LRCX

Avoid: at $230.16 Lam trades 41% above our $163.44 fair value. Lam makes the machines that etch and deposit the layers of a chip; its sales rise and fall with what chipmakers spend on such tools ("wafer-fab equipment", WFE). The September quarter was a record ($8.22B of revenue, a 40% operating margin), Lam guided December to $8.60B, and it gave a first view of 2027: more than $200B of industry tool spending, about 29% above 2026's mid-$150B, led by memory (DRAM). We take that view in full. It moves spending earlier but not higher by 2029, because the latest dated forecast for 2029 (SEMI's, eight days before the call) has not moved. That leaves fair value about where the earlier inputs put it. The price fell 11.8% today (from $260.81), and it still needs something close to the Street's forecasts, which put Lam's 2031 profit about half again above ours.

Why our value is $57 below the last report's $220.31: the method, not the quarter. The earlier report valued Lam's 2031 profit partly at what peers traded at on next year's Street profit (25.3 times operating profit). Our shared method no longer uses a current or forward market multiple. It uses the average of what Lam and its peers (Applied Materials, KLA, ASML) traded at at their fiscal year-ends over the past ten years, 17.5 times. Restated that way, and at a cost of equity of 10.75% instead of the 9.75% the earlier report used (it cut the rate by a point to sit nearer the Street; that cut has since been removed), the earlier report is worth $159.97: about $16 of the $60 is the discount rate and $44 the multiple. The quarter, the guide and Lam's 2027 view net to +$0.27 on top of the earlier inputs (section 3).

How to read the cases. The central case is our best estimate and alone sets the call. The pessimistic case is a severe but plausible bad outcome (tool spending +18% in 2027, then down 7% in 2028 and 22% in 2029 led by memory, with Lam losing share in China); the optimistic case is the matching good one (+34% in 2027, growth through 2031, and gross margin rising to 55.5%). Each case values Lam in mid-2031 as the average of a "long fade", which lets profit growth slow to 3% over 15 more years while new investment earns a return well above the cost of capital (45% in the central case, against a 10.6% cost of capital; Lam's own median is 54%), and an "outside view", the ten-year history multiple above.

1. What Lam sells and what limits it

Chipmakers buy Lam's tools to etch and deposit films. In fiscal 2026 (to June 2026) revenue was $23.23B. In the September quarter systems (new tools) were 70% of revenue and customer support (spares, service, upgrades and tools for mature chips) 30%; systems split by market as foundry 46%, memory 46% (NAND flash 26%, DRAM 20%) and logic 8% (10-Q and the CFO, call of 2026-10-21). What limits it: customers' cleanroom space and tool budgets; China (35% of revenue in September, 26% in June) turning to local toolmakers and to export rules; and memory makers cutting orders when memory prices turn. What drives value per share over a cycle is tool spending, Lam's share of it and its prices. Cost of goods fits $1.52B fixed plus about 43 cents per revenue dollar over fiscal 2024-26, and that fit gives 52.4% gross margin for the September quarter against 52.3% reported (output.txt). Lam's after-tax operating profit has run at a median 54% of operating capital over fiscal 2015-26.

2. The September quarter against our forecasts and Lam's guide

Both of our forecasts held their ranges; both came in below our point.

Source: Lam's release of 2026-10-21; graded in ../2026-10-16-lrcx/grades.json. Both points came from Lam's habit of beating its own guide. This quarter's beats were the smallest of the last nine: 1.5% on revenue (median 2.4%) and 0.3 point on gross margin (median 0.8). Lam gave a cause for a smaller December (supply limits, below) but none for September.

What else the quarter said (release and 10-Q, facts):

• Systems revenue rose 35% on the quarter to $5.73B, a record for both memory and foundry; customer support was flat at $2.49B.
• Operating margin 40.0%, EPS $2.19 (guide $2.15). The tax rate was 16.9%.
• Free cash flow $3.24B; $1.35B returned in buybacks and dividends; cash and restricted cash $7.45B against $3.75B of notes.

December guide and management's claims (call of 2026-10-21):

• Revenue $8.60B ± $0.4B, gross margin 52.5%, operating margin 40.5%, EPS $2.33. The guide "would have been a bit higher if not for some near-term supply chain constraints"; the CFO said that revenue moves to March.
• China's share "meaningfully down" in December, on domestic customers.
• Calendar 2026 tool spending in the "mid $150 billion range"; 2027 "more than $200 billion", led by DRAM, then leading-edge foundry and logic, then NAND; a number limited by cleanroom space and parts, with demand "higher"; customer visibility through 2028.
• Systems revenue up "well over 40%" in 2026; customer support "more than 30%" in 2026 and "over 15%" in 2027. Our December estimate gives +49% and +34%.
• Tax about 16.5% "for the December quarter and likely beyond".
• Gross margin "plus or minus the level that we are at" near term; mid-50s over several years remains a goal, which we quote and do not model.

