting parts of the business.
The company's strong cash generation provides an alternative path to returns even without significant multiple expansion, as management intends to return excess cash.
Despite intraday volatility, the stock closed near its highs, with investors becoming incrementally positive on the memory space due to better visibility, long-term agreements, and supply/demand dynamics.
The closing of HBM's margin gap is expected to add significantly to quarterly EPS.
The gross margin guide for FQ1 is lower than buyside expectations, primarily due to the mid-shift to HBM, although management expects it to be the floor for GMs in '27.
The revenue beat did not translate into equally clean operating leverage, as OpEx increased significantly quarter-over-quarter due to incentive compensation and community investments.
November OpEx is guided higher than Street expectations, and FY27 OpEx is projected to increase further, mainly due to R&D and incentive compensation.
The higher November costs are not just a one-quarter add-back, as higher FY27 compensation will affect manufacturing margins more fully in FQ2, along with startup costs and depreciation.
The implicit FY27 net CapEx is significantly higher than FY26, with most of the increase for construction rather than immediate equipped wafer capacity, meaning first output is not full production.
Consumer demand is less robust, with PC and smartphone units potentially declining double digits in CY26, and the company's mobile/client bits declined sequentially.
Exceptional scarcity profits might be mistaken for a permanently different earnings model, as customers have incentives to optimize memory content and use lower-cost tiers.
The disclosed protection from strategic customer agreements is narrower than suggested by the '75% committed' figure, as it refers to total output and only a portion of SCA revenue has defined pricing frameworks.
The contracts have not been tested through a downturn, and customer financial commitments are not equivalent to guaranteed profit, raising questions about their resilience when spot prices fall.
Cost and supply risks become more significant as pricing growth moderates, with higher revenue not automatically translating into higher gross margin, and R&D, compensation, startup costs, and depreciation all increasing.
Construction spending for new fabs does not immediately create bits and eventually removes a constraint that currently supports pricing, potentially impacting future pricing strength.
Premium mix is masking weaker consumer-unit demand, and Chinese competition remains a potential source of global pricing pressure.
Bears express uncertainty about 2028 and beyond, anticipating that new tools and increased supply will lead to a decline in open-market pricing from mid-2028.
The company reported strong financial results for FQ4, with revenue significantly exceeding Street expectations and EPS surpassing both Street and buyside estimates.
The FQ1 revenue and EPS guidance also cleared expectations, indicating a positive outlook for the upcoming quarter.
Management anticipates FQ1 to be the floor for gross margins in fiscal 2027, with expectations for higher gross margins beyond FQ1 for the remainder of the fiscal year.
In Tech, we get SNPS analyst day, HPE Networking event (stock +4% after raising FY27 networking growth to highs teens/low 20s and increasing synergies target to $800M vs. $600M), and MU earnings (bogeys below).
The FQ1 gross margin guidance was lower than buyside expectations, which was identified as the only real negative point in the earnings report.
The lower FQ1 gross margin guidance is attributed to a mid-shift to HBM and increased fiscal 2026 incentive compensation absorbed into inventories in FQ4.
The company's stock performed well ahead of its earnings print, indicating positive sentiment in the memory sector.
SKHY on Bloomberg’s report that Solidigm is weighing a US IPO as soon as 2027 at up to $100B. Sell-side out with bullish MU previews (JPM sees the DRAM/HBM shortage through CY27-28); investors are focused on FY27 margin durability as LTAs get layered on. HD...
The stock performed slightly better ahead of upcoming earnings, despite a general decline in investor preference for memory names.
Memory pricing is expected to settle into a higher, longer plateau, with tight supply and low inventories supporting AI DRAM hikes through 2027.
DigiTimes reports Micron’s Taoyuan union plans a strike-authorization vote after September 21 mediation failed to resolve profit-sharing demands.
An announced one-time NT$1 million employee bonus has not settled the dispute.
Memory underperformed with SNDK -1.4%/MU +2.4%/SKHY +75bps. Some talk out of Asia of Hynix mgmt trying to lower the bear because of x. We showed this chart below in our weekly yesterday which shows that even a significant amount of adoption likely isn’t a h...
CXMT's G5 DRAM production, with 50% more gross dies per wafer than G4, could strengthen its cost position against competitors, including this company.
The company's memory products underperformed, with some discussion from Asia suggesting Hynix management was attempting to lower expectations.
Bulls remain optimistic due to higher-for-longer gross margins and capital return.
The stock is attempting to move above its 10-day and 20-day moving averages, which has historically been a signal for upcoming strength.
SNDK & MU trying to pop their heads above the 10d/20d MAs, which has usually been a decent signal for upcoming strength over the last couple of years.
Sentiment and positioning have significantly decreased compared to a few months ago.
Bears are still concerned about the slowing of 2nd derivative pricing and China supply risks.
Ardo Today at 1:56 AM my view: MU SNDK - clear negative, especially if a JV were to happen between hynix/intc/hyperscalers INTC - neutral / perhaps even negative; shows they cannot find better use for Ohio fab for either producing own CPUs or as a foundry;...
A potential JV between SK Hynix, Intel, and hyperscalers would be a clear negative for the company.
Memory investors reacted negatively to news of SK Hynix discussing a deal with Intel to make memory chips in the US by 2030, as it raised concerns about increased supply.
DRAM ASPs are pointing to a significant increase of +20-25% in Q3.
SNDK/MU: Hearing Clev pointing Q3 NAND/DRAM ASPs +20-40%/+20-25%, respectively SNDK/MU Clev Q3 NAND/DRAM ASP
Memory weaker (MU/SKHY/SNDK -4-5%), potentially on the back of some Deepseek V4.1 improvements:
UBS Asia raised DRAM estimates, anticipating a widening supply-demand gap in 2027 with blended ASPs increasing significantly and the sufficiency ratio falling, driven by unconstrained DRAM bit demand exceeding supply growth.
Mizuho says SemiAnalysis sees future HBM content potentially capped closer to 4– 8Hi, with 16–20Hi stacks/hybrid bonding still uncertain, which could free enough DRAM supply to support ~20–30% more GPU/ASIC shipments and actually raise aggregate DRAM bit de...
