Lesson 24 · Chips & LLMs · Capstone arc (2 of 3)

Valuation Frameworks

The investor half, made operational. Every technical insight you've built — moats, bottlenecks, the inference cost stack — now has to show up in one of three places: the TAM, the margin, or the multiple. This lesson is how your engineering taste becomes a price.

Builds on: the whole course + THESIS.md Skill: value a name without anchoring to consensus

You can now read these companies better than most analysts at the technical level. This lesson converts that into the thing the mission is about: a defensible, independent valuation view. The trick is that valuation is not a separate skill bolted on — every technical fact you've learned maps directly to one of three valuation inputs. A moat shows up as margin and as a durable multiple; a bottleneck shows up as pricing power (margin) and share of TAM; cyclicality shows up as a multiple discount. Learn the map and you can price a name from first principles.

Core thesis: Value = TAM × share × margin × multiple. The first three are the business; the fourth is the market's judgment of growth × durability × cyclicality. Your technical edge is knowing which input each company's moat actually moves — and spotting the single most dangerous error in semis: valuing peak-cycle earnings at a peak-cycle multiple (double-counting the boom). Get the map right and the per-name signals tell you when the thesis is working.

01 — The Skeleton

Four Inputs, One Price

Market value ≈ TAM × market share × net margin × P/E multiple TAM × share = revenue · × margin = earnings · × multiple = value. Decompose any name into these four and you can see where your view differs from consensus.

Why decompose rather than quote a P/E? Because each input fails differently, and your technical knowledge gives you an edge on a specific one. You might agree on TAM but believe a moat protects margin longer than consensus (NVDA gross margin); or agree on margin but think the multiple is too low because the cut is socket-agnostic (ASML, EDA); or think the market is valuing peak share as permanent. Naming your disagreement is the whole game.

02 — Input 1 & 2: TAM and Share

How Big, and How Much Do They Keep?

Two ways to size a TAM: top-down (analyst market forecasts) and bottom-up (your favorite — build it from physical reality). You already have the bottom-up engine: hyperscaler capex (~$750B in 2026) flows into known buckets (L20's dollar trace). Anchor numbers, all grounded:

MarketNowForecastRead-through
AI accelerators (GPU+ASIC)>$1T/yr by 2030NVDA + AMD + AVGO-designed ASICs fight over this
Data-center GPUs~$100B (2024)~$215B (2030)The merchant-GPU slice specifically
Data-center semis (all)$209B (2024)~$500B (2030)Compute + memory + networking + power silicon
HBMtight; +130% units '25~$100B TAM by 2028SK Hynix ~50% share → the cleanest memory bet

Sources: accelerator TAM, DC semi TAM, HBM.

Share is where your moat knowledge pays. A monopoly (ASML EUV) keeps ~100% of its TAM; a duopoly (SK Hynix/Samsung/Micron in HBM) shares it; a contested socket (NVDA vs AMD vs ASICs) has share that can move — which is exactly what the THESIS.md falsifiers track. The durability of share is a technical question you can now answer.

03 — Input 3: Margin = the Moat on the Income Statement

Gross Margin Is the Moat, Quantified

Gross margin is the single cleanest financial fingerprint of a moat: it's what pricing power looks like in numbers. Map the moat types you learned to the margins they produce:

Moat type (from the course)Gross marginExample
Software / IP toll booth~85–90%Cadence/Synopsys (EDA), NVDA software value
Monopoly hardware~50%+ASML (EUV) — high for capital equipment
Full-stack systems leader~71–75%NVIDIA (data-center)
Process monopoly (capital-heavy)~64–66%TSMC
Cyclical commodity (yield-gated)~30–60% (swings!)SK Hynix / memory

NVDA ~71–75%, TSMC ~64–66% per recent quarters. [TSMC/NVDA margins]

Diligence use: a change in gross margin is the earliest quantitative signal that a moat is shifting. NVDA gross margin slipping toward the low-60s would be your first hard evidence that ASIC/AMD competition (the L16 falsifier) is biting — visible in the financials before it's in the narrative. Watch the trend, not the level.

04 — Input 4: The Multiple, and the Trap

What Justifies a Premium — and the Double-Count Error

The multiple (P/E, EV/EBITDA) encodes three things: growth (faster → higher), durability (more certain → higher), and cyclicality (more cyclical → lower, because the market discounts earnings it doesn't trust to persist). This is why a socket-agnostic toll booth deserves a higher multiple than a memory maker even at similar growth:

The single most dangerous mistake in semis — the double-count: valuing peak-cycle earnings at a peak-cycle multiple. In a boom, a memory maker's earnings spike and its multiple expands, so the price looks "cheap on P/E" exactly at the top. The fix: normalize to mid-cycle earnings, and apply a multiple that reflects the through-cycle business, not the peak. The cyclical-double-count is how investors who understood the technology still lost money — they bought peak earnings as if permanent.

TAM → Value Calculator (with the cyclical-trap check)

Build a value bottom-up, then drag the cycle position to see the double-count danger. Illustrative, not a price target.

