$16.68
$40.99
bear $16.68
base $24.79
bull $40.99
margin of safety $15.00
market price* $78.63
Kulicke & Soffa designs and sells the machines that connect semiconductor chips to their packages. Its core franchise is the ball and wedge wire bonder, where it has been the leading supplier for decades; it also sells thermocompression and fluxless thermocompression bonders for advanced packaging (its Advanced Solutions line), dispensing equipment, and a recurring stream of consumables (bonding capillaries and blades) and service. Customers are chip assemblers and packaging subcontractors, so revenue rises and falls with their spending on new capacity. In fiscal 2025 about 90% of revenue shipped outside the U.S., about 54% went to customers headquartered in China, and the ten largest customers took about 55%. In 2025 the company decided to close its Electronics Assembly equipment business. How we measured earnings. We start from operating cash flow, strip out working-capital swings, subtract stock compensation as a real cost, and subtract the capital spending needed to maintain the business. Two judgements matter. First, inventory write-downs are treated as a real cost rather than added back: they were $70 million in fiscal 2024 and $45 million in fiscal 2025, much of it tied to the Electronics Assembly exit and older product lines, against a more typical few million dollars a year. Second, goodwill and asset impairments ($35 million in 2017, $22 million in 2023, $44 million in 2024 and $40 million in 2025) are added back because they record money already spent; the lesson that acquisitions such as the Electronics Assembly business did not earn their keep is carried in the bear case instead. Stock compensation has risen from about $6 to $12 million a year in 2016-2018 to $28.5 million in fiscal 2025. On that basis, owner earnings over fiscal 2016-2025 were: $46M, $144M, $77M (depressed by the one-time 2018 U.S. tax charge), $26M, $58M, $375M, $417M, $59M, a loss of $3M, and $32M. That decade includes two upswings (2017-2018 and the 2021-2022 chip-shortage boom) and two busts (2019 and 2024-2025), so it is a fair full cycle. The median year is about $58 million, which is about $1.10 per share on today's 53.4 million diluted shares; the average of the eight years outside the 2021-2022 boom is a similar $55 million, so ordinary upswings are already inside that figure. We use dollars on today's share count because buybacks cut the share count from about 72 million in 2017 to about 53 million, and the cash spent on them is already gone from the balance sheet. The company has no borrowings, so the cash flow before lenders is essentially the same as owner earnings (interest paid is about $0.1 million a year). Where we are in the cycle. Fiscal 2025 owner earnings recovered from a small loss to $32 million, a rebound that has no meaningful growth rate. Our base case puts fiscal 2027 at $150 million, well above that, for specific, dated reasons: net income for the first nine months of fiscal 2026 was $109 million against a small loss a year earlier, revenue in the quarter ended July 4, 2026 was $330 million against $148 million a year before, and in August 2026 management said customer orders already extend into the second quarter of fiscal 2027. This is a cyclical upswing, not growth, so we credit it only as far as that order book reaches and then return to the through-cycle level. We also include about $45 to $62 million of earnings for the quarter that is now ending. Base case ($24.79 per share): the current upswing produces about $150 million in fiscal 2027, a post-boom downturn brings $40 to $45 million in fiscal 2028-2029, and earnings then run at the $58 million through-cycle level, booms and busts included. Wire bonding has been the dominant way to connect chips for decades and remains the majority of packages, so we hold that level for 25 years, then fade it to zero by year 35 as newer connection methods slowly displace it. The business is worth about $908 million today. It owes nothing, and we add $417 million of the $517 million of cash and short-term investments it held on July 4, 2026, keeping $100 million back as working cash. Equity of about $1.32 billion is $24.79 per share. Bear case ($16.68 per share): the order book covers only the first half of fiscal 2027 ($90 million for the year), followed by a trough as deep as 2024-2025 (about $10 million in 2028 and nothing in 2029). The heavy reliance on Chinese customers then costs share as Chinese assemblers switch to domestic bonder suppliers and export rules tighten, holding the through-cycle level to $35 million for 20 years before fading to zero by year 30. The business is worth about $475 million; with no debt and the same $417 million of surplus cash, equity is about $891 million, or $16.68 per share. Even here the company's cash covers a large part of the value, which is why the downside is limited. Bull case ($40.99 per share): the company's own levers execute. Record Advanced Solutions revenue continues, capacity it has already roughly quadrupled stays full, and the upswing runs through fiscal 2028 ($230 million and $140 million), similar to the 2021-2022 boom. After that, through-cycle earnings average $115 million, close to the full ten-year average of $123 million that includes one supercycle, held for 25 years and faded to zero by year 40. The business is worth about $1.77 billion; with $417 million of surplus cash, equity is about $2.19 billion, or $40.99 per share. What we left out. We give no credit for growth in advanced packaging beyond orders in hand, for AI-driven demand that has not yet shown up in orders, for the expanded Singapore Advanced Solutions plant due in the second half of fiscal 2027, or for savings from closing Electronics Assembly. None of these can yet be measured from the filings. Each would raise the estimate if it shows up in results. Testing the base case. The weakest point is the $58 million through-cycle level, since fiscal 2023-2025 averaged only about $29 million. But those years carried about $120 million of mostly one-off inventory write-downs from the Electronics Assembly exit and product transitions. Adding back the excess over a normal year puts fiscal 2024 and 2025 at roughly $60 to $70 million, consistent with the level we use. The real risks are China (over half of sales), the ten largest customers (about 55%), and the cycle itself. A downturn can cut owner earnings to near zero for one or two years, as it did in 2019 and 2024. Change since our last review. We previously estimated $16.30, $28.70 and $52.50 per share. We now discount every case at a single 6% rate. Earnings no longer grow every year forever; they hold at a level supported by the record and then fade. We measure that level across a full decade that includes the 2024-2025 bust, and we credit the current upswing only as far as the company's order book reaches. That leaves the low case about the same and lowers the middle and high cases.