Q3 DRAM/NAND pricing is finishing about +20% blended, with server leading and mobile holding it back.
Q4 pricing is tracking +5-10% quarter-over-quarter, which is a low triple-digit rate of change compared to the previous year.
There is less upside for WFE from HBM layer count/hybrid bonding, prompting modest price target cuts.
The company is benefiting from surging DRAM costs, which are becoming too large for Nvidia to absorb, underscoring memory makers' growing pricing power.
The stock has fallen 27.5% from its high, indicating a substantial downturn.
BMO initiated coverage with an Outperform rating and a $1,300 price target, based on expectations of a prolonged memory supercycle driven by tight supply and broad-based demand.
ADI/NVDA/MU/MCHP/SWKS: BMO initiates ADI/NVDA/MU/MCHP at Outperform, SWKS at Market Perform ADI/NVDA/MU/MCHP/SWKS BMO ADI NVDA
BMO starts NVDA at Outperform/$340, highlighting 12+ months of largely sold-out capacity, a differentiated full-stack AI platform and strong margins/FCF; MU at Outperform/$1,300 on a prolonged memory supercycle driven by tight supply and broad-based demand....
Management indicates that CY27 memory supply/demand will be even tighter than CY26, with the company still unable to fulfill over half of some datacenter customer requests.
HBM remains a significant upside driver, with management confirming rising gross margins and industry HBM ASPs projected to increase by approximately 79% year-over-year into CY27 as existing contracts reset higher.
Visibility is improving with over 50% of revenue expected under strategic customer agreements, including about 10% take-or-pay, which should make earnings more durable than previously feared.
Management indicated that CY27 memory supply/demand is expected to be even tighter than CY26, suggesting strong future demand.
Management confirmed that HBM gross margins are expected to rise, indicating improved profitability for this product segment.
Visibility is improving, with over 50% of revenue anticipated from strategic customer agreements, including approximately 10% under take-or-pay contracts, providing revenue stability.
The company is currently fulfilling less than 50% of some datacenter customer requests, indicating a supply constraint that could limit immediate growth.
SanDisk’s new long-term model provides a useful framework for thinking about Micron as a structurally higher-earnings business rather than just a peak-cycle memory trade.
Applying SanDisk-like assumptions, Micron's EPS could potentially reach $230+ by FY30, supporting a 12–15x P/E multiple.
Capital returns are an underappreciated upside lever, with the potential for approximately 10% annual share repurchases once CHIPS Act restrictions expire.
The company is a positive read-through from strong 2027 AI-server demand, with Vera Rubin potentially exceeding 50K racks and production ramping from late 2026.
MU: BofA Reiterates Buy and $1,550 PT, Arguing SNDK’s Framework Supports a Structural Re-Rating of Memory
UBS reiterated Buy and a $1,625 price target, as DRAM fundamentals have tightened further, with 2027 HBM blended ASP now expected around $79/Gb, a 67% year-over-year increase.
HBM4E pricing should exceed $30/Gb, and stronger server/storage SSD demand still supports NAND revenue growth of roughly 23%/26% in 2026/27.
UBS raised CY28/CY29/CY30 EPS to $74.13/$184.89/$265.65, and the company's ability to remain above $160 of EPS in CY29 with over $450B of cumulative FCF through CY28 points to a structurally higher earnings and valuation framework.
Mizuho reiterated Outperform and a $1,375 price target, as DRAM/NAND supply should remain tight through 2027 with little incremental supply until 2028.
Micron’s ~16 strategic customer agreements should cover over 50% of bits through 2030 and support 80%+ gross margins, with newer products carrying pricing premiums.
CXMT remains technologically behind in advanced DRAM/HBM, and HBF adoption is more likely a 2–3 year story, reinforcing Micron’s favorable near-term positioning.
TD Cowen's FMS discussions reinforced confidence that DRAM remains the most viable near-term medium for AI KV cache, supporting resilient DRAM demand through at least 2028.
Management's core message is that the memory shortage is worsening, with CY27 looking tighter than CY26, and DRAM becoming a key AI infrastructure constraint.
Customers are requesting substantially more memory even at elevated prices, and Micron can only fulfill about half of some data center requests, indicating strong underlying demand.
Agentic workloads are estimated to consume significantly more tokens than chat, and HBM continues to cannibalize conventional DRAM capacity, further tightening the market.
The memory cycle is becoming structurally less cyclical due to long-duration/take-or-pay SCAs and qualification-driven HBM programs, which should support stickier share.
Longer term, physical AI/robotics could potentially dwarf today's data center memory opportunity, presenting significant growth potential.
NAND pricing is somewhat softer than previously expected, although stronger server/storage SSD demand still supports NAND revenue growth.
The stock declined despite positive commentary at a Keybanc conference, suggesting that the positive outlook on memory shortages did not fully translate into stock performance.
Morgan Stanley remains bullish on memory, arguing that the recent correction reflects an aging cycle rather than a break in the AI demand thesis, and the rapid expansion of multi-year LTAs improves visibility and reduces volatility.
The company has a significant number of deals covering a substantial portion of its DRAM and NAND volume through 2030, which should better anchor pricing and supply discipline, supported by AI capex and agentic workloads.
The stock declined in sympathy with SNDK's disappointing results.
SNDK -7% / MU -1.3% after SNDK’s disappointing results. We covered our thoughts in the morning wrap here in case you missed it and some good discussion in Slack here
The recent stock pullback offers an attractive entry point, as long-term AI demand and hyperscaler capex support pricing power and utilization.
The analyst sides with the bulls, believing that contract pricing is still rising, long-term agreements indicate sustained strength, valuation is attractive, supply/demand remains in undersupply, and the strategic value in data centers is increasing.
MU +1% / SKHY -70bps. Memory narrative more muddied than May/June as bears continue to point to 2nd deriv pricing rolling over, Chinese capacity coming online in the future (eg, Reuters talking up CXMT’s 2nd chip plan in Beijing this morning), and pricing b...
Citi removed its 90-day upside Catalyst Watch due to mixed Q2 memory peer earnings, suggesting a less clear positive outlook for the company.
