Cohu Is An Underfollowed AI Testing Play

Analyst rates Cohu a Strong Buy with an $83 price target, implying ~57% upside. The thesis centers on expanding AI processor testing and related capabilities (e.g., high-bandwidth memory inspection, power management, analytics) targeting a $1.6B serviceable market, with software (Eclipse/Neon/Diamondx/PAICe) and higher utilization projected to add ~$1.05 of incremental EPS and lift 2027 non-GAAP EPS to $1.63.

Analysis

The market is likely still valuing COHU like a cyclical legacy test-tools supplier, but the mix shift toward AI processors and HBM changes the earnings elasticity: when utilization turns, incremental gross margin can re-rate faster than revenue because the software/inspection layer carries lower marginal cost than hardware shipments. That makes the stock less about near-term unit growth and more about whether management can prove that content-per-package is structurally higher in advanced packaging and memory inspection.

Second-order winners are the back-end ecosystem names tied to test intensity and yield management: OSATs such as AMKR/ASX, and equipment peers with exposure to advanced test and inspection like TER and FORM. The loser is the market’s old assumption that test is a late-cycle, low-multiple business; if AI/HBM requirements persist, COHU can earn a better multiple than a pure cyclical because its revenue becomes more tied to complexity per chip than to handset/PC unit volumes.

The main risk is timing: analyst math can outrun bookings for 1-2 quarters, and the stock can stall if customers delay capex or if yield improvements reduce test steps faster than AI content expands. The thesis is falsified if management does not convert the addressable-market story into order growth and gross margin expansion over the next two earnings prints, or if AI spend broadens but test intensity concentrates with competitors rather than COHU. For now this looks more like a 6-18 month rerating story than a clean immediate catalyst trade.

Key Decisions for Investors

• Go long COHU on weakness rather than chasing strength; use a 3-6 month horizon and require evidence of order acceleration or margin inflection before adding. Risk/reward improves materially if the stock stays near prior resistance while fundamentals re-rate.
• Pair trade: long COHU / short a broad semiconductor equipment basket via SOXX or a larger-cap test peer on a relative basis. This is a cleaner way to express a mix-shift thesis without taking full sector beta.
• Use a call spread in COHU for the next 1-2 earnings cycles if liquidity allows; the setup is asymmetric because upside depends on multiple expansion plus estimate revisions, while downside is cushioned if the AI narrative merely proves incremental.
• Watch AMKR and ASX as confirmation indicators over the next quarter: if backend packaging margins and commentary improve, it supports the idea that test content is rising across the chain; if not, treat COHU as a story stock until bookings validate.