3. Our forecast and value: the quarter changes little, the method a lot

The model is model.py with output.txt; the inputs and their sources are in assumptions.md. The central case:

• Tool spending (SEMI's industry measure) grows 29% in 2027, Lam's view [management claim], split DRAM +38%, foundry and logic +26%, NAND +24% in the order Lam gave. 2028 grows 14-15%, SEMI's July forecast [SEMI]. 2029 grows 3.5%, which keeps 2029 at the level SEMI's 300mm outlook gave on 2026-10-13 ($227B) [our inference]. That outlook covers 300mm fabs only and starts from $149B in 2026, not the $144B total our path starts from; we assume the two grow alike to 2029. SEMI calls for double-digit growth each year to 2029; our path front-loads it into 2027-28 and leaves 2029 small. 2030 is flat and 2031 grows 6-8% [our inference; no source covers them].
• Lam's share rises 2.5% a year in 2027-28 and 1% after; customer support grows 16% in 2027.
• Gross margin stays near 52.5-53% (no pricing gain after fiscal 2027); operating margin about 40%; tax 16.5%.
• Fiscal 2031 revenue $45.6B, operating margin 39.6%, after-tax operating profit $15.1B, trimmed to $14.2B for depreciation catching up with capex and for the end of Lam's Malaysia tax ruling in fiscal 2037.

From $220.31 to $163.44, in order (output.txt, "Bridge"):

The last two rows are one decision. Lam's 2027 view is new evidence that more of the build lands in 2027; it is not evidence that the build is bigger by 2029, and SEMI's 2029 level is the newest dated figure for that year. Adding Lam's raise on top (the old +9.6% step) would put 2029 at about $241B, 6% above SEMI's level, and would be worth +$8.10. If we are wrong here, we are wrong toward too low a value (section 6).

The value is mostly judgment. The rest of fiscal 2027's payout and fiscal 2028's owner cash flow are worth $12.96 a share today; net cash after operating cash is $0.10. The other $150, 92%, rests on years nobody can date.

The two halves of the exit disagree by 39% (fade $178, outside view $248). The fade discounts Lam's 2031 profit at its 10.6% cost of capital with growth fading from 6% to 3%, which is worth 14.3 times the next year's after-tax profit. Lam and its peers have traded at 17.5 times pre-tax operating profit over ten years, richer than that fade, which history can justify only with a lower discount rate or longer growth. The input behind the gap is the discount rate the fade uses against the one the market has used. We do not pick one: the method averages them, and the odds below draw the exit across the spread history shows.

What moves value, from the central (output.txt, "Sensitivities"):

4. Where we differ from the Street

The Street (Street average, 2026-10-22) is close to us for fiscal 2027 and then pulls away (market.py, market.txt):

The Street's path, run through our model step by step, closes most of the gap to the price:

So the price is about the Street's forecasts at our discount rate and our exit method. The difference is the forecast, not the valuation step: the Street has tool spending growing through 2031 with no pause and gross margin passing management's mid-50s. The Street's average target, $368, discounted one year at our cost of equity is $333; our value is 0.49 times that. The part of the targets above $241 is a multiple richer than ten years of history, which our method does not use.

5. What the price expects

No single driver reaches $230.16 even at the edge of its range (the odds command's reverse solve, odds.json): tool spending growth of 14% a year in 2028-31 gives $201; 40% growth in 2027 gives $176; a 5%-a-year share gain gives $186; 1 point a year of pricing gives $175. On our operating path the price needs an exit 1.51 times ours, $320 in 2031, which is 21.8 times fiscal-2031 operating profit against our 14.3 and the ten-year history's 17.5.

Odds. The simulation draws our seven drivers and one factor for where the exit multiple lands, in two worlds: the build holds (75%) or a memory-led downturn hits in 2029-30 (25%; Lam's revenue fell in four of the last seventeen years). The value is at or above the price in 16% of draws. About seven tenths of the spread comes from the exit factor, which is history's spread of multiples five years on, not our operating doubt. The simulated median ($143) sits 12.5% below the central because the downturn world pulls the left side down further than any draw pushes the right.

Alpha's view. Our quantitative model ranks Lam 139 of 702 (tier Strong, score 80.3, 2026-10-22): quality A, momentum B, growth B, expectations C, valuation D. It likes the business and the trend in estimates and rates the price as expensive, which is our view too: the call here is about price, not quality.

Technical read (beside the view, not an input to it; daily bars): D, in a downtrend. The stock is 47% below its 30 June high of $433.33, below its 50-day average ($306) and 200-day average ($283), and fell 11.8% today against 6.8% for the semiconductor fund SOXX.

6. The pessimistic case, and why the other side could be right

The pessimistic case ($71.02) has 2027 spending up only 18%, then down 7% in 2028 and 22% in 2029 led by memory, Lam losing share, gross margin slipping to 49-50% and a 30% return on new capital. Our odds put about one draw in twenty at or below it.