Bears point to a potential rollover in second derivative pricing, future Chinese capacity coming online, and pricing nearing consumer elasticity limits.
Susquehanna's checks indicate that memory remains structurally tight and is expected to reaccelerate in September/October as Chinese customer budgets reset.
US names following earlier with SNDK/MU/WDC/STX -7% early. Many other names down 4-5% as momentum looks to take another hit early. Good chart from our friends at Jefferies showing the magnitude of the momentum unwind, the worst going back to 2000:
Hard looking at the GPL log chart on Bloomberg, but SNDK now down close to about 50% from the peak at sitting at 61.8% fib retracement on the chart and just blew past what we thought was a decent r/r near $1,500. Memory trade continues to suffer from downti...
Micron is pushing back against Apple's lobbying to use Chinese memory chips, warning that opening the door to subsidized Chinese suppliers could undermine the domestic memory industry and conflict with U.S. semiconductor-security goals.
Memory stocks, including MU, suffered due to Trendforce data indicating slowing ASP increases for NAND in Q3 compared to Q2, and rumors circulating about NAND.
Not a good day for memory with SNDK -11%, SKHY -8%, MU -2% after Chinese DRAM maker CXMT was +466% overnight.
KeyBanc reiterated an Overweight rating and raised its price target, citing benefits from memory pricing improvements.
DRAM and NAND pricing are expected to rise significantly quarter-over-quarter.
HBM resets are driving substantial year-over-year price increases.
KeyBanc reiterates Overweight, raises estimates and lifts its price target to $155 from $110, citing outsized server CPU demand from agentic AI, with Intel capacity supporting 25%–30% unit growth this year and more than 50% next year. The firm says 18A yiel...
AMD/MU/MRVL: KeyBanc Raises Targets as AI Compute, Memory Pricing and Custom Silicon Visibility Improve
SNDK +5/MU +5% - This chart was getting passed around from a mid-day note by Shawn Kim:
Memory sentiment has declined following the company's earnings report, driven by concerns about limited upside in stocks and pricing, flat spot prices, a negative inflection in contract prices, and increasing supply from China.
If the projected year-over-year memory pricing holds and the stock trades at approximately 2x next twelve months price-to-book by mid-2027, the stock price would remain around its current level.
The stock could be range-bound in the near term, requiring careful consideration for entry points.
The company is positioned as a key technology and memory enabler in the AI data center market.
Memory names were the weakest, with the stock down 5% as Asian names experienced double-digit declines.
Korea fell 9% and is dragging down semis -3% with it early while Software +50bps and Internet +10bps are green. Memory names the weakest as Asian names were all down double digits: SNDK/WDC/MU/STX -5%.
The stock experienced a decline.
MU +1%: As we wrote yesterday, SNDK one of the few names where stock was at a place you could really put some nice round numbers on the r/r bear case math ($300 x 5x = $1.5k). $1,500 was hit early again in the premarket this morning and stock bounced nicely...
Memory pricing is expected to peak and normalize from 2HCY27 into CY28 as new long-term agreements are signed and supply-demand conditions rebalance.
Citi added an Upside 90-Day Catalyst Watch, expecting stronger AI-driven DRAM pricing to drive upside into earnings.
The firm raised its 2Q/3Q/4Q26 DRAM ASP forecasts to +44%/+20%/+13% QoQ (from +37%/+13%/+11%).
Upside remains for the approximately 60% of Micron’s sales not yet covered by long-term agreements.
In the U.S., Software +1% outperforming Semis -1% early with the former helped by a few upgrades from Difucci at Guggenheim (CRM, NOW, CHKP). Memory following Korea memory names lower (SNDK -5%, MU -3%). HDDs and Semicap also weak to start the new month.
Jefferies says new Strategic Customer Agreements fundamentally transform Micron’s business model by replacing short-term, non-binding contracts with 3-5 year take-or-pay agreements backed by $22B of customer financial commitments.
Management expects LTAs to ultimately cover ~50% of company sales, providing significantly greater demand visibility and more predictable capex planning while reaffirming that memory demand will remain above supply through 2027 and into 2028.
MU IR highlighted growing AI-driven demand from agentic AI, KV cache, and Storage Next workloads, emphasizing that HBM and conventional DRAM should be viewed as complementary rather than substitutable.
Cantor raised its price target, arguing that the company's long-term Strategic Customer Agreements fundamentally extend the memory cycle by locking up a significant portion of future supply.
The firm sees a path to $200 EPS in CY27 with upside to $250 in CY28, believing more predictable pricing, sustained high gross margins, and significant buybacks create a materially higher earnings framework than investors currently appreciate.
The company is a key beneficiary of the raised global semiconductor TAM forecast, driven by stronger AI infrastructure demand and memory pricing.
Needham materially raised estimates and its PT to $1,550, arguing robust AI-driven demand, limited industry capacity additions, and favorable long-term supply agreements are extending the memory upcycle.
The firm expects DRAM pricing to continue rising through FY28, with contract pricing potentially moving another 50-100% higher and DRAM pricing reaching roughly $0.27-0.28/GB over the next year.
Needham now forecasts FY27/FY28 EPS of $140/$155 versus prior estimates of $92.50/$116.50, believing investors remain too conservative on both ASPs and earnings power as hyperscalers continue increasing AI infrastructure spending.
The stock was strong ahead of earnings, with positive pricing checks circulating, including Jefferies' expectation of DRAM pricing increasing 30% in Q3 and another 40% in Q4, exceeding buy-side expectations.
Micron and Anthropic announced a strategic agreement covering memory and storage AI architecture design, supply and demand, enterprise adoption of Claude, and a strategic investment in Anthropic's Series H funding round, directly linking frontier AI model demands to infrastructure design and deployment.
Memory strong with MU +7% and SNDK +4% ahead of MU’s earnings on Wednesday as more positive pricing checks make the rounds — this time it was Jefferies saying they see DRAM pricing increasing 30% in Q3 and another 40% in Q4, which is above where the 20 and...
