←The mapResearch bank
Theses
Solid-state transformer: unrelieved constraint carrying semiconductor revenue as rent, not as scalePackage Perimeter as Hyperscaler Negotiating Leverage Against NVIDIACoWoS Gatekeeper Paradox: TSMC's Margin Reflects Crowding-Out, Not ConversionNVLink's Demand Is a Fragile Coalition of Neoclouds Under Collateral PressureWolfspeed: High-Voltage Moat Evaporates Under Thermal ScrutinySK Hynix: The Memory Supplier Suffering Its Own Customer's ConstraintAccelink: The Valuation Anomaly in China's Optical Chip IndigenizationCoWoS Binds the Buyer, Not the LandlordHBM5: The Transmission-Line Bottleneck Transfers Margin to the Memory SuppliersArm: The Revenue-Mix Pivot Consensus Hasn't PricedInnolight: The Hyperscaler Pricing Ratchet Disguised as GrowthLong-Context Inference: Infrastructure Debt Becomes the LeverageInterconnection queue converts datacenter operator debt into utility equity premiumAmazon: The Capex-Margin InversionAixtron: Customer Capex Discipline Masks Downstream GrowthShin-Etsu: Hidden Exposure to Datacenter Grid FailuresTransformer scarcity is pricing power for industrial conglomerates, not a hyperscaler cost problemMarvell's custom-ASIC rent accrues upstream, not in the fabless P&LKLA: the annuity thesis the equipment sell-side ignoresLG Innotek: Substrate Ambition Subsidised by the Wrong CustomerElectrical Steel Asymmetry: GE Vernova Captures Transformer Scarcity Without Steel ExposureHBM4: equipment leverage swamps memory maker concentrationABF substrate dual-gate arbitrage: substrate converters capture the spread the market assigns upstreamNAND Flash: Consumer Demand Collapse Disguises Enterprise Pricing PowerTesla: AI Capex Subsidy Disguised as Automotive TurnaroundGas Turbine Bottleneck Inverts the Natural Gas ExposureUALink: Consortium Activity Masks Structural IrrelevanceCo-packaged optics is a packaging play being misfiled as an optics transitionApplied Optoelectronics: Loss-Funded Market Access Captures Transceiver Supply ScarcityMicrosoft: Free Cash Flow Collapse Telegraphs Margin Compression Before It Hits the P&LRack Power Density: The Liquid Cooling Adoption MirageGPU collateral decay transmits to NVIDIA demand before physical constraints clearxAI: Vertical Integration Theater Masks Structural GPU DependencyAdvanced Packaging: The Scarcity Premium Flows Upstream, Not DownSilicon Wafer: Duopoly Ships into Oligopoly Margin Explosion, Captures NoneEaton: Margin expansion telegraphs demand scarcity more than growth can showMoE inference: stranded-power miners own the scarcest input, hyperscalers rent itApplied Materials: The Hidden Margin Trap in a Structural UpswingMurata: The Margin the Bottleneck Hasn't ReachedAlphabet: The Capex Ratchet Liquidity Mispriced as Platform OptionalityLam's service annuity masks exposure to memory's margin conversionEUV scarcity is priced into ASML, invisible in AlphabetConstellation: Hyperscaler Monopsony Masks Nuclear Fuel TransmissionOracle: Monopsony Rent Capture Masked by Consolidated AccountingFoxconn: Customer Concentration Absorbs AI Margin Before It Hits the P&LThe Conventional DRAM Squeeze: HBM Conversion Creates a Consumer Margin Crisis Through 2028Palantir: The Margin-Protected Infrastructure PlayDISCO: Memory oligopoly capex collapses the monopolistCXMT: The Supplier-Margin Windfall Hiding Inside the Subsidy StorySMIC: Subsidy Converts to Capacity Under Obscured Margin PressureQualcomm's low capex masks an IP-only future neither consensus nor bears have pricedTransceiver margin expansion is a mirage: buyer concentration at 1.6T resets pricing power the market prices as durableTokyo Electron: Memory Rent Disguised as Equipment MarginSilicon capacitors expose Intel's package productivity deficitArista: Hyperscaler capex intensity conceals margin compression riskGlobalFoundries: Customer Concentration Masks Structural Insulation from AI Capex Whiplash
Mechanisms
ABF substrate and build-up film supplyCoWoS advanced-packaging capacityConventional DRAM and NAND supply (HBM crowding-out)EUV tool capacity — the lowest rungGPU residual value as loan collateralGrid interconnection queue positionHeavy-duty gas turbine delivery slotsLarge power transformer lead timesSamsung memory long-term agreementsBuyer concentration tighteningHBF consortiumRent converting into capacityRent migrating upstreamRent not being competed awayUALink ConsortiumUltra Ethernet Consortiumco-packaged-optics displaces copperco-packaged-optics displaces optical-transceiveremib displaces cowosglass-substrate displaces abf-substratehybrid-bonding displaces euvsilicon-capacitor displaces mlccChip designDatacenter mathEfficiency arrives in steps, not trendsInference shapeMemory economicsPhotonicsTau scalingToken mathTokenomicsCapacity arriving — CoWoS (Chip-on-Wafer-on-Substrate)Capacity arriving — HBM4Capacity arriving — Silicon Wafer

