Sean
@sean_________
TMT catalyst/event-driven trader with a focus on research.
3K Following 8.6K Followers
UBS on $MU: "now model MU HBM ASP up ~72% Y/Y (vs. +61% Y/Y prior) along with broadly unchanged MU HBM bit shipments of ~11.65B Gb for C2027. This comes despite changes to $NVDA Rubin Ultra specs, where our checks suggest a first VR300 variant is expected to use 384GB of HBM4 per GPU (8-Hi × 24Gb × 8 × 2 dies), with a subsequent version moving to 512GB of HBM4E per GPU (8-Hi × 32Gb × 8 × 2 dies) around C3Q27 - and the latter being below our previous estimate of 768GB of HBM4E per GPU."
"slightly trim estimates in the near-term due to a wider range of potential pricing outcomes for NAND through year-end and into C1H27."
"For DRAM, we believe HBM supply/demand has actually tightened with pricing for both HBM4 and HBM4E even stronger than our prior expectations."
"Overall HBM consumption for C2027 is actually increasing from 58.7B Gb to 61.5B Gb. We are also getting confirmation that memory suppliers are re-widening the HBM/DDR price premium..."
Show more
Fubon- "the production push out in Rubin will bring some negative impact to the supply chain as Rubin has higher dollar content compared to Blackwell"
$MU: "recent checks suggest that $NVDA HBM4 cost will increase to US$31-32 per GB from US$17-18 per GB (HBM 3e) in 2026."
"We estimate nVidia’s Rubin GPU price will skyrocket to US$78-80K assuming nVidia is still able to maintain its GM at 75-80% in chips."
"For other GPU and custom-ASIC projects, we believe their HBM cost will be higher than nVidia’s and may reach US$35-36 per GB, meaning the HBM cost will double in 2027."
"Despite the market’s AI inflation concern recently, in our view, the token cost is still a more important factor in driving CSP’s capex, and we remain positive about AI market demand."
"Our checks suggest nVidia still plans to put the same amount of compute dies in its new rack architecture or system"
"Between the two racks, nVidia plans to use NPO/CPO for scale out connection while inside the rack, cable connection is still used for scale up connection"
"Based on our checks, Google plans to have 12-15mn TPUs in 2028. Entering 2028, Google’s TPUs will enter the V9 generation with four compute dies, suggesting that their capacity consumption will more than double in 2028 vs. 2027."
"Based on our checks, $INTC EMIB capacity will reach 10-12k by the end of 2026, and the plan is to double its capacity to 24-25k per month by the end of 2027."
"we see $TSM slowing down its SoIC expansion to prioritize the CoWoS expansion"
$AVGO: "we think it is possible that MediaTek could further raise its custom-ASIC market share in 2027."
Show more
Wolfe on $MSFT: "We think that the timing and size of MSFT’s disclosure could support the possibility that MSFT is one of $MU’s early SCA customers"
"We estimate ~$10B of our $15B CY26 Capex increase reflects higher memory prices, based on our estimate of memory accounting for ~30% of our prior C2H26 short-lived asset Capex at a ~40% price increase, with the remaining $5B reflecting additional caution around component costs and potential investments to bring capacity online earlier."
"we now estimate FY27 FCF to be negative, impacted by our $40B increase to FY27 Capex"
Show more
Jefferies on $NVDA Kyber/Backplane Delay- "The supply chain began to witness such possibility since May, and in recent few weeks, the absence of Kyber in 2027 has become a highly likely event, which means Rubin Ultra in 2027 will stick to Oberon structure (i.e., NVL72)"
"-ve for overall PCB supply chain, though we expect upstream to keep outperforming downstream, driven by ongoing near-term price hike, while +ve for copper cable makers." $CRDO $APH $TEL
Show more
1. Bullish for the NPO supply chain.
2. The spec downgrade of NVIDIA Rubin Ultra signals the erosion of NVIDIA’s performance moat.
3. Bearish for the CPO supply chain.
4. Big winners: AMD and the TPU ecosystem?
Show more
Analogs-
Mizuho: "we see as many as 6 OEMs implementing price hikes into 2H26E as demand for select components remains very tight. We see $NXPI, $TXN, IFX, Molex, $STM and Shanghai Belling all noting price increases with some noting price increases of up 5-30% as AI servers drive increased demand. We believe some analog suppliers are having to absorb some of the cost increases being handed out by wafer suppliers. Additionally, we continue to see suppliers of memory and MLCC raising prices, further straining the supply chain and increasing the cost for end consumers across auto and industrial end markets. We believe memory/storage component and AI DC power price increases will be able to be passed on to customers, with only high-end MCUs and HV MOSFET/IGBTs seeing higher prices passed on."