Citi raises FY26/FY27 EPS estimates due to stronger memory pricing, with FY27 EPS now above Street expectations, as AI-driven demand continues to outstrip supply.
Investor focus is increasingly shifting toward long-term agreements (LTAs), which Citi believes can sustain elevated pricing and earnings power beyond the current cycle.
(MU/SNDK +2-3%; INTC/AMD/ARM +2-3%).
The stock showed positive performance within the memory sector.
Memory mixed with SNDK -1.5% but MU +2%.
The company's stock showed strength, contributing to the overall market's upward movement.
In Techlandia, strength in semis continues although more bifurcated this morning with HDDs/Memory leading us higher (WDC/STX +8-10%) and SNDK/MU +2-3% while CPU names softer (ARM -2%, INTC -1% although AMD +50bps)
RBC raised its price target to $1,200 from $525, arguing the current DRAM cycle is in the early stages of a longer AI-driven upcycle.
The firm expects DRAM pricing to remain stronger for longer due to limited cleanroom additions, ongoing HBM conversion, and robust demand from inference and agentic AI workloads, with contract pricing tracking ahead of expectations.
RBC sees HBM becoming a larger portion of Micron's revenue mix, supporting structurally higher margins and a higher valuation multiple through the cycle.
While LTAs improve visibility, RBC believes the bigger story is persistent supply constraints and secular growth in memory content driven by AI.
Alethia raised its price target to $1,600, arguing investors underestimate the magnitude and duration of the memory cycle.
The firm expects server DRAM pricing to rise another 30% in Q3 2026 and HBM pricing to more than double in CY27, driven by sustained AI demand, limited cleanroom expansion, and growing memory content.
Alethia estimates memory's share of AI hardware value could exceed 70% by CY27, up from roughly 40% today, as HBM, DRAM, and memory-rich AI architectures become increasingly important.
The combination of stronger pricing, rising memory content, expanding LTAs, and accelerating FCF generation leads Alethia to believe memory deserves a higher valuation multiple than in prior cycles.
Memory is becoming critical in AI systems, indicating a strong demand driver for the company's products.
The DRAM cycle has plenty of room to run, suggesting sustained positive momentum for the company.
Analysts raised price targets, with Alethia setting a $1,600 PT and RBC setting a $1,200 PT, reflecting increased confidence in the stock.
Memory strong as well: MU +11% and SNDK +6%. Alethia and RBC had positive notes on MU raising PT to $1,600 and $1,200 arguing memory is becoming critical in Ai systems and DRAM cycle has plenty of room to run.
Edge is raising prices again, with Q3 server DRAM now expected to increase by 15-20% compared to the prior expectation of 6-8%.
Wolfe raised its FY26/FY27 estimates, driven by higher DRAM pricing in FY26 and higher HBM pricing in FY27, arguing supply remains constrained through at least CY27.
Wolfe expects DRAM pricing to rise ~200% in CY27 and NAND pricing ~17%, supported by limited cleanroom capacity additions and strong demand from AI GPUs and agentic CPU-driven workloads.
Long-term agreements (LTAs) are becoming a structural positive for the industry, improving supply visibility, reducing cyclicality, and potentially supporting higher valuation multiples.
HBM pricing is seen as a key CY27 catalyst, with contract negotiations later this year likely providing another leg of estimate upside.
In Tech, semis are leading the way higher with INTC (+5%) up on a dbl ug at BofA, Memory following Asia strength overnight (SNDK/MU +3-5%), and semicaps looking to be the first AI semi group to be regaining 52wk highs (ASML/AMAT +5%).
SNDK +30bps outperforming MU -1.4%. We had some notes yesterday around why the SoCAMM DRAM cuts are actually bullish NAND.
Cantor reiterated Overweight ratings and Top Pick status on MU, raising its price target, as it believes the memory cycle is structurally different from prior upcycles. The firm expects DRAM and NAND to remain undersupplied through CY27 due to robust AI-related demand, emerging CPU-driven memory growth, and an inability to add capacity fast enough. Growing adoption of LTAs, covering roughly 50% of industry bits by year-end with prepayments and pricing corridors, improves visibility and reduces cyclicality. Stronger-for-longer pricing assumptions led to raised CY26/CY27 estimates, with further upside expected if pricing accelerates beyond current expectations.
Wells Fargo more than doubled its price target on MU, arguing the market continues to underestimate the pace of memory pricing and the durability of the current upcycle. The firm lifted FY27/28 EPS estimates significantly, driven by materially higher DRAM and NAND pricing increasingly supported by Strategic Customer Agreements (SCAs) that provide pricing visibility and duration. Wells Fargo expects DRAM and NAND supply to remain below demand through CY27, with the first meaningful DRAM greenfield capacity not arriving until mid-CY27, and believes peak earnings power could approach $200/share. Upcoming commentary around SCAs and HBM-related agreements are seen as key catalysts for a re-rating.
Mizuho sees growing TPU, OpenAI, and MTIA deployments as incrementally positive for memory and storage demand, reinforcing its bullish view on MU.
MU/SNDK: Cantor Raises PTs Again, Says LTAs and Structural Undersupply Are Rewriting the Memory Playbook
Cantor reiterated Overweight ratings and Top Pick status on MU and SNDK, raising PTs to $1,500 and $2,900, respectively, as it becomes increasingly convinced the
Memory stocks, including MU, are bouncing back, with the street defending the sector given recent SoCAMM reductions at NVDA. Analysts expect no meaningful reduction in overall SoCAMM demand as memory suppliers remain capacity constrained and Nvidia prioritizes shipment volume, which also helps NAND demand. With less DRAM on board following cuts, the CPU tier now maxes out at smaller batch sizes, pushing near-real-time workloads onto NVMe offloading earlier than planned, increasing NAND demand through NVDA’s Context Memory eXtension (CMX) technology. Buy-side expectations for CY27 EPS are $160+, and a unique valuation bull framing suggests that even with a 50% price ding in 2028, EPS could easily exceed $100 in 2029, leading to significant cash generation and a sub-10x earnings multiple.
Memory bouncing back nicely with MU +9% and SNDK +5% as street was out defending what was obvious late last week given the SoCAMM reductions at NVDA. Citi anod others said they expect no meaningful reduction in overall SoCAMM demand as memory suppliers rema...
