DISCO (6146) Stock Valuation: A Great Business Priced Above the Bull Case, Reduce at ¥49,540

DISCO earned a 47.4% operating margin in the September quarter, a record, and its shares hit a 2026 low the next day. Both are true because management guided the December quarter's operating income to ¥58.1 billion from ¥67.6 billion, a fall of 14%. At ¥49,540 the shares pay 32 times a run-rate that is already below the quarter just reported.

At ¥49,540 on October 30, 2026, DISCO trades 45.6% above my weighted Base of ¥34,025, which is 31.3% below the price. The Bear value of ¥21,078 is 57.5% below the price and the Bull value of ¥49,112 is 0.9% below it, so even the Bull case does not reach the price. I assign no probabilities to the cases. The verdict is Reduce.

Valuation route: a semiconductor equipment maker with a leading position in precision cutting and grinding (a 71.7% gross margin is the evidence), cyclical in demand and unusually profitable, valued on five-year net income with an exit multiple and cross-checked on through-cycle earnings. Values are present fair-value estimates, not 12-month price targets.

What DISCO is

DISCO makes the machines that cut silicon wafers into chips (dicing saws), grind them thin (grinders) and polish them, and the blades, wheels and other consumables that wear out in those machines. Its customers are chip makers and the outsourced assemblers that package chips. Management says the annual mix is roughly half high bandwidth memory and half logic, and that memory was just under 40% of shipments in the September quarter.

The current fiscal period is guided as April to December 2026. All growth rates below compare with the same period a year earlier unless stated.

The latest results

September quarter 2026 (¥ billion)	Result	Comment
Net sales	142.8	up 36.4% from a year earlier; up 24.9% from the June quarter
Shipments	148.6	a record; planned at 141
Gross margin	71.7%	a record, about 1.5 points above the company's own assumption
Operating income	67.6	47.4% margin; 42.9% in the June quarter
Net income	46.1	32.3% of sales
First half net income	80.4	earnings per share ¥740.84 against ¥515.70

December quarter guidance (¥ billion)	Guidance	Comment
Net sales	132.4	down from 142.8
Operating income	58.1	margin 43.9%
Net income	41.5	31.3% of sales
Shipments	149.6	above the September quarter, mainly outside memory
Gross margin	just under 2 points lower	about 1 point from currency, just under 1 point from other factors

Shipments are higher in the guidance but sales and profit are lower, because sales trail shipments and the gross margin is expected to fall. The company guided the nine months to ¥389.5 billion of sales, ¥174.7 billion of operating income and ¥121.9 billion of net income (¥1,123 a share), and the December quarter ends the guided period; I found no outlook beyond it. It assumes ¥151 to the dollar.

Management said high bandwidth memory shipments will pause in the December quarter after a high September quarter, and called that quarter-to-quarter variability and not a change in demand. It also said it could not clearly explain its forecast for the quarter after that. Both statements are fair, and neither is something to value as a trend.

What the cash flow is made of

Cash conversion is lower than the margin suggests. First-half operating cash flow was ¥111.7 billion on net income of ¥80.4 billion, but investing outflow was ¥54.8 billion, so free cash flow was ¥56.9 billion, 71% of net income.
Spending is rising. Capital spending for the period is about ¥33 billion against depreciation of about ¥16 billion, including a new plant in Hiroshima (about ¥33 billion over three years) and a research center in Haneda (about ¥14 billion).
Inventories are large. ¥160.8 billion of inventories is about 1.1 times the September quarter's net sales, a number to watch if orders slow.
Hiring is accelerating. About 100 more support staff will be added by March 2027, for a total of about 200 in that group.
Dividends are formulaic. The policy is 25% of half-year net income, so the ¥186 interim dividend rises and falls with profit.

Debt, shares and financing

Item (September 30, 2026)	Value
Cash and deposits	¥342.4 billion
Total assets	¥824.6 billion
Total liabilities	¥195.6 billion
Equity	¥629.0 billion, an equity ratio of 76.1%
Net cash used in this analysis	about ¥340 billion (borrowings not separately read)
Shares	about 108.5 million (net income divided by earnings per share)

The balance sheet is a strength: equity is three-quarters of assets and cash covers all liabilities. I count the cash once, at about ¥340 billion, or ¥3,100 a share, and do not add cash that will build in the future, which is conservative. I did not read the borrowing line, so net cash could be slightly lower.

Primary valuation: five-year net income value

I project net income to 2031, apply an exit multiple to the 2031 figure, add dividends along the way at the company's 25% payout and add today's net cash, discounted at 9%. The Base value is ¥35,046.