Evercore: Our $ON checks remain positive, and SYNA accelerates ON’s development of edge-AI processors and accelerates its push into edge and physical AI."
Show more
$QCOM: Great work by @StreetSignal__ on bogeys:
JPM Tech Spec: "most investors think this is a management that sets aggressive targets, so many expect an aggressively bullish tone – and quite a number say they expect to be shorting after the smoke clears from the event. Most folks I’ve spoken with expect either an update of the guide through F29, or discussion through C30 (though some did correctly underscore that they already informally guided F31 revenue during their investor day two years ago) and bulls sound like something at the higher end of $10~15b would be viewed positively. Bears seem very skeptical about the complexion and size of new order wins, with questions about who ASIC customers will be (particularly given a seeming lack of AI networking expertise, and SerDes that are viewed as lower-tier); more constructive investors are just looking at the TAM and assuming that as it disaggregates, and QCOM naturally wins some orders. Interestingly, a number of folks said they found some comfort in just how skeptical folks sound in conversations, suggesting expectations are not high."
Show more
Kudos to Wells and @sean_________ if $qcom announces $amzn tomorrow!
JPM pounding table on $AVGO: "we believe that Broadcom is on track to ramp its next-gen TPU v9 2nm (4 compute die, 16 HBM stacks, 400Gbps SERDES) in CY28 – contrary to recent sell-side, Asia supply chain, and various news outlets that have suggested delays or cancelations of this program"
Show more
Mizuho ASIC channel checks: pounding table on $AVGO
"potential >35M TPU shipments in 2028E (up 8x from 2026E ~4.3M and 2025 ~2.4M), which is significantly above our estimates of ~7M ASIC shipments for AVGO"
"We believe OAI ASIC is in development with 10GW AVGO project (Nexus) adding to TPU/MTIA/ARM roadmap."
"Our call noted $ARM AI ASIC chip development ongoing as we continue to expect a launch in late-2026E/early-2027E"
" $MU : With potential for TPU shipments to reach a cumulative 50M units in 2026-28E, we think ASICs could command an increasing share of the HBM market, especially as some of the next-gen ASICs move to HBM4e"
Show more
Mizuho semi supply chain tracker: no delays 800VDC & strong demand $NVDA CPO switch:
"In the HVDC/800VDC investment case, the deployment schedule should be an average of 12-month lead-time from the initial system (like power rack) shipment by vendors, in our view"
"Lumentum ( $LITE ) highlighted a rising adoption rate of InP-based NPO from its clients and we also expect stronger Google TPU demand to boost $AVGO VCSEL-based NPO sales.
"micro-LED (uLED) based optical solutions that are developed and promoted by other chipmakers, such as Credo ( $CRDO ) in ALC (LED cables) and Avicena (private), to capture future scale-up (XPU to XPU) and scale-in (XPU to HBM) opportunities."
Show more
Aletheia on $MU:
"We forecast MU’s EPS to jump 8.5x in CY27E, followed by a further 1.8x expansion in CY28E. This implies roughly 15x cumulative EPS growth and $350–400bn of FCF generation in FY26-28E."
"We now expect server DRAM ASP to jump a further 30% in C3Q26 (vs previous expectation of 10- 15%); this is likely to rise by another 10-15% in C4Q26 (same as previous expectation)"
"we now expect HBM ASP to double YoY in C2027"
"Our analysis shows that memory devices are becoming the most critical components in the AI hardware system as their combined content value are expected to cross over 70% in 2027 vs mid-40%s in 2025. For DRAM-intensive device such as Vera CPU, the SoCAAM alone contributes over 70% of BOM in 2H26; the full spec Vera CPU rack could reach a staggering $26M ASP per rack..."