The case for owning Lam at $230.16 is that the build is longer than any dated forecast says. Two facts support it: Lam has raised its 2026 view three times this year (mid-$130s to mid-$150s), and it calls 2027 a constrained number with customer visibility through 2028. If 2029-31 grow at about 12% a year instead of our +3.5%, flat and +7%, the value is about $191; at the 14% top of our range, $201; with management's mid-50s gross margin on top, near $210. That world needs no new multiple, only a cycle with no pause before 2031, which Lam's record (four down years in seventeen) has not yet shown. The single assumption that would flip the call is our 2029-30: SEMI's December forecast or its next 300mm outlook lifting 2029 well above $227B, together with a margin path toward the mid-50s.

7. Next results preview: the December 2026 quarter, about 2027-01-27

Lam has not announced the date; 2027-01-27 is our estimate. Revenue, gross margin and EPS ranges run from the smallest to the largest beat of Lam's last nine quarters against its own guide: with nine past outcomes the next lands inside that span about 80% of the time ((n-1)/(n+1)), if the next guide error is drawn like the last nine. This is a change of method: the September ranges were the mean of eight beats with a 90% interval from their spread. They held; ranges built this way would not have held, because September's beats were the smallest of the nine (output.txt, "Guidance record").

What a beat is worth: each $100M of December systems revenue above ours, if it lifts the base rather than shifting timing, is worth about $0.77 a share (market.txt). Each point of 2027 tool-spending growth in Lam's January view is about $1.2.

8. Dates to watch, and what would change the view

• 2026-11-10, US export rules: the suspension of the Affiliates Rule ends; a lower China share in 2027 is the pessimistic case's share loss.
• December 2026, SEMI's year-end equipment forecast: 2027 near Lam's +29% confirms our central; under +20% points toward the pessimistic case.
• About 2027-01-27, December-quarter results and Lam's refined 2027 view: the six forecasts above.

What would move the view to Hold (fair value at least 0.85 times the price, $195.64, about +$32): 2028-31 spending growth near 12% a year (+$27) plus a pricing gain of 0.4 point a year (+$5); or 2029 above SEMI's $227B by 5% or more together with gross margin above 53.5% for two quarters with China under 30%. What would move it further from the price: a 2027 view under $190B, or China below 20% of revenue for two quarters.

9. What changed from the 2026-10-16 report, and why

• New evidence: the September quarter (reported), the December guide, Lam's 2027 spending view, the tax rate and customer-support growth (section 2).
• Forecast judgment: the December estimate is the guide at Lam's median beat, as before; the 2029 step moves so that 2029 stays at SEMI's level.
• Company-model structure: depreciation catch-up now corrects the fade only (the outside view's history multiples are on reported profit); the change in working capital sits in the bridge, as the method asks (counting it in reinvestment as before gives $162.66); valuation date 2026-10-22.
• Shared method: the outside view at the ten-year history multiple, each case's multiple scaled by its own growth after 2031, and odds that draw one exit factor (so the old outside-multiple driver is dropped) with every range centred on our value; and no cut to the cost of equity (9.75% to 10.75%). These are the $60 step: about $16 the rate, $44 the multiple.
• After the price was read: the low end of the pricing range moved from a 0.40 to a 0.45 point slip in gross margin a year. The odds command requires each range's median to sit near its middle, and -0.40 failed that check; the move is to pass it, not on evidence. Its effect is small: that end of the range is worth $158.15 instead of $158.74.
• The price: $230.16 against $263.94 at the last report, down 12.8%; our value fell 25.8% from $220.31, though only 2% from the $159.97 restatement of the earlier report's inputs under today's method. The call is Avoid both times.

Sources

• Lam Research 8-K, 2026-10-21, Exhibit 99.1 (September quarter results and December guide), in data/edgar/. Primary.
• Lam Research 10-Q for the quarter to 2026-09-27, in data/edgar/ (read masked). Primary.
• Lam Research earnings call, 2026-10-21 (Quartr transcript via market data), data/calls/lrcx-q1fy27-call-2026-10-21.txt. Secondhand (a vendor's transcript of Lam's call).
• Lam Research releases, September 2024 to June 2026 quarters (guidance record), in data/edgar/. Primary.
• SEMI, 2026-07-14 equipment forecast and 2026-10-13 300mm Fab Outlook, as cited in the 2026-10-16 report's assumptions. Third party.
• Market data statements, peer betas (2026-10-18), fiscal-year-end multiples (data/exit-anchor.json, 2026-10-22), analyst estimates and targets (2026-10-22). Secondhand.
• Market quote and daily bars, 2026-10-22, data/market-2026-10-22.json, data/history-2025-10-01-2026-10-22.json.
• Alpha model rank, 2026-10-22 (read-only).