Memory: Hearing Cleveland positive as they raise ASPs on SNDK and MU citing strong demand/pricing, calling out LTAs for SNDK well above current market prices (near 50c/GB)
Morgan Stanley raised its price target to $1,050 from $520, driven by significantly higher DRAM and NAND pricing assumptions.
Supply-chain checks indicate DRAM pricing rising roughly 40% quarter-over-quarter in May and another 20%+ in August, with no near-term fix to industry shortages.
Morgan Stanley lifted CY27 EPS estimates by 48%, citing stronger pricing, expanding margins, and accelerating demand from AI and hyperscale customers.
The stock remains inexpensive at sub-10x CY27 earnings despite a multi-year favorable supply-demand backdrop, with multiple expansion expected to be the larger driver of upside.
MU/SNDK: Morgan Stanley Raises PTs Sharply on Higher Memory Pricing, Sees Supply Tightness Lasting 2-3 More Years
MU +3% / SNDK -2.5% as SK Hynix said they planned to 2x capacity. Good discussion here in Slack on potential hardware risks
The company's stock was negatively impacted by SK Hynix's announcement to double its capacity, raising concerns about potential hardware risks and increased competition.
Memory: Mizuho Raises PTs Across MU/SNDK/ARM/DELL/ON on Agentic AI + HBM Tightness
Mizuho reiterated MU as its top pick and raised PTs across $MU ($1,150), $SNDK ($1,825), $ARM ($360), $DELL ($350), and $ON ($150), arguing the market still underestimates the scale/duration of the Agentic AI infrastructure and memory buildout.
The company's stock was up significantly, nearing a key price target, indicating strong market confidence.
Barclays remains bullish on the broader memory/storage trade, arguing LTAs are materially improving pricing visibility and reducing historical cyclicality.
The firm highlighted continued DRAM/HBM tightness for the company.
Barclays raised its price target, arguing the market still underestimates the durability of AI-driven storage demand and structural pricing improvement.
MU +8% getting close to $1k. Lots of your favorite AI semi names up 4-6%. Internet and Software down early, the latter led lower by ZS’s -25% horrid print, which is bringing down other cyber names as well (PANW/CRWD -4%). Interestingly, hyperscalers largely...
Memory: Barclays Says MU/STX/WDC Still Have Another Leg Higher as LTAs Improve Visibility
The firm highlighted stronger TSMC 2nm capex, Samsung/MU HBM ramps, Japanese foundry expansion, and continued China domestic WFE growth as key tailwinds. Mizuho believes current estimates for $LRCX, $AMAT, and $MKSI remain too low versus the emerging AI/HBM...
Memory stocks generally outperformed after Barclays' upgrade of SNDK, with MU showing strength.
UBS raised its price target significantly, arguing that new long-term agreements (LTAs) are fundamentally changing the memory industry by improving visibility, smoothing cyclicality, and supporting structurally higher earnings power.
UBS believes enhanced LTAs now cover a significant portion of industry DDR bit shipments and server DDR5 volumes, which helps stabilize pricing and secure supply for hyperscalers.
UBS expects the company's EPS to remain comfortably above $100 through at least C2029 and forecasts substantial cumulative FCF from 2027-2029, while also highlighting continued upside from HBM pricing/tightness and AI-driven DRAM demand growth.
Mizuho believes AI-driven memory demand is stronger and more durable than investors appreciate, with undersupplied DRAM/NAND markets likely extending through 2026-27.
Mizuho highlighted potential significant increases in HBM4e pricing, ongoing customer shortages, and accelerating LTA activity as hyperscalers secure supply.
Mizuho argued that agentic AI and CPU-driven memory demand are increasingly important, estimating a significant incremental DRAM demand opportunity from NVIDIA's potential CPU revenue, and maintained the company as its top memory pick with an $800 PT.
A sell-side analyst raised their price target significantly, reiterating a bullish case based on long-term agreements and memory becoming a more strategic component of the data center stack.
Meetings with management led to an increase in financial projections.
The market is currently rewarding believers in the ultimate bull case, and a 10-15x multiple for the stock is not unreasonable given the tailwinds for NAND/DRAM over the next few years.
The firm highlights NVDA’s NVL SuperPOD scaling to ~17 exabytes of memory (~240x density vs. NVL72), underscoring memory as a key bottleneck and value driver. Aletheia adds the evolving balance between compute and memory will define AI factory economics, re...
Memory underperformed (MU -1.4%/ SNDK -4.5%) as Samsung strike gets resolved (means no supply disruption there) and some Asian press had some things to say about Chinese Memory players. Here’s cwnewsroom via Jukan:
The company underperformed as the Samsung strike was resolved, removing concerns about supply disruption, and due to Asian press commentary on Chinese memory players.
The company exhibited lower highs on negative RSI divergence, similar to other memory and HDD names.
Memory underperformed (MU -1.4%/ SNDK -4.5%) as Samsung strike gets resolved (means no supply disruption there) and some Asian press had some things to say about Chinese Memory players. Here’s cwnewsroom via Jukan:
INTC +5% leading the way higher following Lip Bu Tan’s bullish JPM Conference appearance yesterday (we sent out notes here in case you missed it). The rest of AI semis strong as well with MRVL +4%, MU +4%, ALAB +3% among others as we head into NVDA earnings...
AI server demand continues driving powerful memory tailwinds, with HBM4 ramps in 2H26 and increasing CPU-attached DRAM content pushing DRAM/NAND pricing materially higher into 2027.
AI racks are driving greater than 4:1 trade ratios versus traditional DRAM nodes alongside rising LPDDR5/SSD content for inference workloads, while supply remains tight amid constrained NAND/DRAM capacity.
Mizuho raised its price target, citing stronger HBM/eSSD demand, tightening NAND supply, and potential additional upside if HBF ramps accelerate.
The memory sector, including the company, showed strength, with the stock up 2.5% ahead of its appearance at the JPM TMT conference.