Show more
$MU:
Wolfe: "never been before existed in memory, suggesting some multiple expansion opportunity."
MS: "in our view, this implies DRAM stocks could re-rate to at least 8-10x PE vs. 5x PE now"
Show more
$QCOM Wells Fargo:
“Based on company comments / our analysis, we see AWS as the potential lead hyperscale ASIC partner”
$QCOM: Digitimes reporting:
"Qualcomm has "more than one" ASIC project customer."
"It is understood that another of the four major US cloud service providers (CSPs) is currently collaborating with Qualcomm on a project."
Show more
$MU:
Wolfe: "never been before existed in memory, suggesting some multiple expansion opportunity."
MS: "in our view, this implies DRAM stocks could re-rate to at least 8-10x PE vs. 5x PE now"
Show more
MS on $GEV:
"Our view remains that gas turbine industry orders decrease in 2027 (vs 2026 levels of ~117GW). This view is driven by (1) an already high level of behind the meter gas turbine data center orders relative to build rates, and (2) our view that Middle East power infrastructure investments may see some delays"
"But, any way you cut the data, the Gas turbine market looks balanced at best in 2030, and more likely it starts to look over-supplied into the 2030"
Show more
Mizuho ASIC channel checks: pounding table on $AVGO
"potential >35M TPU shipments in 2028E (up 8x from 2026E ~4.3M and 2025 ~2.4M), which is significantly above our estimates of ~7M ASIC shipments for AVGO"
"We believe OAI ASIC is in development with 10GW AVGO project (Nexus) adding to TPU/MTIA/ARM roadmap."
"Our call noted $ARM AI ASIC chip development ongoing as we continue to expect a launch in late-2026E/early-2027E"
" $MU : With potential for TPU shipments to reach a cumulative 50M units in 2026-28E, we think ASICs could command an increasing share of the HBM market, especially as some of the next-gen ASICs move to HBM4e"
Show more
JPM catalyst, $QCOM: "we expect the company to outline targets for data center revenues which will include: 1) revenues in excess of $3 bn in FY27, and 2) revenues of $35 bn in FY31"
Show more
Cowen on $QCOM: 6/24 catalyst
"already has purchase orders in hand across multiple products, including custom ASICs, CPUs, and customer-specific engagements, such as its HUMAIN deal"
"Management pointed to a material revenue ramp beginning in the DecQ (F1Q27) and accelerating more meaningfully in fiscal 2027, with the potential to reach a multibillion-dollar scale."
Show more
Cowen on $QCOM: 6/24 catalyst
"already has purchase orders in hand across multiple products, including custom ASICs, CPUs, and customer-specific engagements, such as its HUMAIN deal"
"Management pointed to a material revenue ramp beginning in the DecQ (F1Q27) and accelerating more meaningfully in fiscal 2027, with the potential to reach a multibillion-dollar scale."
Show more
$QCOM: Digitimes reporting:
"Qualcomm has "more than one" ASIC project customer."
"It is understood that another of the four major US cloud service providers (CSPs) is currently collaborating with Qualcomm on a project."
Show more
Cowen TMT memory/storage takeaways:
"DRAM/NAND is becoming a big percentage of CapEx (went from 10-15% to +40%). This requires a meaningful shift in how hyperscalers operate given that they have used to quarterly pricing negotiations."
" $STX and $WDC both indicated that NAND pricing increase should not be a derivative for HDD pricing, and investors are clearly most enthusiastic about the pricing opportunity for HDDs"
"For $SNDK, management highlighted the unique structure of the LTAs it already signed. The floor gross margin is higher than what our checks for DRAM suggest (80% for SNDK vs our checks for 60% in DRAM, here)."
"SNDK already locked in 30% of its capacity next year under LTAs and the goal is to potentially increase that to 80% over time. With this level of margin structure, durability matters more than continued price increases."
Show more
GS on memory ( $MU / $SNDK ):
"The key demand driver of both the 2017-2018 upcycle and the current upcycle is strong server demand. While the 2017-2018 upcycle was driven by cloud data center expansion, the current cycle is driven by massive AI server build out...memory demand is now closely tied to server demand where server segment comprises around 50% of DRAM demand and ~40% of NAND demand."
