Japan Deep Value
Japanese Equity Research, Deep Value & Below-Book Opportunities
Advantest's valuation is distorted by passive flows, which pushed the stock up 6% after lunch for no reason to 79x earnings, or 69x forward. Passive investors keep the market from behaving normally even as Nikkei tries to reduce its weight. Uniqlo's parent is similarly stretched at 45x forward earnings while its chairman sells. x.com
Accelerating earnings mean ASML is not expensive right now. It trades at 24x forward earnings and 19x forward EV/EBITDA. x.com
Indie Semiconductor is expensive against its earnings at 36x forward earnings, while chipmakers usually get about 18x. Fair value sits around $1.31-1.66, or 18x next year's $0.08 EPS, so at $2.97 buyers pay about 78% over fair, and the valuation leans on a forecast 61% above trailing earnings. This read rests on valuation levels from reported earnings and estimates only. x.com
Advantest proved it deserves a premium valuation: its SoC tester market share rose from 56% in 2024 to 66% in 2025, widening its lead over Teradyne. It also received its first high-volume ATE order for silicon photonics. x.com
Low multiples give AI infrastructure good risk/reward as demand outruns supply and agentic workloads have runway. TSMC at 17x 2027E EPS appears to have all the accelerator and CPU demand it can handle this year and next; Ultra Clean is at 8x, MKS 12x, TTM 13x, Nvidia 14x and Celestica 15x. Kioxia at 3x and Samsung preferreds at 2.5x carry a big discount on pricing and cycle fears. Multiples imply a 2027 capex peak and meaningful drop; 2028 capex growth is expected, but below 2027's rate. x.com … Show more
ASML is cheap against its earnings at 29x next year's, versus the 36x chip-equipment makers usually get. Fair value is $1,883 to $2,397, or 36x next year's $60.08 EPS give or take 12%, leaving the stock 9% below the low end. The read leans on forecasts, with forward earnings 104% above trailing, and is a valuation read only, with no fundamental work. x.com
Advantest beat because it simply grew at its own historical upcycle pace, which Wall Street underestimated. Revenue came in at ¥367B versus ¥336B consensus and a ¥360B forecast; analysts modeled just 2.3% sequential growth, while Advantest's median in an upcycle is closer to 10%. Earlier reports from Teradyne, TSMC and Disco had already signaled HBM4 pulling tester demand forward; Teradyne memory-test revenue below $170M or an AI mix under 65% would have broken that chain. x.com
Lam Research is cheap against its earnings, trading at 30x next year's earnings when chip-equipment makers usually get 38x. Fair value is $392.57 to $499.63, or 38x next year's $11.74 EPS give or take 12%, so at $347.49 it trades 11% below where fair starts. The valuation leans on the forecast, with forward EPS 100% above trailing. This is a levels-only read off reported earnings and revenue estimates, with no fundamental work. x.com
Onto Innovation looks cheap against its earnings, trading at 31x forward earnings versus the roughly 38x chip-equipment makers usually get. Applying 38x to next year's forecast $9.95 a share, give or take 12%, puts fair value at about $333–423, versus $308.30 now. The case leans on that forecast, with forward EPS 102% above trailing, and rests on valuation levels from reported earnings and estimates, not fundamental work. x.com
Aehr Test Systems is key to AI manufacturing and can even be a bottleneck, but at almost 20x revenue through 2028 at the current price, it is stratospherically priced. The risk is giant if something goes wrong for the company or for the AI ramp, and even a slowdown could be deadly at such high valuations. Tech moats are not the same as financial moats. x.com
ASML's valuation is almost back to its 2021 high. x.com