Memory strong (MU +2.5%/SNDK +3.75%) as both are at JPM TMT’s conference tomorrow. Citi raised SNDK PT to $2k+highlighting ASPs more than doubling q/q at Kioxia, continued supply discipline, and growing LTA adoption with financial guarantees that improve ea...
SNDK -5% despite Kioxia being limit up although finished well off the lows; MU -6%. Both companies will be at JPM on Wednesday.
Memory weaker ahead of Kioxia’s print abroad: SNDK -4.5% / MU -3.5%. A couple negative things we called out in our morning note:
The stock experienced a negative movement due to weaker memory market conditions ahead of Kioxia's print abroad.
Memory weaker ahead of Kioxia’s print abroad: SNDK -4.5% / MU -3.5%. A couple negative things we called out in our morning note:
The stock was down early in pre-market trading, aligning with a broader weakness in tech semiconductors following a decline in the Korean market.
In Tech semis -2% following Korea weaker early in the pre market while sw the main outperformer. Recent AI leaders all down early: SNDK/MU -3%, INTC -5%,
The stock showed weakness following overnight trends in Korea's memory market.
Memory weaker with SNDK -6% and MU 3.6% following the weakness in Korea overnight…not a ton of news here.
AI is fundamentally reshaping memory demand and supply dynamics, supporting a more durable and profitable cycle.
Rising memory intensity across AI workloads (HBM, DRAM, NAND) and structurally tighter supply from slowing Moore's Law and HBM wafer tradeoffs are expected to keep industry conditions tight through CY26.
Strategic Customer Agreements are improving visibility and through-cycle margins, and Micron's execution and mix shift position it to disproportionately benefit.
The firm raises its price target to $1,000.
The stock was strong early, leading the way higher in AI semiconductors.
In tech land, HUBS -22% blow up showing app sw not in same league as consumption sw after DDOG’s beat yesterday, AKAM rallying on a big deal, NET down on an expects miss, MCHP the latest analog to beat, CRWV down on a weaker guide, ABNB looks solid, QCOM co...
The stock performed strongly, indicating positive sentiment in the memory sector.
Memory names are leading the market higher, with the stock up double digits.
AI inference creates exponential memory demand, driven by the need to store KV-cache for token generation and the growing context window lengths.
The company is innovating across the entire memory hierarchy, from HBM to context storage SSDs, to address the memory wall bottleneck in AI inference.
The company is undertaking five global fab expansions and moving to newer process nodes to meet demand, despite current capacity constraints.
The company is driving innovation with HBM4, delivering significantly higher bandwidth, and engaging in deep co-design with partners across the industry.
AI is 'warming' cold storage, driving SSD demand as AI agents access previously archived data, and future agentic workflows will require persistent memory stored on SSDs.
Concerns about AI monetization are overblown, as the technology's long-term potential to transform society will eventually lead to monetization.
The company is already capacity-constrained despite significant fab expansions, indicating that demand is outstripping supply.
TD Cowen indicates that long-term agreements (LTAs) are structured around a gross margin framework of 60% at the low end and high 80s at the high end, suggesting durability.
AI-driven demand, including CPU-side inference, could extend the memory cycle, and the firm remains constructive on multi-year durability.
DA Davidson initiates with a Buy rating and a $1,000 PT, arguing that AI is extending the memory cycle through a positive feedback loop between compute deployment and demand, supporting structurally higher pricing and earnings power.
The market is seen as underestimating memory demand relative to the broader semi complex, with Micron's node leadership reinforcing its positioning and enabling long-duration earnings upside.
The stock has already priced in much of the upside, with approximately $110 CY27 EPS, and incremental AI-driven demand is not considered game-changing.
Near-term EPS upside is limited, and the firm is cautious due to gross margin normalization towards ~60% and transition timing risks.
Melius initiated coverage with a Buy rating.
Google's TPUv8 reinforces strong attach for ARM-based CPUs and increasing HBM/SRAM content, driving upside across Micron.
AVGO/ARM/MU/CRDO/LITE: Mizuho Highlights TPUv8 as Catalyst for Networking and Memory Upside
The firm says GOOGL’s TPUv8 reinforces strong attach for ARM-based CPUs and increasing HBM/SRAM content, driving upside across AVGO, ARM, and MU. The analyst notes architectural shifts (boardfly topology, OCS, Axion CPUs) improve bandwidth and reduce latenc...
The stock showed strong performance, driven by positive sentiment around HBM pricing.
Melius initiated coverage with a Buy rating and a 2-year target, believing the AI memory cycle could continue through the end of the decade, leading to higher market valuation for the durability of margin and demand profiles in HBM, DRAM, and NAND markets.
In Tech land, semis broke their 18-day win streak although plenty of names were green on my screen. The early morning semi swoon was catalyzed by a POET -50% blow up early morning, one of retail’s favorite “undiscovered” optical plays after MRVL pulled a pu...
SNDK +8% slicing through $1k / MU +5% strong today: On DRAM, heard Asia chatter around HBM pricing being better than expected. Morgan Stanley and Cantor were bullish on SNDK raising PT (MS to $1k and Cantor to $1.4k) both pointing out the usual goodness: #s...
We initiate coverage of Micron (MU) and Sandisk (SNDK) with Buy ratings and a view that the AI memory cycle could keep going through the end of the decade and that the market will eventually be willing to pay more for the unusual durability of the margin an...
Memory names like MU are viewed more skeptically despite improving fundamentals, offering a more attractive way to play the AI cycle given better risk/reward and re-rating potential.
STX/WDC/MU: UBS Prefers Memory Over HDD; Sees Valuation Disconnect and Cyclical Risk
Morgan Stanley favors the company as a cleaner way to play the CPU strength trend, citing valuation and upside in memory.
INTC/AMD/MU/SNDK: MS Raises INTC PT to $56; Prefers Memory as Best Play on CPU Strength
Morgan Stanley says INTC and AMD should benefit from improving CPU demand driven by agentic AI, but sees memory as the cleaner way to play the trend, favoring MU and SNDK on valuation and upside. The firm raises INTC PT to $56 and estimates on stronger serv...
The analyst heard positive feedback from Clev.
