Macrontology
Research bank
Theses
Power Caps the AI Buildout — the bottleneck moved from chips to megawattsThe ABF Chokepoint Does Not Pay Its Owners — Buy the Bottleneck Fails HereASML / China-DUV displacement is overpriced — the sell-off prices near-full substitution off a 5-machine data pointThe China chokepoint moves from lithography to bonding — and the substitution trade is priced in the wrong categoryCost Per Token Is Set Outside the ChipCustom Silicon Pays a Different MerchantEMIB Shifts the Packaging Bottleneck to SubstratesHBM moat vs. DRAM commoditization — is SK Hynix's HBM-mix 'miss' actually bullish?HBM4 Qualification Concentrates Share, Not Spreads ItSilicon cannot emit light — indium phosphide is the chokepoint CPO relocates but cannot removeThe Neocloud Rent Is Consumed By The Asset Before It Reaches EquityNVIDIA's Rent Compresses Through Software, Not SiliconOptical attach is set by topology, not by shipments — linear optical TAM models are mis-specifiedThe Megawatt Is the Unit of AccountThe Upstream Is Single-Sourced and UnpricedChipflation Is a Wafer Allocation, Not a Demand ShockMemory-demand durability — does algorithmic efficiency cap the memory super-cycle?Neocloud margin re-rating — do open-source deployment + rising rental rates break the thin-margin cap?Semicap service is an installed-base annuity, and the market still prices these names as capex cyclicalsABF substrate dual-gate arbitrage: substrate converters capture the spread the market assigns upstreamAccelink: The Valuation Anomaly in China's Optical Chip IndigenizationAdvanced Packaging: Equipment Suppliers Capture the Scarcity PremiumAlphabet: The Margin Inversion Nobody PricesAmazon: The Capex-Margin InversionApplied Optoelectronics: Hyperscaler Procurement Arbitrage Driving Negative SpreadsArista: Hyperscaler capex intensity conceals margin compression riskArm: The Royalty Stream That Scales With Everyone's Margin but Its OwnCo-packaged optics transfers margin to substrate integrators, not optics specialistsConstellation: Hyperscaler Monopsony Masks Nuclear Fuel TransmissionCoWoS Constraint Prices Into the Wrong CompaniesEaton: Margin expansion telegraphs demand scarcity more than growth can showElectrical Steel Asymmetry: GE Vernova Captures Transformer Scarcity Without Steel ExposureFoxconn: Margin Squeeze Masked by AI Server Revenue MixGlobalFoundries: Capex Intensity Signals Strategic Exclusion from AI, Not WeaknessGPU collateral decay gates buildout faster than physical supplyHBM5: The Signal Integrity Tax Inverts the Memory Supplier Power BalanceTransformer lead times gate hyperscaler returns, not start datesInnolight: Negative Operating Leverage Hiding in Hyperscaler ConcentrationInterconnection queue converts hyperscaler capex into generator rentLong-context inference: memory constraint compresses cloud margin before GPU-poor plays feel itMicrosoft: Burning Cash to Rent Margin It Cannot OwnMoE inference: converted miners capture memory scarcity, hyperscalers leak itMurata: The Margin That Doesn't MoveNAND Flash: Datacenter Crowding-Out Misread as Memory Cycle RecoveryTransceiver margin expansion is a mirage: buyer concentration at 1.6T resets pricing power the market prices as durableOracle: Monopsony Rent Capture Masked by Consolidated AccountingPackage Perimeter Arbitrage: Optics Suppliers Capture Value NVIDIA Cannot RetainPalantir: The Margin-Protected Infrastructure PlayQualcomm's low capex masks an IP-only future neither consensus nor bears have pricedRack Power Density: The Liquid Cooling Adoption MirageShin-Etsu: The Hidden Rent in the DenominatorSilicon capacitor adoption arbitrages package economics invisible to component analystsSilicon Wafer Duopoly: Record Downstream Margins Signal Pricing Power Collapse, Not TightnessSK Hynix: The Customer Concentration That Validates the MoatSMIC: Bottleneck Arbitrage Disguised as National ChampionTesla: Capex Surge Hides Margin Compression at Suppliers

Arm: The Royalty Stream That Scales With Everyone's Margin but Its Own

gen-arm · conviction — · status open · horizon — · as of 2026-08-10

Arm is converting its best cycle into capacity at the fastest rate in its history—capex intensity now 11% versus a five-year median of 2.6%, ranking 100th percentile against itself. The royalty model that makes margin look structural actually ties revenue to the volume and ASP decisions of thirteen oligopsonists whose own capex intensity just fell 21% in four days.
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 21 days old (oldest input: alphabet).
How to read the numbers on this page

A range instead of a point. 903 of 1,848 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.
55193331index100 = startSOXX 221ARM 190GOOGL 172AMZN 123NVDA 12012mo, indexed to 100 at start · dashed = SOXX benchmark

Series available as data/gen-arm.csv

Exhibit 2Who pays CoWoS advanced-packaging capacity, and who keeps the moneyCapturers average 47.1% operating margin against payers' 47.0% — the owners of the scarce thing capture the rent, as expected.
Taiwan Semiconductor Manufac56.1%Analog Devices, Inc.38.1%SK Hynix68.0%NVIDIA Corporation64.0%Broadcom Inc.44.2%Advanced Micro Devices11.8%

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

Arm is the AI enabler everyone pays, a toll-booth on the hyperscale buildout capturing 26% growth as datacenter and edge AI proliferate. The 124× forward multiple prices in datacenter licensing wins and the transition from smartphone royalties to higher-value cloud and automotive silicon. Gross margin near 100% and single-digit capex intensity make this the cleanest levered play on semiconductor unit growth.

