Cohu: AI Test Exposure Can Still Pull Earnings Higher

Analyst rates COHU a buy, arguing its AI/HPC chip-testing pipeline and recurring revenue are underappreciated. Management targets $600–800M revenue, 48–50% gross margin, and $1.50–$3.25 non-GAAP EPS mid-term, with potential upside if the AI/HPC funnel converts into revenue.

Analysis

AI/HPC exposure matters here less as a volume story and more as a content-per-unit story: advanced chips require more test insertion, tighter thermal control, and more binning iterations, which can lift ASPs and gross margin faster than headline wafer growth. That creates a cleaner operating leverage setup than a generic semi-capex rebound, and it can also pull share away from broader ATE vendors whose platforms are optimized for higher-volume, less complex devices. The second-order winner is the installed base/service stream: once COHU is embedded in a production flow, recurring consumables and support can smooth the cycle and make the revenue mix look less volatile than investors assume.

The near-term catalyst is not the strategic narrative; it is order conversion over the next 1-3 quarters. The market will care whether management can prove that AI/HPC-related wins are translating into backlog and not just pipeline, because a missed conversion window would quickly re-rate the stock back toward cyclicality. Key risks are customer concentration and capex digestion: if hyperscaler/custom-ASIC programs slip, or if test intensity proves lower than expected, the margin model can disappoint even if revenue grows.

Contrarian view: the market may be underestimating that testing, not wafer fab, becomes the bottleneck in advanced packaging. If that is right, COHU’s true TAM expansion is broader than a simple AI spend proxy. But the flip side is that the stock is vulnerable to being priced on hope before proof, so the cleanest risk/reward is to buy only after evidence of order inflection and margin leverage; otherwise this is a watch item, not a blind momentum long.

Key Decisions for Investors

• Buy COHU on pullbacks into the next earnings cycle; target a 6-12 month hold with thesis validation from book-to-bill and gross margin inflection. Falsifier: two consecutive quarters without order growth or margin expansion.
• If you want cleaner risk, use a 6-12 month call spread on COHU instead of common stock to cap downside while keeping exposure to an AI/HPC re-rating.
• Pair trade: long COHU / short TER on a relative basis if you want to express rising test intensity in AI/HPC. This works best if advanced-node demand is the driver; it breaks if broad-based semi capex rebounds and dilutes the relative story.
• Set an alert on management guidance: if mid-term revenue and EPS targets are not narrowed upward within the next 1-2 quarters, treat the pipeline thesis as unproven and reduce exposure.
• Avoid chasing after an initial spike; the better entry is a post-earnings reset or a failed selloff that holds prior support, since the market is still likely underpricing execution risk.