What Does Lam Research's Price Assume About the China-Licence Stuck Step?
Lam Research closed at $263.45, about $329.7 billion. It already shipped $4.250 billion of etch tools. What does the price assume about getting paid on the etch while two China NAND fabs wait on a US licence?
ANALYSIS · MEMBERS CONTINUE · 5 min free / 15 min full · Data as at 21 Oct 2026
Lam Research closed at $263.45 on 20 October, about $329.7 billion for the shares. It already shipped $4.250 billion of etch and deposition systems in the quarter to 28 June. This morning's Wire: Samsung is converting Xian — about 40% of its NAND — toward 286-layer chips, and SK Hynix wants 240 layers or more at Dalian. What nobody has disclosed is how far Washington has licensed the American etch tools those conversions need. What is the price already assuming about getting paid on the etch while two China fabs still wait on a signature?
In short
- The question: what is Lam Research's price already assuming about the China-licence stuck step — chambers recognised in Taiwan, Korea and on the fleet, while Xian and Dalian wait on an annual US tool licence?
- The answer: Scenario B. About $329.7 billion is 45.7 times trailing earnings and 27.8 times forward earnings (ours). The price is paying for etch to keep shipping, and for service on more than 40,000 chambers, not for a Commerce upgrade stamp
- Why: June printed $6.722 billion of revenue and $4.250 billion of systems; NAND 23% of systems, China 26% of total. September guided $8.10 billion ± $400 million. Reuters: Xian 5,000 → 40,000 V9 wafers a month by December 2027; Dalian 192 → 240+ layers. BIS still points to 2025: operate existing fabs, not upgrade technology
- The size of the gap: September's midpoint is 20.5% above June (ours). Eight times the December V9 start-rate is a conversion, not a new fab. That conversion is not on the $8.10 billion guide
- What would break it: a September print well below ~$8.10 billion, or NAND mix rolling from 23% (A); or a dated US upgrade-without-capacity statement plus a NAND beat that looks like 40,000 wafers (C)
- Nothing here is a recommendation. This is a map of what the price requires, with the working shown
The numbers that matter
| What | Figure | Date | What it tells us |
|---|---|---|---|
| Share price / equity value | $263.45 close; about $329.7bn | 20 Oct 2026 | The price we test |
| Trailing earnings / sales | NI $7.27bn; sales $23.23bn; EPS $5.76 | StockAnalysis, same close | About 45.7× / 14.2× (ours) |
| Forward multiple | Forward PE 27.77 | Same | Implies about $9.49 of EPS, +65% vs $5.76 (ours) |
| June revenue / systems | $6.722bn / $4.250bn (63.2%) | 8-K, 29 Jul 2026 | Who is paid on the chamber |
| June CSBG | $2.472bn (36.8%) | Same | Service, spares, upgrades, Reliant |
| NAND / DRAM / foundry mix | 23% / 23% / 44% of systems | Company slides, 29 Jul 2026 | About $0.98bn / $0.98bn / $1.87bn (ours) |
| China / Taiwan / Korea | 26% / 27% / 20% of total | Same 8-K | About $1.75bn / $1.82bn / $1.34bn (ours) |
| September guide | Revenue $8.10bn ± $400m; EPS $2.15 ± $0.15 | Outlook, 29 Jul 2026 | Not a print. Call 21 Oct, 2:00pm PDT |
| Xian V9 plan | 5,000 wpm Dec 2026 → 40,000 Dec 2027; ~40% of Samsung NAND | Reuters, 21 Oct 2026 | The conversion that needs US etch |
| Dalian plan | 192 → 240+ layers; Solidigm's only NAND fab | Same | Same licence clock |
| TSMC Q3 / Q4 guide | US$46.72bn; US$52.3–54.3bn | 15 Oct 2026 | Taiwan still invoicing |
Sources: Lam 8-K exhibit 99.1, 29 Jul 2026; June 2026 slides; StockAnalysis, LRCX, 20 Oct 2026; Reuters via CNBC, 21 Oct 2026. Mix dollars and multiples are ours.
Why does Lam sit at the China-licence stuck step?
Lam sells etch, deposition and clean tools. It gets paid when a system is recognised, then again when the installed fleet is serviced. It does not sell the EUV scanner, and it does not issue the US export licence. What Does Lam Research Do? is yesterday's free Stack read of that model. This morning's Wire asked what binds at Xian and Dalian. The scarce object there is a signature in Washington. Today is the company next to that factory.
Demand outside those two fabs looks clear. Slides put NAND at 23% of June systems mix, doubled sequentially on 256-layer and above. DRAM held 23%. Foundry was 44%. This morning's Taiwan Pulse: TSMC printed US$46.72 billion of third-quarter sales and guided US$52.3–54.3 billion for the fourth. Lam's own September output is guided $8.10 billion ± $400 million, 20.5% above June (ours) — a ship plan, not a 40,000-wafer Xian conversion. Cash and restricted cash $5.598 billion less debt $3.735 billion is about $1.86 billion of net cash (ours). The print is due on this afternoon's Pacific call, after this desk.
What is left is the licence. Extra layers at Xian and Dalian still need American etch boxes, and those boxes now need an annual US licence. Reuters named etching machines as the equipment that needs that licence, and named Lam, Applied Materials and KLA when the 2025 waivers were pulled. Commerce, asked this morning, pointed back to its 2025 line: license the existing fabs, not technology upgrades.
This is not Applied's scanner-slot question. On 15 October we asked what Applied's $354 billion already needed from chambers that ship around a scarce ASML tool. Lam sits on the carve seat of the same layer, but today's stuck step is a yearly stamp for two China NAND fabs. Do not read $4.250 billion as a shipped Xian conversion.
What does ~$330 billion already assume?
Three answers compete:
- Chambers slip with the licence. A "no upgrades" year freezes Xian and Dalian a generation behind. NAND mix rolls from 23%. September prints well below ~$8.10 billion. The forward earnings step does not arrive
- Chambers keep shipping around a scarce licence. NAND mix holds in the low-20s because Korea and the US still convert. China 26% is multinational customers on nodes they are already allowed to run, not a 286-layer unlock. Taiwan 27% is TSMC still invoicing. Lam is still paid on the etch
- The licence eases. A dated US statement allows technology upgrades without new capacity. The 5,000-to-40,000 V9 ramp becomes a 2027 systems print. NAND mix steps up
A tip sheet would pick a side. Analysis asks which answer today's price already needs.
Our read, in one line: the price already assumes Scenario B. About $329.7 billion is 45.7 times trailing earnings (ours). The forward multiple of 27.77 times needs about $9.49 of EPS, 65% above $5.76 (ours). That is a claim on chambers that still ship — not on BIS adding an upgrade sentence. The licence is what binds Xian. The invoice is what binds the price. The scenarios and the arithmetic are below.
The free preview ends here.
MEMBERS CONTINUE HERE — who gets paid · three scenarios · takeaways