"given the long lead time (takes at least ~3 years to construct a fab and produce output), we expect conventional DRAM capacity addition pace by 2030 to be slower than the 2017-2018 upcycle."
"Based on our checks, LTAs are being discussed to include sizable prepayments and meaningful penalty clauses for non-compliance which could guarantee stronger binding power. Pricing is being discussed within a pre-determined range with a floor price that guarantees a high margin level on a sustainable basis."
Expect supply shortage to persist at least until 2028 and next year to be tighter than this year
"We now expect 2026E/2027E/2028E DRAM S/D to be in a larger undersupply of 5.0%/5.9%/3.9% (vs. prior expectations of 4.9%/2.5% undersupply in 2026E/2027E)"
"We now expect 2026E/2027E global DRAM demand to grow 28%/20% yoy (vs. +25%/+17% yoy previously), and introduce 2028E estimate of +19% yoy growth."
"We now expect 2026E/2027E/2028E NAND S/D to be in a larger undersupply of 4.4%/4.6%/3.0% (vs. prior expectations of 4.2%/2.1% undersupply in 2026E/2027E)"
"We now expect 2026E/2027E global NAND demand to grow 23%/20% yoy (vs. +22%/+15% yoy previously), and introduce our 2028E estimate of +19% yoy growth."
HBM: Expect a tighter 2027 S/D, and a significant ASP increase to lead to a +100%+ growth in TAM to US$116bn
We now expect 2026E/2027E/2028E HBM S/D to be in a larger undersupply of 5.4%/6.0%/4.3% (vs. prior expectations of 5.1%/4.0% undersupply in 2026E/2027E)
Significantly raise HBM ASP forecast as we expect a catch-up to conventional DRAM
Raise 2026E/2027E/2028E HBM TAM to US$56bn/US$116bn/US$168bn
Show more
GS on memory ( $MU / $SNDK ):
"The key demand driver of both the 2017-2018 upcycle and the current upcycle is strong server demand. While the 2017-2018 upcycle was driven by cloud data center expansion, the current cycle is driven by massive AI server build out...memory demand is now closely tied to server demand where server segment comprises around 50% of DRAM demand and ~40% of NAND demand."
"given the long lead time (takes at least ~3 years to construct a fab and produce output), we expect conventional DRAM capacity addition pace by 2030 to be slower than the 2017-2018 upcycle."
"Based on our checks, LTAs are being discussed to include sizable prepayments and meaningful penalty clauses for non-compliance which could guarantee stronger binding power. Pricing is being discussed within a pre-determined range with a floor price that guarantees a high margin level on a sustainable basis."
Expect supply shortage to persist at least until 2028 and next year to be tighter than this year
"We now expect 2026E/2027E/2028E DRAM S/D to be in a larger undersupply of 5.0%/5.9%/3.9% (vs. prior expectations of 4.9%/2.5% undersupply in 2026E/2027E)"
"We now expect 2026E/2027E global DRAM demand to grow 28%/20% yoy (vs. +25%/+17% yoy previously), and introduce 2028E estimate of +19% yoy growth."
"We now expect 2026E/2027E/2028E NAND S/D to be in a larger undersupply of 4.4%/4.6%/3.0% (vs. prior expectations of 4.2%/2.1% undersupply in 2026E/2027E)"
"We now expect 2026E/2027E global NAND demand to grow 23%/20% yoy (vs. +22%/+15% yoy previously), and introduce our 2028E estimate of +19% yoy growth."
HBM: Expect a tighter 2027 S/D, and a significant ASP increase to lead to a +100%+ growth in TAM to US$116bn
We now expect 2026E/2027E/2028E HBM S/D to be in a larger undersupply of 5.4%/6.0%/4.3% (vs. prior expectations of 5.1%/4.0% undersupply in 2026E/2027E)
Significantly raise HBM ASP forecast as we expect a catch-up to conventional DRAM
Raise 2026E/2027E/2028E HBM TAM to US$56bn/US$116bn/US$168bn
Show more
Mizuho on Agentic AI CPU/servers: see Agentic AI driving demand higher, but memory/CPU supply potentially limit upside in 2H26E.
"believe the demand for traditional servers could come from both x86 (+ve $INTC, $AMD) and $ARM-based CPUs with ARM gaining overall share into 2027E."
