DRAM — expert X mentions

Mizuho Securities: CPUs & GPUs Market Forecasts & Growth > Shipment Growth: Industry server CPU shipments are forecasted to reach 35 million units in 2026 and grow to 50 million units by 2027, representing a 40% year-over-year increase. > Long-Term TAM: The long-term Total Addressable Market (TAM) estimate for 2030 has been raised to $170 billion (up from the previous $107 billion forecast), driven by higher CPU-to-GPU ratio assumptions for AI inference servers. > CPU-to-GPU Ratios: The CPU-to-GPU ratio on AI servers is accelerating and is expected to approach 1:1 by the end of 2027 or 2028. Supply Chain & Technical Bottlenecks > DRAM Constraints: A critical bottleneck exists in DDR5/LPDDR5 supply, with a projected fulfillment ratio of only 70% over the next 12–18 months. > Demand vs. Supply Gap: Based on current models, the 2027 demand for DDR5/LPDDR5X (over 300 billion 1Gb equivalents) significantly exceeds the projected supply (220–250 billion 1Gb equivalents). > Potential Risks: The shortage of key materials—DRAM, substrates, and passives—is expected to persist through 2027 and could pose downside risks to downstream server assemblers, potentially leading to lower server rack output. Key Player Insights (2027 Forecasts) > Nvidia: Expected to reach 5.0–6.0 million units for the Vera CPU, including 2.0–3.0 million units specifically for agentic AI stack racks. > Google: Axion CPU production is projected to increase more than 2x year-over-year, aligning with the growth trajectory of TPU units. > AMD: The N2 Venice CPU is forecasted to exceed 6.0 million units. GPUs/ASICs Market Growth Projections > Rapid Expansion: The total AI ASIC market is projected to grow from 4.1 million units in 2025 to 24.0 million units by 2028. > Volume Drivers: The growth is driven by substantial increases in deployment by major hyperscalers including Google, Amazon (Annapurna), Meta, Microsoft, and OpenAI. > External Demand: The market for external (non-Google) AI ASIC units is expected to surge from 0.6 million in 2025 to 7.2 million by 2028. Key Hyperscaler Activity > Google (TPU): Continues to be a dominant player, with total shipment units increasing from 2.5 million in 2025 to 7.1 million by 2028. > Anthropic: Significant ramp-up is forecasted for their "TPU Ironwood/Sunfish" chips, moving from 0.6 million units in 2026 to 6.2 million units by 2028. > Amazon/Annapurna: Shipments for the Trainium line are projected to double from 1.5 million in 2025 to 3.6 million by 2028. > Meta: Rapid scaling of MTIA chips is expected, growing from 0.1 million units in 2025 to 2.7 million units by 2028. Technical Trends > Advanced Packaging & Nodes: There is a heavy reliance on sophisticated packaging technologies like CoWoS-L and CoWoS-S, and advanced foundry nodes including N2, N3, N4, N5, and A16. > HBM Integration: Nearly all listed high-performance ASICs utilize High Bandwidth Memory (HBM), with a transition toward newer generations such as HBM3E and HBM4/4E to meet performance demands. > ASP Variance: Average Selling Prices (ASP) range significantly, from approximately $2,000 for entry-level models to as high as $40,000 for top-tier specialized chips like the TPUv10. $DRAM $EWY $MU $GOOGL $AMKR $TSM $ASE $NVDA $AMD $AVGO $MRVL $INTC $MSFT $META

"Even if the memory shortage persists throughout 2027, research firms, sell-side analysts, and other prominent industry observers broadly agree that the supply-demand imbalance will begin to ease in 2028." Yes, that is correct. I actually have an interview transcript (I cannot disclose due to sensitivity) where the senior employee of one semiconductor company in Taiwan states: "The supply and demand for DRAM will reach a balance in H2 2028 to stabilise the price." He explains it along the lines of: 2026: The current demand gap is hovering around 4% to 5%. Original 2025 projections of an 8% to 9% shortage proved exaggerated because major manufacturers like Samsung and SK Hynix successfully transitioned from 1b to 1c process technologies. This node upgrade boosted bit output and filled much of the anticipated shortfall. 2027: The supply gap is projected to shrink to under 3% by the second half of the year as Micron and Samsung ramp up entirely new production facilities. 2028: The market will face a minor 2% to 3% deficit in the first half of the year (H1) before finally achieving a complete supply-demand equilibrium in the second half (H2), driven by the massive volume of new capacity deployed between late 2027 and 2028. $MU $EWY $DRAM

$DRAM all time high daily volume by a mile today. Nice deep red to green close. Should be a start of a decent DCB next week.

