Jefferies raises Cohu stock price target on AI test momentum

Jefferies raised Cohu’s price target to $60 from $55 while keeping a Buy rating, citing accelerating AI and high-performance computing demand. Cohu also lifted June-quarter revenue guidance to $144 million versus the Street’s $127 million estimate and increased its 2026 compute revenue midpoint to $90 million from $72.5 million. Offsetting that strength, Q1 2026 EPS came in at $0.01 versus $0.03 expected, although revenue beat consensus at $125.1 million.

Analysis

The market is starting to re-rate Cohu less as a cyclical test-equipment name and more as an embedded AI infrastructure lever. That matters because the revenue mix is shifting toward a few high-value customers and a broader compute stack, which usually compresses the path from pilot wins to meaningful backlog visibility. The second-order effect is that suppliers tied to AI packaging, HBM, and advanced-node test may see a faster multiple reset than the broader semiconductor capex basket, because the spend is moving from discretionary to qualification-driven.

The near-term risk is margin optics, not demand. When a business is ramping new platforms, gross margin can look “stuck” for several quarters even as unit economics improve underneath, and that often creates a sell-the-news window after upside guidance. If the compute pipeline converts slower than expected, the stock can de-rate quickly because a lot of the current move is already discounting 2026 earnings power rather than current cash generation.

Contrarian take: consensus may be underestimating how narrow the current demand driver really is. If AI/HPC broadens into CPUs, ASICs, and networking silicon as suggested, Cohu’s addressable market expands; if it stays GPU-centric, the growth narrative becomes more fragile and more concentrated than the headline implies. The biggest tell over the next 1-2 quarters is whether industrial/autotesting utilization stays near the level that triggers incremental system purchases—if it does, the stock can keep working even if AI revenue is lumpy.

On the HBM side, the positive read-through is to capital equipment names with exposure to memory quality control and advanced packaging, but the timing is important: the upside is more visible in 2H26 than immediately. The move in Cohu also raises the bar for peers with weaker design-win momentum, so relative value should matter more than beta exposure.

Key Decisions for Investors

• Go long COHU on any 5-8% post-upgrade pullback; target a 3-6 month hold with upside to a higher 2026 earnings multiple if compute pipeline converts, but size modestly because current valuation already discounts execution.
• Use a pair trade: long COHU / short a weaker semi-capex or test-equipment peer with less AI exposure over 1-3 months; the spread should widen if the market keeps rewarding clear AI monetization and punishing slower ramp stories.
• Sell near-dated covered calls against an existing COHU long into the next earnings/guidance update; margin expansion skepticism creates attractive premium capture while preserving upside through the 2026 thesis.
• Add a small tactical long in HBM-related test/inspection beneficiaries on weakness, but only as a 6-12 month trade; the cleaner catalyst is the 2026 ramp, not the next quarter.
• If COHU fails to hold gains after another revenue beat, consider a short against the move via puts or a partial hedge; the risk/reward flips if growth remains concentrated and margins do not inflect by the next two quarters.