Why Cohu Stock Raced Nearly 6% Higher Today

Baird initiated coverage on Cohu (COHU) with an “outperform” (buy) rating and a $65 price target (~18% upside). The stock jumped nearly 6% on the day as the thesis centered on Cohu benefiting from AI data-center/hardware validation demand, with services spanning software analytics, power management, and hardware testing. The analyst also argued the business can hold up if semiconductors soften, supporting a medium-to-long-term positioning play.

Analysis

COHU is a second-order AI beneficiary: the spend is not on the compute chips themselves, but on the increasingly expensive validation step that comes with advanced packaging, higher power density, and tighter yield requirements. That creates leverage to AI capex even if headline semiconductor unit growth slows, because test/diagnostic content per shipped device can rise faster than the market assumes. The better read-through is to OSATs and advanced packaging ecosystems (ASE, Amkor, TSMC-related capex) than to generic semiconductor demand; if those budgets stay elevated, COHU can compound faster than the broader equipment group.

The catch is timing. This is usually a lagging capital-spending story: orders and backlog tend to confirm 1-2 quarters after the AI capex wave is announced, so the immediate move is mostly multiple expansion on narrative. If hyperscalers digest spend, or if customers shift mix toward fewer but larger deployments, test intensity can underwhelm and the stock can de-rate quickly because COHU is still a cyclical industrial tech name, not a pure software-like AI compounder. The most important falsifier is a lack of booking acceleration and no margin lift in the next 1-2 earnings prints.

Contrarian view: the market may be overstating the durability of the AI tailwind and understating how much of this rally is beta to semis sentiment. The business does have downside support from service/software and a more specialized moat than many small-cap hardware names, but that does not make it immune to a semiconductor digestion phase. My base case is modest outperformance if AI capex broadens, but not a straight-line rerating unless management can prove that AI-related orders are moving from story to backlog.

Key Decisions for Investors

• Tactically long COHU on pullbacks over the next 1-3 weeks; use it as a higher-beta AI infrastructure proxy, but size smaller than NVDA/AMAT because the earnings translation is slower and more cyclical.
• Pair trade: long COHU / short SOXX for a 1-3 month relative-value position to isolate AI test-equipment upside from broad semiconductor beta; thesis breaks if SOXX outperforms on a memory or foundry-led rally.