$META in talks to rent up to $10B in AI compute to Anthropic over two years. GPU+HBM clusters are becoming enterprise revenue engines. $MU $SKHY $DRAM. Healthy Economics Incoming. Claude is the enterprise standard for complex, high value problems. Most expensive frontier model per task. Highest KV cache footprint. Highest HBM demand per token. Enterprise doesn't mean casual usage. It means contracted, always on workloads running around the clock. Every dollar Meta charges Anthropic for enterprise compute flows directly through GPUs and HBM.

Morgan Stanley: ZAM Memory > Massive Cloud Capex Exposure: Cloud memory spending is projected to reach US$418 billion by 2030, representing an 8% CAGR from 2026. Memory spending could account for 40% of total cloud capex in 2027 (up from just 12% in 2023, prior to the AI boom). > Bandwidth Lags Behind Processing: Memory bandwidth improvements (e.g., a +14% increase from DDR5-5600 in 2024 to DDR5-6400 in 2026) are failing to keep pace with explosive token growth, which expanded over 320x between April 2024 and June 2026 on major cloud platforms. > Absolute Spend Peaking: Absolute cloud memory spending is forecasted to hit its cycle high in 2028e at roughly US$520 billion. > The Post-2028 Softening: From 2028e to 2030e, even though overall cloud capex (excluding memory) is projected to rise to its highest level of $1.28\text{ trillion}$ by 2030, memory spending falls as a percentage of the total budget back down to 25% by 2030e. DRAM Supply/Demand (S/D) Imbalance > Historical Context: The severe supply shortage of 2023 is clearly visible, bottoming out in 4Q23 near -28%. > The Current 2026/2027 Tightness: Following a brief macro recovery, the market enters a sustained period of deep supply deficit. Through all of 2026e, the supply deficit hovers continuously between -10% and -16%. > Supply Shortage to Persist: Looking ahead, the green bars indicate that the market remains structurally undersupplied, keeping DRAM conditions historically tight all the way through 4Q28e. NAND Supply/Demand Dynamics > The COVID/Post-COVID Glut: The massive NAND oversupply of late 2022 is prominent, peaking in 4Q22 with a +17% surplus. > Structural Repricing Phase: After hitting a deep shortage in 1Q26e (dipping near -9%), the NAND deficit begins to gradually moderate, though it remains marginally in negative territory (undersupplied) near -3% to -4% through late 2026e. > Transition to Enterprise Storage: This sustained tightness supports the report's thesis that NAND is moving from cheap consumer sockets up the memory hierarchy to act as massive LLM KV caches. $DRAM $MU $SNDK $WDC $EWY

BofA: 3Q Memory Key Takeaways from the 3Q Contract Price Check > Strong Server DRAM Pricing: Up 20–30% QoQ, spearheaded by high-speed LPDDR5 (compared to the market consensus of 20% or less). > Upbeat Spot Demand: Spot demand is healthy, driven by commodity DDR5 and legacy DDR4, with July prices increasing Month-on-Month (MoM). > Transition to HBM4: There is a near-term rise in orders shifting toward higher-priced HBM4 over the cheaper HBM3e. > Lower Long-Term Agreements (LTAs): LTA-based sales represent well below 50% of total DRAM sales (non-LTA portion is at 60–70%), allowing for solid QoQ price increases (e.g., 5–10%) that are often settled even under existing LTAs. > Rush Orders: Sudden OEM rush orders confirm a QoQ price increase of over 20% for both commodity DRAM and NAND. > Optimistic ASP Outlook: BofA forecasts a 21% QoQ rise in 3Q DRAM ASP, which is significantly more optimistic than TrendForce's assumptions (13–18% QoQ for conventional DRAM, or only 8–13% including HBM). Spot Market Dynamics > Price Rallies: DRAM spot prices have risen for eight consecutive weeks, defying earlier expectations of a price cap at US$35–40 for 16Gb DDR5. > OEM Buying Behavior: Having previously cut device production to resist the spot-price rally, OEMs are now actively purchasing memory chips to prepare for high sales in September and the peak 4Q season. > NAND Recovery: NAND spot prices rebounded, driven by a 4% Week-on-Week (WoW) rise in 1Tb wafers, supporting expectations of a 10%+ NAND ASP rise in 3Q. $DRAM $MU $SNDK $WDC $EWY

Appetite for the memory trade is still super high: Record inflows for the South Korea ETF $EWY this week. Tuesday: $814 million (previous record high) Wednesday: $1.1 billion (new record high) Investors are back to using $EWY as a proxy for the South Korea SK Hynix listing, just like in Q1. Rather than paying the ~25% premiums for the US listed SK Hynix ADR. And for added context on the demand for memory stocks: The memory ETF $DRAM is currently at record-high AUM levels at ~$25 billion, despite only being released three months ago.