Base case net income (¥ billion)	2027	2028	2029	2030	2031
Net income	166	180	192	202	212
Earnings per share (¥)	1,530	1,659	1,770	1,862	1,954

2027: ¥166 billion, which is the December quarter guidance of ¥41.5 billion multiplied by four. It is below the September quarter's ¥46.1 billion, and it is not a forecast of a downturn.
2028 to 2031: growth of 8%, 7%, 5% and 5%. Packaging for artificial intelligence chips keeps demand high, and I assume the equipment cycle does not repeat 2026's pace.
Exit multiple: 24 times 2031 net income, for a business with a 47% operating margin, net cash and 5% growth. That is a premium to the market because the franchise is better than the market's.
Terminal share: the exit value is 86% of the Base value over a five-year horizon, so the multiple matters more than the path, and I show the sensitivity.

Primary value per share by 2031 net income (rows) and exit multiple (columns):

2031 net income	18 times	21 times	24 times	27 times	30 times
¥180 billion	¥24,075	¥27,266	¥30,458	¥33,649	¥36,840
¥212 billion	¥27,529	¥31,287	¥35,046	¥38,804	¥42,563
¥250 billion	¥31,630	¥36,062	¥40,495	¥44,927	¥49,359
¥290 billion	¥35,947	¥41,088	¥46,230	¥51,371	¥56,513

The price of ¥49,540 needs about ¥313 billion of net income in 2031 at 24 times, or an exit multiple of 36 on my Base ¥212 billion. Both are possible. Neither is a base rate.

Secondary valuation: through-cycle earnings

The equipment industry has downturns. I take net income of ¥135 billion as through-cycle in the Base, 82% of the December quarter's run-rate, and apply 24 times, then add net cash. That gives ¥33,005. The Bear uses ¥105 billion at 18 times and the Bull ¥165 billion at 28 times.

Growth diagnostic

Sales rose 32% in the first half and net income 44%, so profit grew faster than sales as the gross margin reached a record. This is structural at the level of the franchise and cyclical at the level of the year: the company expects consumables, the part that does not depend on new machine orders, to be roughly flat excluding currency in the December quarter. I classify the growth as strong but cyclical, with the margin close to a peak.

Valuation triangulation

Primary (five-year net income value): ¥35,046, weight 50%.
Secondary (through-cycle earnings): ¥33,005, weight 50%.
Divergence: 6% of the midpoint, below the 20% warning level.
Weighted Base: ¥34,025.

The methods agree because both assume 24 times earnings and net income in the low ¥130 billion to ¥210 billion range. Their agreement is not independent evidence, since both rest on my multiple. The gap to the price is the main finding, and it does not depend on which method is used.

What is the market already pricing in?

Reverse valuation	Result
2031 net income that equates the price at 24 times	about ¥313 billion (my Base ¥212 billion; 2027 run-rate ¥166 billion)
Exit multiple that equates the price on my Base path	36 times (my Base 24 times)
Earnings multiple on my 2027 run-rate	32.4 times
Return if the Base is right	-31.3% from the price

The market is paying for net income to rise almost 90% from the current run-rate by 2031 at a premium multiple, which is what an AI packaging build-out could deliver, but it also needs the cycle to be absent.

Bear, base and bull scenarios

Case	Five-year net income value	Through-cycle value	Blended	Versus price
Bear	¥21,597	¥20,559	¥21,078	-57.5%
Base	¥35,046	¥33,005	¥34,025	-31.3%
Bull	¥52,497	¥45,727	¥49,112	-0.9%

Bear: net income of ¥150 billion in 2027, ¥130 billion in 2028 as memory and AI packaging spending pauses, then recovery to ¥160 billion by 2031 at 18 times. Through-cycle ¥105 billion at 18 times.
Base: as described above.
Bull: net income of ¥175 billion in 2027 rising to ¥285 billion in 2031 on AI packaging and new products, at 28 times. Through-cycle ¥165 billion at 28 times.

The discount rate is 9% in all cases. The spread is wide, from a 57% fall to a Bull case that is still slightly below the price, because the multiple at the top and the cycle at the bottom both matter.

How the buy price is set

Confidence is C, which has a default margin of safety of 42.5%. The two methods differ by less than 20% so I add nothing.

Method	Value	Buy price at a 42.5% margin
Five-year net income value	¥35,046	¥20,151
Through-cycle earnings	¥33,005	¥18,978
Weighted Base	¥34,025	¥19,564

The entry range is ¥18,978 to ¥20,151 and the conservative limit is ¥18,978. The price is 146% above the top of the range. The range states what I can defend as a margin of safety, not an expectation that the shares will fall there.

Dividend and buybacks

The interim dividend is ¥186 a share, up from ¥129, and the year-end dividend has not been decided. The policy of 25% of half-year net income gives a yield of about 0.8% on an annualized basis at the current price. I found no buyback in what I read.

Reasons to own DISCO

A 71.7% gross margin and 47.4% operating margin are exceptional for equipment makers, and they reflect a position that competitors have not displaced.
Net cash of about ¥340 billion and an equity ratio of 76% mean no financing risk.
Shipments were a record ¥148.6 billion, ahead of plan, and the December quarter guidance for shipments is higher still.
Demand is broad: logic, power chips, China and new fields such as co-packaged optics are all expected to rise in the December quarter.
Consumables, which are tied to chips already being made, are expected to hold roughly flat excluding currency in the December quarter, a floor under the cycle.