The Conventional DRAM Squeeze: HBM Conversion Creates a Consumer Margin Crisis Through 2028

gen-dram · conviction computed 0.63 · status open · horizon — · as of 2026-08-19

Conventional DRAM shortage is structural, not cyclical, because HBM and conventional DRAM share the same wafer base and HBM conversion is subtractive. SK Hynix M15X and Samsung P4 remove 130 kwpm of conventional capacity in 2026-Q4 while demand compounds at 2.3% quarterly. The only material addition (120 kwpm, speculative, 2027-Q4) arrives after the deficit opens.
Robust to undisclosed shares. 1 derived input under this thesis; redrawing every supply weight the industry does not publish moves none of them by more than 25%. Computed from evidence at most 29 days old (oldest input: micron).
How to read the numbers on this page

A range instead of a point. 980 of 1,929 supply weights are not disclosed by anyone. Where redrawing them across their plausible range moves a figure by more than 25%, the figure is shown as a range and marked. A tight number is ground you can stand on; a wide one is not.

Not every premise is scored. A premise citing something the model verifies on every rebuild — a filed figure, a graph edge, a computed cell — is a PRECONDITION, not a risk. It gates the conclusion but contributes no uncertainty, because charging a conclusion for being verifiable made well-evidenced arguments look weaker than vague ones.

Composed two ways. Where several premises gate a conclusion, the figure is given as "X% if independent, Y% if they move together". They are claims about one industry, so the truth is between and nobody can say where. Treat it as an ordering device, not a calibrated probability.

Dated. Each figure is computed from facts, and the page states the age of the oldest one beneath it. The full arithmetic runs from a published model config to company revenue exposure.

Exhibits

Exhibit 1Relative performance, indexed to 100How the names in this thesis have traded against SOXX.
116041196index100 = startMU 764000660.KS 635005930.KS 388SOXX 213AAPL 13712mo, indexed to 100 at start · dashed = SOXX benchmark

Series available as data/gen-dram.csv

Exhibit 2Who pays Conventional DRAM and NAND supply (HBM crowding-out), and who keeps the moneyCapturers average 58.6% operating margin against payers' 32.4% — the owners of the scarce thing capture the rent, as expected.
SK Hynix68.0%Micron Technology65.6%NVIDIA Corporation64.0%Samsung Electronics36.9%Apple Inc.32.4%

Green/blue = model marks it as CAPTURING the rent (unbound and supplies the scarce good); faded = PAYING it (bound severe or moderate). Operating margin, live.

The variant

Consensus

Markets price memory suppliers on HBM exposure and datacenter AI demand. Apple's gross margin compression is read as iPhone mix shift or services deceleration. The conventional DRAM tightness visible in mobile contract prices (+90% QoQ early 2026) is treated as a transient supply hiccup that capacity additions will relieve by late 2027.

Variant

Conventional DRAM shortage is structural, not cyclical, because HBM and conventional DRAM share the same wafer base and HBM conversion is subtractive. SK Hynix M15X and Samsung P4 remove 130 kwpm of conventional capacity in 2026-Q4 while demand compounds at 2.3% quarterly. The only material addition (120 kwpm, speculative, 2027-Q4) arrives after the deficit opens. Consumer OEMs face sustained BOM inflation through 2028, but only Apple has balance-sheet depth and operating margin (32.4%) to absorb multi-year memory cost escalation without pricing power or margin collapse.