The company's stock saw a slight increase, partly due to Cleveland's note, which, while not as positive for DRAM, indicated expected price increases for NAND.
MU +22bps as Clev also had a note but didn’t sound as positive for DRAM as NAND with price increases expected to moderate. HDDs mixed with STX +3% / WDC
DELL +9% / HPE +5% helped by the AI trade but also signs that rate of change DRAM price increases starting to shift (see Clev MU comment above)
Conventional DRAM price forecasts are stabilizing somewhat after meaningful raises through Q1 2026.
Cleveland's note was not as positive for DRAM, with price increases expected to moderate.
Mizuho reiterated Outperform and raised its price target, citing tight NAND/HDD supply driving pricing through 1H27E.
The KVCache ramp is driving NAND demand as eSSDs deliver significantly higher read/write speeds than HDDs and contain more NAND content in inference racks compared to training-only AI racks.
The upcoming High Bandwidth Flash (HBF) in 2H26E, offering 6x HBM capacity at 100x TLC SSD bandwidth, could create new NAND supply constraints as 16-layer SLC-NAND stacks consume conventional capacity.
Memory/Storage: Mizuho Raises PTs Across MU/SNDK/STX/WDC — KVCache, QLC eSSDs and HBF Drive Inference Storage Inflection
UBS maintains conviction in the company's upcycle despite recent negativity, citing continued DRAM and NAND pricing strength, with C1Q26 DDR contracts up approximately 95% quarter-over-quarter and NAND up 80%.
Hyperscalers are actively negotiating longer-term HBM contracts with pre-payments, and the company, along with SK Hynix and Samsung, aims to rebuild HBM pricing premiums into C2027, with DRAM shortages potentially persisting into C2028.
Memory names leading the way higher (SNDK/MU +9%) but plenty of names up 5%+ across the board.
The stock experienced a significant decline in early trading.
The stock was leading early with other AI names.
The firm maintains a Buy rating, arguing that TurboQuant concerns about memory demand are overdone, similar to DeepSeek, where efficiency gains ultimately unlock more usage.
Memory peers have begun negotiating long-term strategic agreements with hyperscalers, which should provide support to contract prices despite spot weakness.
Memory remains structurally tight, with AI-driven demand sustaining supply/demand imbalance and supporting higher pricing for DRAM and NAND over the next 2-3 years.
Hyperscale/AI datacenter demand is receiving priority allocation, and suppliers are shifting mix toward higher-value segments, driving margin expansion.
DRAM pricing is accelerating sharply while bit supply remains constrained, reinforcing favorable pricing dynamics.
SNDK +10% / MU +4% spread continues to widen as investors shift preference to NAND. Added some here (SNDK) as r/r $600 too hard to ignore when market turns risk on. Think this whole NAND over DRAM preference gives it even more juice as most “savvy” investor...
The target price was cut to $425 from $510 due to a pullback in DRAM spot prices.
Investors shifted preference away from DRAM towards NAND, causing the spread between SNDK and MU to widen.
The stock was down significantly, with the spread between MU and SNDK approaching double digits since MU's earnings, and the narrative continues to shift in favor of NAND over DRAM.
There is increasing discussion about declines in DDR5 prices, mainly driven by consumer de-stocking, with specific examples of significant price drops for Corsair memory modules on Amazon US.
DDR5 spot prices have been flat to down over the last couple of months, contrasting with NAND holding up.
The stock is already down 35% from its peak, with the 150-day moving average 12% lower and the 200-day moving average over 20% lower.
SNDK -7% as spread here now approaching double digits since MU’s earnings and we sense the narrative continues to shift in favor of NAND (a shift that started 2-3 weeks ago). A lot of talk on X over the weekend about declines in DDR5 prices, mainly driven b...
FundaAI views KV cache compression as supportive of sustained memory demand, as efficiency gains are typically reinvested into more state, longer context, and higher concurrency, rather than signaling structural demand erosion.
Morgan Stanley believes the memory selloff is overdone, as AI demand makes the current cycle different, with memory increasingly becoming the primary constraint on AI demand.
AI is consuming so much DRAM that it is crowding out PC and smartphone supply, cloud customers are paying expedite premiums, and HBM4 complexity will absorb new capacity.
Google's TurboQuant KV cache optimization is an evolutionary development with no surprises for memory, as the 6x gain applies only to KV cache, not total memory, and there is no evidence of universal demand reduction.
Aletheia reiterated Buy, stating that GTC 2026 reinforced that AI inferencing—especially in the agentic era—requires a far more diverse hardware stack beyond GPUs + HBM, with increasing reliance on DRAM, LPDDR, SSD, and SRAM to support low-latency, multi-step workflows.
Aletheia highlighted NVDA’s NVL SuperPOD scaling to ~17 exabytes of memory (~240x density vs. NVL72), underscoring memory as a key bottleneck and value driver.
Aletheia added that the evolving balance between compute and memory will define AI factory economics.
The company's HBM4 roadmap shows material throughput gains, reinforcing memory as a key bottleneck and beneficiary of the AI scaling cycle.
The memory trade is considered fine, and the risk/reward under $500 skews upwards, with bulls targeting $120+ for CY27 EPS.
In particular - as we mentioned in our morning wrap - we preferSNDK +2.5% even more after the MU print as 1) MU print pointed out NAND upside and 2) the print — and overnight conversations among investors — made clear that SNDK still has more room for earni...
Despite objectively better numbers, there was noise overnight about "peak GMs" and FY26/FY27 capex raises potentially leading to an overbuild.
Wolfe expects DRAM prices to rise approximately 100% year-over-year in CY26 and NAND prices to rise approximately 95%.
Wolfe projects blended HBM pricing to increase approximately 8% annually through CY27, driven by long-term agreements despite legacy price declines.
The stock performed strongly despite reports of losing out on HBM4 in Vera Rubin, suggesting resilience or skepticism about the reports.
SNDK +12% with a nice bounce, strongest out of the memory group although rest were also strong with MU +5% despite press reports out of Asia they lost out on HBM4 in Vera Rubin (debate how true this is); STX/WDC +6-7%
The company is investing $200 billion to expand its manufacturing capacity, including two new chip factories in Boise and a $100 billion fab complex near Syracuse, to address the AI memory bottleneck.