What could go wrong?

The multiple. At about 32 times run-rate earnings, a re-rating to 24 times with unchanged profit is a 25% fall.
The cycle. Management says it cannot clearly explain its forecast for the quarter after December, and high bandwidth memory is about half of the mix.
Gross margin. It is expected to fall just under 2 points in the December quarter, partly from currency.
China. China demand is described as strong and centered on outsourced assemblers; export rules could change that.
Currency. The yen assumption of ¥151 to the dollar matters for about 1 point of gross margin.
Capital spending. Spending of about ¥33 billion against ¥16 billion of depreciation raises the cost of a downturn.

Management execution

Guidance was beaten: September shipments were ¥148.6 billion against a plan of ¥141 billion and the gross margin came in 1.5 points above assumption.
The company now publishes a nine-month forecast, which is new, and it is cautious: December operating income is guided below the September quarter.
Management disclosed that it cannot clearly explain its forecast for the following quarter, which is an honest statement and a sign of limited visibility.
The dividend policy ties payments to profit, and the company keeps most of the profit as cash.

Stock Analyza scorecard

Business quality	Score
Moat and pricing power	20/20
Economic return over cost of capital	20/20
Balance sheet and resilience	15/15
Earnings and cash quality	11/15
Growth and runway	11/15
Management and capital allocation	8/10
Dilution and governance	4/5
Total	89/100

The valuation score is 4 of 100.

Valuation score	Score
Discount to the weighted Base (price 46% above it)	0/50
Protection against the Bear (-57.5%)	0/30
Market-implied asymmetry (Bull -0.9%, Bear -57.5%)	4/20
Total	4/100

Business quality and valuation are shown separately and are not averaged. One of the best businesses in this sector, at a price that even the Bull case does not reach, is a Reduce.

Confidence score

Confidence	Score
Data and sources (company results presentation read; accounts not read)	4/5
Business predictability (cyclical equipment demand, memory mix)	2/5
Valuation robustness (6% divergence, one multiple behind both)	4/5
Accounting transparency	4/5
Scenario dispersion (blended ¥21,078 to ¥49,112)	2/5
Total	16/25, grade C

The grade measures how well the model is identified, not the chance of a loss. The cycle and the exit multiple are the reasons it is a C.

What I would watch from here

Green

December quarter operating income comes in above ¥58.1 billion.
The company gives a full 2027 outlook with shipments above ¥600 billion a year.
High bandwidth memory shipments resume after the December pause.

Yellow

Gross margin below 70%.
Inventories rising faster than sales.
Chinese assembler shipments falling for a second quarter.

Red

Operating margin below 40%.
A cut to the shipment plan from a major memory customer.
Free cash flow staying below 60% of net income.

The three most important thesis breakers

The equipment cycle turns down in 2027, as in my Bear case.
The multiple falls to 18 to 20 times earnings on any growth miss.
Competitors close the technical gap in thinning and dicing for advanced packaging.

Stress test and independent check

I recomputed every value, the sensitivity grid and the reverse valuation with a separate script. That checks the arithmetic, not the forecasts.

Input	Status
Sales, profit, shipments, balance sheet, cash flow, guidance, dividend	Reported in the company's September 2026 quarter results presentation
Management commentary on mix, customers and outlook	Summary of the earnings call from a data service, not the transcript
Price	Tokyo close on October 30, 2026 from a broker data page; the November 2 close was not collected
Shares	Derived from net income and earnings per share
Net income path, exit multiples, through-cycle earnings	My assumptions
Net cash	Cash and deposits; borrowings not read

The weakest point is the exit multiple, because both methods use 24 times, so their agreement is partly circular. The strongest bull argument is that AI packaging sustains growth above my 5% to 8% fade, which is the ¥49,112 Bull. The strongest bear argument is that equipment demand reverses and the multiple falls to 18 times, which is the ¥21,078 Bear.

Final verdict: reduce at ¥49,540

The question is not whether DISCO is a great company. It is whether a business that guides profit down 14% from one quarter to the next deserves to be valued as if profit only goes up.

Verdict: REDUCE at ¥49,540. Weighted Base ¥34,025, 31.3% below the price, cross-check ¥33,005, Confidence C (16/25). Required margin of safety 42.5%, buy range ¥18,978 to ¥20,151, conservative buy limit ¥18,978.

For an existing holder, the original cost is irrelevant. The price is above even my Bull value, so I would trim a large position and keep a core holding for the franchise, which is the part of the business I trust. New money waits for a price at or below ¥20,151. The next test is the December quarter result and the guidance for the quarter after December.

Sources

Primary sources are DISCO's second quarter results presentation (October 22, 2026), a summary of the earnings call and a results review from VLSI Korea.

The reference market price is ¥49,540, the October 30, 2026 Tokyo close, from Matsui Securities. The share count is derived from reported net income and earnings per share. The call summary is not a transcript.