$MU $SKHY $DRAM "DRAM export value in early July increased 437% compared to the same period last year." "Analyst Ryu Hyung-geun projects that 2027 HBM ASP will rise more than 100% year-over-year. The base scenario is a 100% increase, and the upside could be even larger depending on negotiations with customers." "The reason is not limited to HBM supply-demand alone. As general-purpose DRAM becomes more profitable, the opportunity cost for memory companies when allocating the same capacity to HBM also rises. [...] Customers may accept higher prices to secure stable HBM volumes." "SK Securities analyst Han Dong-hee stated that if long-term supply contracts expand and price increase rates moderate, subsequent performance growth must be driven by increased sales volume."

$MU $SKHY $DRAM New private (non bank) memory research just landed. July reports. Some of these reports cost five figures annually. I will be sharing my commentaries on these reports on Patreon. A bargain if you're in the memory complex. One data point jumped out immediately. "DRAM capex is expected to reach a record high in 2026, currently projected at near $65bn, driven not only by HBM and node migration, but increasingly by higher construction spending for new fabs and cleanroom expansion." "Most of the current capex surge is unlikely to contribute materially to production until late 2027 or 2028, leaving near-term supply growth still dependent on migration and selective line optimization." "Record-high DRAM capex increasingly reflects construction-led investment, but most new capacity is unlikely to contribute meaningfully until late 2027 or 2028." https://t.co/LYqoeOF7k1

There are 3 avenues I explored for hyperscalers in the coming months and how memory costs may impact their earnings. Hardware is a small part of their overall business, but has the potential to create the same impact as Samsung's mobile business did in their preliminary earnings. Check it out 👇 $GOOGL $AMZN $MSFT $NVDA $MU $DRAM $EWY

Morgan Stanley: HBM & NAND Supply/Demand HBM (High Bandwidth Memory) Market & Supply > Market Size Expansion: The total HBM market value is projected to explode from US$3 billion in 2023 to US$94 billion by 2027e, representing a massive 128% CAGR. > NVIDIA Market Share: NVIDIA is expected to maintain overwhelming dominance in the GPGPU market, keeping its market share consistently at 90% or above through 2027e (peaking at 92% in 2026e). > Soaring HBM Intensity: The average HBM content per GPGPU is modeled to increase from 80 GB in 2023 to 317 GB in 2027e. Average HBM content per ASIC is also scaling up, from 40 GB to 238 GB over the same period. HBM Supply Shares (2027e Implied Production): SK Hynix: 22,226 million Gb Samsung: 21,239 million Gb Micron: 10,372 million Gb > Widening Deficits: While HBM-specific supply appears to outpace standalone HBM demand, the broader DRAM market faces a severe squeeze. The Total DRAM sufficiency ratio (combining both HBM and commodity DRAM) is projected to plunge to -17% in 2026e and remain deeply negative at -15% in 2027e. AI NAND Market & Supply > Surging AI NAND Demand: Demand for AI NAND is modeled to nearly triple, rising from 205 EB (Exabytes) in 2025 to 609 EB in 2027e. > AI as a Percentage of Total NAND: AI's share of total global NAND demand is expected to climb sharply from 18% in 2025 to 41% by 2027e. Key Hardware Drivers (2027e): ASIC eSSD Usage: Major demand is driven by Google TPU (33 EB) and AWS Trainium (19 EB). GPGPU eSSD Usage: Driven heavily by NVIDIA platforms (63 EB extra deployment per tray + 34 EB in-rack eSSD) and AMD's MI-series (37 EB). > NAND Sufficiency Deficit: After a slight supply surplus of +2% in 2025, the total NAND market is projected to flip into a severe deficit of -15% in 2026e and remain under-supplied at -9% in 2027e. > Non-AI Demand Remains Flat: Meanwhile, non-AI NAND demand (PCs, smartphones, general enterprise SSDs) is projected to remain relatively flat, moving from 905 EB in 2025 to 874 EB in 2027e, showing that AI is entirely driving the market's tight supply dynamics. $DRAM $MU $EWY $SNDK

$MU "Our view is consistent with management, that AI will keep DRAM demand well above supply BEYOND 2027" -Morgan Stanley Holy Structural $DRAM Shit