The stock experienced weakness along with other memory and HDD names.
The company has seen significant stock appreciation, outperforming Samsung by a large margin since the fall, despite concerns about its HBM progress.
The analyst is uncertain whether HBM fears will persist given the ongoing tailwind of DDR and NAND pricing.
The stock declined after reports that Samsung plans to restart mass production of HBM4 for NVIDIA's Vera Rubin, raising concerns about Micron's HBM progress.
The company, along with other major memory chipmakers, is reportedly scrutinizing customer orders to prevent intentional inventory hoarding amid persistent supply tightness.
The company has tightened requirements by asking customers to disclose end customers and order volumes to ensure genuine demand and prevent overbooking or excessive stockpiling.
HSBC raised its price target and maintained a Buy rating, citing accelerating DRAM pricing, inference-driven demand, and supply discipline.
The firm raised its FY2Q operating profit estimate and sees NAND as the next upside catalyst, with AI workloads driving incremental server DRAM demand.
Capex expansion from 2H27E is expected to extend tight supply and support earnings momentum.
The stock rose, following the trend of other memory and storage companies.
MU +6%, STX/WDC +6-8% following along as well
A director's purchase of shares, the first since 2022, is a positive signal, exceeding the mandated ownership requirement for board members.
The director's share purchase, while positive, represents a negligible portion of his net worth, suggesting limited personal conviction despite exceeding the mandated ownership.
The analyst reiterates a constructive super-cycle view with execution intact and a path to $40+/share in EPS.
AMAT / LRCX / KLAC / ADI / MPWR / HDDs / MU / SNDK: Wells Fargo Raises Multiple Names to OW on WFE, Analog, and Memory Upside
Citi removes the stock from its US Focus List, citing expected deceleration in DRAM pricing momentum after a strong 1Q increase, though longer-term memory supply/demand remains balanced into 2026–27.
Meaningful production from Micron's Idaho facility, including new fabs, customer certifications, and meeting AI client standards, is not expected until after all approvals are completed, effectively pushing substantial output gains into 2028.
The company is expected to benefit from a memory supercycle with strong pricing in 2026E.
AI strength is expected to persist across GPUs and ASICs (positive NVDA, DELL, AVGO), AI interconnect and optical with scale-up/out/across and 800G/1.6T (positive LITE, CRDO), and continued tailwinds for WFE (positive LRCX) alongside a memory supercycle wit...
UBS raised its price target on Micron and reiterated a Buy rating, reinforcing confidence in the durability of the current memory cycle after meetings with management.
Investors are underestimating how AI has structurally changed memory, particularly DRAM, into a more strategic, less cyclical asset, and are mispricing the sustainability of Micron’s earnings power.
Successive HBM generations become obsolete faster than legacy DRAM, discouraging inventory builds and dampening historical boom-bust dynamics.
The firm lifted CY26 and CY27 EPS estimates, arguing the improved earnings profile supports a re-rating.
AI-driven demand and supply constraints are driving a structural upcycle with HBM fully contracted through 2026.
Capex increases are focused on capability and transitions rather than oversupplying the market, supporting tight supply/demand dynamics.
The memory industry remains highly cyclical and commodity-like, meaning pricing could reverse quickly.
Rising DRAM prices could hurt demand in consumer electronics, and Samsung remains a competitive threat in HBM.
The market is underestimating Micron's earnings power from the DRAM rally and transitions to advanced technologies like HBM4.
Robust cloud service provider capex supports a strong operating profit growth forecast.
Investors are concerned about financial risks associated with neo-CSPs and specific projects like Stargate.
DRAM is expected to secure long-term contracts across the AI supply chain, mirroring other major AI hardware beneficiaries.
Worsening DRAM shortages and rising pricing power are expected to drive gross margins and EPS to record highs.
DRAM is expected to become a critical long-term contract component in the AI supply chain, driving significant EPS growth.
DDR gross margins are overtaking HBM margins due to capacity reallocations.
MU +5.5% as DRAM prices continue to spike. Citi was out raising PT to $240 saying that DRAM is poised to become the next key component to secure long-term contracts across the AI supply chain, mirroring the trajectory seen with NVDA, AMD, and AVGO — they sa...
MU/SNDK: Hearing Clev positive on both saying Hyperscale demand driving upside to pricing through 2026
DRAM prices are surging with no signs of cooling off due to constrained supply, with contract prices expected to rise significantly in Q4.
Al trade all getting a bid: CRWV +5% on positive ISI note; NBIS +2.8% on a bullish GS note; ANET +6% (didn't see anything new here); MRVL 5% #RSRI) Ip) ; ANET +6% (RMIAE) ; MRVL MU +5% as DRAM prices continue to climb [News] Korea, U.S.
MU +1% as Asian press reported Hynix raising HBM prices Al names broadly lower: NVDA -1%; AMD -1.49; AVGO -1.69; ANET -2%; ARM -29 AMD -1.4%; AVGO -1.69/; ANET -2%; ARM -2%/ TXN+1.4% as heard chatter about potential px hikes GFS -9% after guiding lower SOFT...
MU/Memory: Goldman Sachs cut its HBM TAM forecast for 2026 by 13%, from $51B to $45B, as Nvidia’s NVDA next-gen Rubin GPU is expected to keep the same HBM content as the Blackwell Ultra at 288GB, instead of a significant increase
Goldman cut its 2026 HBM TAM forecast by 13%, and Morgan Stanley warns of weakening HBM/DRAM pricing and oversupply risks in 2026.
Morgan Stanley warned of downside to HBM pricing once Samsung's supply enters the market.
Goldman Sachs reduced its HBM total addressable market forecast for 2026.
Research through 2026-11-02 · 2025-07-28 to 2026-11-02 · 114 source reports.
| Range | Prices through |
|---|---|
| 1W | |
| 1M | |
| 3M | |
| 6M | |
| 1Y | |
| 2Y | |
| 3Y |
Research through 2026-11-03. Artifact validity does not establish factual accuracy or research freshness.