Variant

Arm is converting its best cycle into capacity at the fastest rate in its history—capex intensity now 11% versus a five-year median of 2.6%, ranking 100th percentile against itself. The royalty model that makes margin look structural actually ties revenue to the volume and ASP decisions of thirteen oligopsonists whose own capex intensity just fell 21% in four days. Arm's operating margin sits at 18.5% while customers like NVIDIA, Broadcom, and Qualcomm run 64%, 44%, and 26%; the spread says Arm is not capturing the end-market rent—it is subsidizing its customers' IP cost to defend the architecture against RISC-V.

Differentiator

Earnings-focused analysts see gross margin and miss that Arm is spending into R&D and capex at unprecedented rates to keep the architecture relevant. Supply-chain models see diversified exposure across thirteen customers and miss that the HHI of 879 is a full standard deviation below peers, meaning no single win moves the number—and that customer capex pull, which drives licensing upsells, just deteriorated 21% common-mode. The consensus reads the royalty as a margin story; the BOM says it is a volume and mix story, and both are decelerating structurally.

Falsifiers

Open questions

Reasoning chain

Arm is converting current operating leverage into future capacity faster than at any point in the past five years, implying margin expansion is being deferred to defend architectural ubiquity rather than harvested. VALID
premises
  • Arm Holdings plc — capex (TTM) $545m1.00 strong
    Disclosed TTM capex, unchanged for twenty days.
  • Arm Holdings plc — revenue ttm usd $4.92bn1.00 strong
    Disclosed TTM revenue, unchanged for twenty days.
  • Arm Holdings plc0.88 strong
    3%11%
    Capex-to-revenue is proxy for capacity investment; licensing IP typically runs <3%, so 11% signals either foundry partnership or defensive core development at scale.

Composed 0.88 via and over 1 gating premise · 2 supporting premises shown but not multiplied in — citing a filed figure should not cost conviction · Weakest link Arm Holdings plc at 0.88

100th-percentile capex intensity versus own history, during a 26% growth year, means Arm is reinvesting the windfall rather than dropping it to EBIT—which contradicts the toll-booth thesis and implies competitive pressure the multiple ignores.

Customer capex pull—the intensity with which Arm's thirteen downstream customers are themselves investing—fell 21% in four days and now sits below the peer median, indicating the licensing upsell cycle that drives Arm's growth is structurally decelerating. VALID
premises
  • Arm Holdings plc — capex pull at least 20.87791.00 strong
    Computed metric with zero spread, aged twenty days.
  • Alphabet Inc.0.82 strong
    Arm customer; any change in Google's capex affects weighted pull given edge supply relationship.
  • Amazon.com, Inc.0.82 strong
    40%
    Graviton licensee; AWS capex cycle is 40% of cloud spend and directly gates Arm datacenter royalties.
  • Ai Capex Cycle0.71 moderate
    21%
    Moderate upstream exposure; 21% four-day drop and common-mode flag imply cycle is rolling, not accelerating—falsifiable if hyperscale capex guides up next quarter.

Composed 0.71 via and over 1 gating premise · 3 supporting premises shown but not multiplied in — citing a filed figure should not cost conviction · Weakest link Ai Capex Cycle at 0.71

Capex pull below peer median plus deteriorating trend plus common-mode flag means the tailwind licensing revenue depends on is fading industry-wide. If customers slow silicon investment, Arm's per-chip royalty compounds on a smaller base and licensing NRE dries up.

Arm's 18.5% operating margin versus 44–64% at Broadcom, NVIDIA, and Qualcomm indicates it is not monetizing architectural control; it is subsidizing IP cost to prevent RISC-V displacement, which makes the current margin cyclical rather than structural. VALID
premises
  • Arm Holdings plc — operating margin 18.5%1.00 strong
    GAAP disclosed, depressed by R&D and stock comp but still far below customers who license the same IP.
  • NVIDIA Corporation — operating margin 64.0%1.00 strong
    64%
    NVIDIA Grace uses Arm; 64% operating margin shows rent accruing to integrator, not architecture owner.
  • Broadcom Inc. — operating margin 44.2%1.00 strong
    25.7 pp
    Broadcom custom AI chips use Arm cores; margin spread of 25.7pp says licensing cost is single-digit percent of COGS.
  • QUALCOMM Incorporated0.79 strong
    26%
    Qualcomm's 26% margin on smartphone and automotive SoCs, both Arm-based, implies royalty is low-single-digit revenue share—or pricing power would show up in Arm's margin, not Qualcomm's.

Composed 0.79 via and over 1 gating premise · 3 supporting premises shown but not multiplied in — citing a filed figure should not cost conviction · Weakest link QUALCOMM Incorporated at 0.79

The architectural moat is real, but rent capture is weak. Customers earn 2–3× Arm's margin selling Arm-based silicon, meaning royalty rates are set by RISC-V's zero-cost threat rather than by switching cost. Current margin is competitive equilibrium, not monopoly pricing.

Sources

Write-up

Pre-filled skeleton: gen-arm.md