Morgan Stanley: Semiconductors Memory Market Debates & Pricing Cycles > AI Spend (Constructive): Driven by hyperscale capex spend, the intensity of top LLMs, funding, and Annual Revenue Run Rate (ARR). The team notes that 2Q26 capex is the true "tell" and that monetizing infrastructure does not equal excess compute. > LTA Re-rating (Neutral): There is no structural re-rating on Long-Term Agreements (LTAs). The market is pricing memory earnings rationally rather than euphorically, drawing parallels to the analog case study during COVID where LTAs were renegotiated or inventory was forced. > Memory Cycle (Neutral Inflection): The cycle is seeing a peaking year-over-year (YoY) rate of change, but it is expected to elongate rather than collapse. DRAM contract pricing is forecasted to peak in 4Q26. > Inventory Trends: DRAM and NAND inventory picked up in 2Q26, a shift primarily driven by module houses rather than suppliers or OEMs. NAND Supply vs. Demand Scenario (YMTC Focus) > Capacity Expansion: China's YMTC is building Fab4 and Fab5 (~100kwpm capacity each). Dedicating all five announced fabs to NAND could theoretically elevate YMTC to a 24% global NAND market share. > Outlook to 2028: Assuming non-AI NAND demand grows at +5% YoY and AI SSD demand expands by +30–60% YoY, the NAND market is projected to stay tight into 2028—provided YMTC remains disciplined. Faster greenfield expansion remains the primary oversupply risk. $MU $DRAM $EWY

$TSM $MU $SKHY $DRAM TSMC's SoW-X: 64 HBM stacks on a single wafer. Today's best? 12. Today (CoWoS): 8 to 12 HBM stacks 2028 (CoWoS): 20 HBM stacks 2029 (CoWoS): 24 HBM stacks 2029 (SoW-X): 64 HBM stacks The leap is staggering. https://t.co/TsRqiHjq1z

KOSPI opened (South Korea index) at 8pm EST… started slightly green, then went red and now back to flat. I do think SK Hynix and/or SK Square are very compelling buys after a -40% pullback. SK Hynix (SK) / $SKHY is now trading below 4.5x NTM EPS and below 4.0x 2027 EPS NFA. DYOR. We own SK Hynix through SK Square and $DRAM.

According to this report, memory supply coming in 2028 is greatly overestimated. "Recent supply chain reports suggest that SK Hynix’s actual new memory production capacity in 2028 may be only one-sixth of the amount initially planned." $DRAM $EWY

J.P. Morgan: Memory > Memory at the Center of AI Debate: Memory has become a focal point for investors tracking Cloud Service Provider (CSP) hardware capital expenditure (capex). > Surging Capex Share: Memory's share of CSP capex is projected to skyrocket to 52% in 2026 and exceed 70% next year, compared to less than 20% in the pre-AI era. > Investor Skepticism: While investors agree on tight supply-and-demand fundamentals, they want to see concrete AI service breakthroughs and revenue acceleration to justify memory's massive value share. > Rapid Forecast Revisions: Memory value within 2026 CSP capex forecasts jumped from 20% to over 50% in the past year, driven by stronger-than-expected DRAM and NAND price momentum in the March quarter. > Lagging Consensus: Analysts revise semiconductor forecasts faster than CSP hyperscaler forecasts, meaning it may take a few months for the consensus gap between the two sectors to close. $MU $DRAM $EWY

Inside Memory's New Playbook: SCAs, Equity Stakes, and the End of the 'Boom-Bust' Cycle @mzuhair123, one of the best researchers, and I did a collaborative article We talked about these aspects: - The Current Industry State: Micron Earnings, SK Hynix ADR & Frontier Lab Collaboration - Hyperscaler & Rising Memory Costs: A Huge Risk To the Infrastructure Buildout - Qualcomm’s HBC: High Bandwidth Compute, But It’s Just Another Stacked DRAM Approach - It’s Not All “Rainbows & Sunshine”: The Risks Worth Flagging With the DRAM Industry $MU $DRAM $EWY Read on substack or X 👇

$MU $SKHY $DRAM Just Remember HBM pricing is expected to nearly TRIPLE by Q4 2027. Not stabilize. Not hold. Triple. Volatility is driven by fear. The thesis is driven by data. https://t.co/WvfQMowb3o

$MU $SKHY $DRAM Just Remember 86% operating margins projected to hold through 2027. Volatility is psychologically driven. The investment thesis is data driven. https://t.co/1c4Gj9WuSz