"Estimate total server shipments for 2026E up 17% y/y to 14.6M units (prior: ~14.0M units) with non-AI server growth at up ~6% y/y (prior flattish y/y). Estimate AI servers growing at a 26% 2025-29E CAGR to reach 5.87M units by 2029E (prior: 5.67M units)"
"Continued Blackwell/Rubin ramps providing near-term tailwinds for $CRDO"
"We see AMD with Turin and Genoa continuing to see strong demand ahead of Venice ramps in 2H26E, while INTC continues to see Granite/Diamond Rapids supply constrained ahead of Coral Rapids in 2027E, and ARM continuing to benefit from partner ramps, ahead of upcoming AGI CPU, where demand was noted up 2x vs. launch, and potential ASIC announcement"
Show more
UBS Desk:
"I can’t remember a
time when I have received so many inquiries from one our research reports than this $MU report did not to mention the level of controversy it
created (Tim actually received a few rogue threatening emails not cool…)"
Show more
UBS on $MU:
Our supply chain work on Long Term Agreements (LTAs) across the memory industry suggests that up to 30% of DDR volumes industry-wide will be soon locked in at pricing that is just slightly below current levels
we now expect the DRAM industry to remain undersupplied until at least C2Q28 (vs. 4Q27 prior), and NAND undersupply to last until 4Q27 (was 3Q27).
At this point, we believe hyperscalers alone to have currently secured ~60-70% of the industry Server DDR5 volumes under 'enhanced' LTAs - in other words, guaranteeing MU and others offtake for this volum
Even more important, this scenario yields MU EPS of >$100 even out in C2029E when we do believe DRAM pricing will undergo a fairly significant correction. In other words, we think MU still earns >$100 even if pricing for the floating portion of DDR declines by ~50%.
Show more
Mizuho on Agentic AI CPU/servers: see Agentic AI driving demand higher, but memory/CPU supply potentially limit upside in 2H26E.
"believe the demand for traditional servers could come from both x86 (+ve $INTC, $AMD) and $ARM-based CPUs with ARM gaining overall share into 2027E."
"Estimate total server shipments for 2026E up 17% y/y to 14.6M units (prior: ~14.0M units) with non-AI server growth at up ~6% y/y (prior flattish y/y). Estimate AI servers growing at a 26% 2025-29E CAGR to reach 5.87M units by 2029E (prior: 5.67M units)"
"Continued Blackwell/Rubin ramps providing near-term tailwinds for $CRDO"
"We see AMD with Turin and Genoa continuing to see strong demand ahead of Venice ramps in 2H26E, while INTC continues to see Granite/Diamond Rapids supply constrained ahead of Coral Rapids in 2027E, and ARM continuing to benefit from partner ramps, ahead of upcoming AGI CPU, where demand was noted up 2x vs. launch, and potential ASIC announcement"
Show more
Mizuho $MU NDR takeaways:
Agentic-AI driving CPU-DRAM, we est. $NVDA's $20B CPU revs could imply 3K PB incremental CPU demand at ~6% global DRAM supply
We believe some key customers remain 30-50% under-supplied and growing customer interest in L-T agreements
2027E with pricing potentially UP 70-100% y/y vs current avg. HBM projections of up 17%/up 36% for 2026/2027E
Show more
DRAM/NAND pricing pressure should pick-up the next 2-3 quarters
To date, $DELL supply chain has deftly managed the sharp increases in both DRAM and NAND and pressure across other components as well. However, the impact of the rising costs are expected to be more severe in the second half of calendar 2026 into the first calendar quarter of calendar 2027. While the rate of change moderates beyond C1Q:27, we do not expect a decline in prices further impacting PC, server, and storage margins for longer.
Show more
$DELL upside at Wells Fargo/Evercore:
Traditional Servers - Agentic Driven Upcycle: With $INTC & $AMD recently noting very strong server CPU demand, expect Dell to report another strong double-digit growth qtr. in trad. server demand.
Resilient demand and higher ASPs support upside to revenue and gross profit dollar estimates, even as memory-related cost pressure weighs on margins.
The key debate remains whether demand can hold up in H2 as additional price increases are passed through.
Show more
