Cohu director Steven Bilodeau sells $460,026 in common stock By Investing.com

Cohu director Steven J. Bilodeau sold 10,257 shares at $44.85 for $460,026, reducing his direct holdings to 52,272 shares including RSUs and DSUs. Offsetting the insider sale, Cohu posted Q1 revenue of $125.1 million, ahead of the $122.07 million forecast, and guided Q2 revenue to $144.0 million, or 15% sequential growth. Analysts turned more constructive, with Stifel, Needham, and Jefferies all raising price targets to $50, $54, and $60, respectively, citing AI and high-performance computing demand.

Analysis

The key read-through is not that Cohu is suddenly a pure AI winner, but that the market is repricing the semiconductor test bottleneck as a second-order beneficiary of the AI capex cycle. As hyperscale spend shifts from GPUs alone toward power conversion, advanced packaging, and reliability testing, the test/equipment layer can see a more durable demand tail than headline AI hardware names, because every incremental watt of AI compute needs validation and yield control. That makes the upside less dependent on a single product cycle and more tied to the multi-quarter buildout of power infrastructure around data centers.

The insider sale is more noise than signal on its own, but it matters in a stock that has already rerated sharply: at these levels, the burden of proof shifts to sustained order conversion and guidance follow-through. The near-term catalyst path is quarter-to-quarter execution, not broad AI enthusiasm; if DiamondX orders or second-half revenue acceleration stall, multiple compression could be abrupt because the stock is now trading on an implied inflection that must be repeatedly confirmed. The biggest fundamental risk is that AI test demand proves more lumpy than investors expect, with customers timing evaluations and capex in bursts rather than a straight-line ramp.

Consensus may be underestimating how much of the move is already in the tape. When a smaller-cap equipment name gets multiple analyst target hikes in quick succession, the marginal buyer becomes momentum and factor-driven, which increases fragility if the next data point is merely good rather than great. In that setup, upside is still intact over months if AI power/test exposure compounds, but near-term asymmetry looks worse than the bullish narrative suggests unless management can convert pipeline into repeatable backlog growth.

Key Decisions for Investors

• Maintain a tactical long COHU, but only on pullbacks toward the low-40s; target a move back to the high-50s over 3-6 months if second-half orders convert, with a stop below the prior breakout area around the high-30s.
• Use a call spread in COHU for event-driven exposure: buy 45/55 or 50/60 calls into the next earnings/guidance update to express upside from AI power-test demand while capping premium at risk.
• Pair trade: long COHU / short a broader semiconductor equipment basket over 1-3 months if you want a purer AI test thesis; the thesis is that COHU’s incremental revenue mix has more near-term AI leverage than the average equipment name.
• Avoid chasing COHU after a strong open or analyst-upgrade gap; the stock is vulnerable to multiple compression if management only reiterates rather than raises, so wait for a post-rally consolidation before adding.
• For NVDA, treat this as a supply-chain validation signal rather than a direct fundamental catalyst; no standalone position change, but monitor whether power/test vendors keep confirming the AI infrastructure spend narrative over the next 1-2 quarters.