Cohu director William Bendush sells $368,550 in shares

Cohu director William Bendush sold 7,500 shares for $368,550 at a weighted average price of $49.14, leaving him with 24,780 indirect shares plus 10,257 RSUs. The article also highlights improving fundamentals: Q1 2026 revenue beat estimates at $125.1 million versus $122.07 million expected, and management guided Q2 revenue to $144 million, up 15% sequentially. Analyst price targets were raised to $50-$60 on AI compute and high-performance computing demand, though the stock is already near its 52-week high and appears overvalued on InvestingPro's Fair Value analysis.

Analysis

The key signal is not the insider sale itself, but the market context around it: COHU is behaving like a late-cycle AI-capex beneficiary with the stock already pricing in a strong second-half ramp. That makes the setup fragile because any disappointment in AI-related test demand, order timing, or gross margin mix can compress a stretched multiple quickly; when a cyclical semiconductor equipment name rerates this hard, the next leg is usually driven more by estimate revisions than by headlines. The insider sale via a 10b5-1 plan is not a bearish tell by itself, but it does reinforce that management sees enough liquidity at current levels to de-risk into strength.

Second-order, the more interesting read-through is to AI infrastructure capex breadth. COHU’s exposure suggests the market is pushing beyond compute silicon into adjacent picks-and-shovels like power devices and test equipment, which often lag the headline AI trade but can outperform on earnings revision momentum when the buildout broadens. If that thesis holds, the beneficiaries are the less obvious enablers of the AI stack; if it does not, they tend to de-rate faster than the semis themselves because their end-market visibility is thinner and their valuation support is weaker.

The contrarian view is that the current optimism may already be pulling forward 2H demand that has not fully converted into durable backlog. In that case, COHU becomes a classic "good narrative, mediocre execution" stock: revenue can look fine while EPS disappoints due to mix, utilization, or operating leverage. For NVDA, this is a softer positive only if the broader AI capex cycle is expanding; otherwise, the market may be over-interpreting adjacent supplier strength as confirmation of an even larger AI spending wave.

Key Decisions for Investors

• Trim or avoid initiating new COHU longs above recent highs; use any rally extension toward the mid-$50s to fade strength with a 1-2 quarter horizon, since valuation can compress 15-20% quickly if guidance does not re-accelerate.
• If seeking AI-capex exposure, prefer NVDA over COHU on a risk-adjusted basis: long NVDA / short COHU as a pair trade for 1-3 months, betting that platform leaders capture the multiple premium while adjacent equipment names are more exposed to order volatility.
• Buy COHU downside protection into the next earnings window: 1-2 month puts or put spreads to hedge the risk that guide-up optimism is already priced in; target a 2:1 or better payoff if the stock retraces from elevated levels.
• Monitor confirmation in AI power and test supply-chain names over the next 4-8 weeks; if they fail to participate, treat COHU’s strength as a single-name rerating rather than a durable sector signal.
• For traders wanting a cleaner AI continuation expression, add NVDA on dips rather than chasing COHU strength, as NVDA has better liquidity, stronger narrative durability, and less execution risk tied to a single product cycle.