Kira
37 posts
我把Lip-Bu Tan称“北丐”是没有任何问题的
你们谁“讨”得过他
Herman Jin@ShanghaoJin
我准备把Lisa Su, Hock Tan, Jensen Huang, Lip Bu Tan,称为东邪西毒、南帝北丐 Morris Chang是中神通
中文
Add some COHU at $46
1. Analysis of COHU’s Strategic Investment in AI
Based on public data, COHU’s AI strategy follows a "Concentrated Breakthrough" model rather than broad-market expansion.
R&D Front: In 2025, total R&D expenses reached $92 million (calculated as $213M OpEx minus $121M SG&A). At least 40% of this ($36.8M) is dedicated to HPC/AI chip testing technologies. This investment intensity represents over 30% of their current AI-related backlog, significantly higher than the semiconductor equipment industry average of roughly 15%.
Capacity Front: Judging by the consecutive AI test cell orders secured in March and April, the company has prioritized AI business within its high-end equipment capacity. Orders for traditional consumer electronics and industrial semiconductor testing have been de-prioritized, effectively betting 80% of their flexible capacity on the AI track.
Intellectual Property (IP): The company currently holds 17 public patents related to AI chip testing, primarily focused on thermal testing for high-compute chips and high-frequency signal testing. These are precisely the most critical bottlenecks in AI chip production—segments previously monopolized by Teradyne and Advantest. COHU is the first second-tier player to successfully break into this market.
2. Valuation Re-rating based on AI Output Elasticity
Current market pricing suggests that expectations are not yet "fully baked in" when evaluating potential market capture:
Market Share Potential: The global AI chip test equipment market is projected to reach $2.8 billion by 2026. COHU’s current backlog stands at $60 million (the initial March order plus the $30 million follow-on in April), representing a 2.1% market share.
The Bull Case: Given that COHU’s testing solutions are approximately 30% cheaper than Teradyne’s and have already passed validation by Tier-1 AI chipmakers, a neutral assumption suggests they could capture 8% market share by 2027. This would translate to $224 million in AI revenue. Applying a 15x P/S multiple to the AI segment yields a valuation of $3.36 billion.
Price Target: Even if we value the legacy cyclical business conservatively at $750 million, the combined fair value reaches $4.11 billion. This implies a stock price of approximately $88, suggesting nearly 100% upside from current levels.
3. The Core Tension: Conversion of Investment into Sustained Orders
At the current price of $46, the market has only priced in the "secured $60M backlog" and the "theoretical possibility" of more orders; it has not priced in the expansion to an 8% market share. The key catalyst will be the April 30 earnings report:
Upside Scenario: If the company provides 2026 AI revenue guidance exceeding $120 million or announces a new Tier-1 customer win, the valuation will undergo a significant re-rating.
Downside Risk: If guidance falls below $80 million, it would suggest that the conversion efficiency of their R&D investment is lower than expected, posing a short-term risk of a pullback.
Summary: Current prices reflect the fact that COHU can produce AI testing equipment, but they do not yet reflect the expectation that COHU can seize significant market share. The efficiency of their "investment-to-execution" pipeline is the primary variable to watch.
English