Differentiator

Supply-chain analysts model memory capacity as aggregate DRAM and miss the zero-sum reallocation between HBM and conventional within a fixed wafer base. The constraint binds consumer electronics, not AI, because hyperscalers can pay; Apple is long scarce input costs and short the OEMs that cannot.

Open questions

Falsifiers

unstructured Conventional DRAM capacity deficit persists through 2027
Industry conventional DRAM wafer starts (300mm-equiv) exceed 1,650 kwpm in any quarter before 2028-Q1, disclosed in supplier earnings or third-party fab utilization data
settles refuted by 2027-12-31 · no machine-readable clauses yet
unstructured Apple sustains gross margin above 30% under memory cost inflation
Apple reports product gross margin below 30% in any quarter through 2027-Q4
settles refuted by 2027-12-31 · no machine-readable clauses yet
unstructured Consumer OEM consolidation accelerates under memory cost pressure
Top-5 smartphone OEM share (ex-Apple) increases by at least 5 percentage points YoY in any quarter through 2027-Q4, per IDC/Counterpoint shipment data
settles confirmed by 2027-12-31 · no machine-readable clauses yet

Reasoning chain

Conventional DRAM capacity contracts through 2027 as HBM conversion subtracts faster than greenfield adds VALID
premises
  • Conventional DRAM and NAND supply (HBM crowding-out)0.88 strong
    Constraint exists and mechanism is documented; uncertainty is conversion pace vs demand growth
  • SK Hynix0.92 strong
    M15X timeline is firm and capacity quantum is disclosed
  • Samsung Electronics0.78 strong
    P4 is planned-grade; execution and allocation between HBM/conventional carry doubt
  • SK Hynix — operating margin 68.0%1.00 strong
    Filed figure; no uncertainty

Composed 0.88 via and over 1 gating premise · 3 supporting premises shown but not multiplied in — citing a filed figure should not cost conviction · Weakest link Conventional DRAM and NAND supply (HBM crowding-out) at 0.88

130 kwpm subtraction in 2026-Q4 is higher-confidence than the speculative 120 kwpm add in 2027-Q4, and demand grows 9.5%/yr compounded. The gap widens before it narrows.

Apple sustains gross margin under memory cost inflation where smaller OEMs cannot VALID
premises
  • Apple Inc.0.85 strong
    Pricing power and services mix can offset BOM rise, but consumer demand elasticity is the gate
  • Apple Inc. — operating margin 32.4%1.00 strong
    Filed margin; absorptive capacity is factual
  • Consumer Electronics Demand0.72 moderate
    10%
    Smaller OEMs face same cost shock with sub-10% margins and no pricing power; doubt is timing of exit/consolidation
  • Apple Inc. — market cap usd $4.81tn1.00 strong
    Balance sheet depth is observable; relevant to multi-year input cost locking

Composed 0.61 via and over 2 gating premises · 2 supporting premises shown but not multiplied in — citing a filed figure should not cost conviction · Weakest link Consumer Electronics Demand at 0.72

32.4% operating margin gives Apple 20+ points of cushion vs OEMs at sub-12%. Memory rising to >30% of BOM is survivable for Apple, fatal for marginal players, so share consolidates to the one name that can pay.

Memory suppliers convert record margins into HBM capacity rather than conventional relief, extending the shortage VALID
premises
  • Conventional DRAM (DDR5 / LPDDR) — input-cost pressure at least 68.17111.00 strong
    Computed from supplier margins; tight band, aged 29 days
  • SK Hynix0.80 strong
    Disclosed HBM roadmap and pricing power in both segments; doubt is political/export-control interference
  • Micron Technology0.80 strong
    65.6%
    Same position as SK Hynix; 65.6% operating margin with HBM ramp ahead
  • Micron Technology — operating margin 65.6%1.00 strong
    Filed operating margin

Composed 0.80 via and over 1 gating premise · 3 supporting premises shown but not multiplied in — citing a filed figure should not cost conviction · Weakest link SK Hynix at 0.80

68% gross margin (z=0.91 vs peers) funds capacity, but every announced project through 2027 targets HBM because datacenter willingness-to-pay exceeds consumer. Conventional tightness is the intended outcome, not a bug.

Sources