Revenue
$150B
TAM × share
Value (as-quoted cycle)
$1.6T
earnings × P/E
Mid-cycle (sober) value
$1.6T
normalized earnings × P/E
05 — The Per-Name Map

Which Input Each Name's Thesis Lives In

NameTierGross marginThe ONE signal that moves it
ASMLtoll~50%Net new bookings (EUV + High-NA); China revenue mix
Cadence/Synopsystoll~85–90%Recurring-revenue % + backlog/renewals; design starts
TSMCtoll~65%Capex guide + advanced-node utilization + CoWoS capacity
NVIDIAleader~71–75%Data-center revenue trajectory and gross-margin trend (erosion = competition)
Broadcomleader~60%+Custom-ASIC/AI revenue + new XPU customer wins
AMDleader~50%Data-center GPU revenue ramp; ROCm/rack-scale wins (the L16 catalyst)
SK Hynixcyclical~30–60%HBM ASP + bit shipments; the broader DRAM cycle (double-count risk highest here)

Notice the pattern: the toll-booth tier (ASML, EDA, TSMC) is valued on durability — their signals are bookings/backlog/utilization, and their risk is a capex air-pocket, not displacement. The leader tier (NVDA, AVGO, AMD) is valued on share + margin durability — their signals are revenue ramp and margin trend, and their risk is competition (the falsifiers). The cyclical tier (SK Hynix) is valued on where we are in the cycle — its signal is ASP, and its risk is buying the peak. Three different games requiring three different valuation lenses.

06 — Putting It Together

From Technical Edge to a Price View

Your edge is input-specific
You rarely beat the market on all four inputs. Pick the one your technical knowledge actually informs — e.g. "I think NVDA's gross margin is more durable than the multiple implies because rack-scale topology isn't copyable yet" — and size the bet to that claim. A thesis is a specific disagreement on a specific input.
Normalize before you judge "cheap"
Always ask: is this earnings number mid-cycle or peak? A cyclical name on a low P/E at peak earnings is a value trap; a toll booth on a high P/E with durable earnings can be cheap. The P/E alone is meaningless without knowing where in the cycle the "E" sits.
Margin trend > margin level
The earliest evidence a thesis is breaking shows up as a margin inflection, often a quarter or two before the narrative catches up. This is where your technical read (is the moat eroding?) becomes a tradable, falsifiable signal — the bridge from THESIS.md to action.
The multiple is a falsifier in disguise
"This deserves a premium multiple because the cut is socket-agnostic" is a falsifiable claim: if a credible second source appears (ASML EUV rival, open-source EDA at advanced nodes), the durability premium should compress. Tie every multiple judgment to the relevant THESIS.md falsifier.
Primary Source

Go Deeper

Read first: Stratechery (in RESOURCES.md) for how technical/business structure becomes durable value — the qualitative side of the multiple. Pair with: SemiAnalysis for TAM/margin teardowns, and primary 10-K/earnings releases (e.g. NVIDIA's filings) to read gross-margin and segment trends yourself rather than via headlines.

Comprehension Check

Quiz — 6 Questions

Select the best answer for each.

1. In the framework Value = TAM × share × margin × multiple, a moat shows up most directly as:

A larger total addressable market for the whole industry
Higher margin and a more durable (higher) multiple
A lower share that is easier to defend over time
A permanently fixed price-to-earnings ratio for the sector

2. The "cyclical double-count" error in semiconductor valuation is:

Counting both training and inference revenue in the TAM
Applying a peak multiple to peak-cycle earnings at the top
Using gross margin and operating margin in the same model
Adding a company's share and its competitor's share together

3. Why does a socket-agnostic toll booth (ASML, EDA) deserve a higher multiple than memory?

Its earnings are more durable and less cyclical to forecast
It always grows revenue faster than every other semi name
It has lower gross margins that must be compensated for
Regulators require a premium valuation for monopolies

4. The earliest quantitative sign that NVIDIA's competitive moat is eroding would be:

A one-quarter dip in total data-center unit shipments
A downward trend in its gross margin over several quarters
An increase in the company's research-and-development spend
A rise in the overall data-center semiconductor TAM

5. A cyclical memory name trading on a low P/E at peak earnings is most likely:

Genuinely cheap, since a low P/E always signals value
A value trap, because the "E" is at an unsustainable peak
Fairly valued, as cyclicality never affects fair multiples
Overvalued only if its gross margin is also at a trough

6. A well-formed valuation thesis is best described as:

A general belief that a strong company will keep winning
A specific disagreement with consensus on one named input
A forecast of next quarter's exact earnings-per-share figure
A judgment that the stock's chart looks ready to move up
From your instructor: The frame to keep — Value = TAM × share × margin × multiple; a moat is higher margin + a durable multiple; a bottleneck is pricing power + share; cyclicality is a multiple discount; and the deadly error is valuing peak earnings at a peak multiple. Your technical edge attaches to one input per name — name it, and you have a thesis instead of a vibe. Ask me anything: how to normalize mid-cycle earnings, why EV/EBITDA can beat P/E for capital-heavy names, or how to read a gross-margin bridge in an earnings release.