Teradyne Strengthens AI Test Portfolio: Can It Beat KLAC & COHU?

Teradyne reported that AI-driven revenue exceeded 60% of Q2 2026 sales, with total revenue surpassing $1.3 billion and Semi Test revenue rising 128% year over year; compute revenue surged nearly 600%. Its new Iris 100 platform expands optical testing for microLED AR displays and AI data-center interconnects, while Q3 revenue guidance is $1.20-$1.30 billion. Consensus 2026 EPS is $9.10, implying 129.8% year-over-year growth, although TER's 10.54x forward sales multiple is above the industry's 7.74x.

Analysis

TER's relevant differentiation is not merely AI-test exposure but the potential to capture a higher-value test-content mix at the optical/electrical boundary. If customers adopt a combined test flow, attach rates to the installed UltraFLEXplus base could lift service, software, and utilization economics faster than system revenue alone. The market will need evidence of qualified production programs and orders—not product introductions—to underwrite a durable multiple premium; initial microLED volume is likely too small to move consolidated earnings over the next 1-3 quarters.

Near term, TER's elevated AI sensitivity makes it vulnerable to even a modest guide-down in compute, HBM, or optical-interconnect demand: a 10% miss versus the high end of revenue guidance would likely matter more to valuation than to earnings because expectations and the revenue multiple are already stretched. KLAC is the lower-beta beneficiary of the same advanced-packaging/HBM investment cycle, with process-control demand tied to wafer starts and yield complexity rather than end-test share gains. COHU has more upside torque if GaN power-device testing scales with data-center power buildouts, but its smaller customer concentration and lower order visibility make it a watchlist name rather than a clean TER substitute.

Contrarian view: the key risk is that AI test demand has been pulled forward by rapid accelerator, HBM, and networking ramps. Test equipment is cyclical and customer capex digestion can follow capacity additions quickly; a normalization in utilization during 2027 would expose TER's premium valuation. Conversely, sustained optical-interconnect adoption and rising test intensity per AI module could make current revenue estimates conservative, but this requires bookings, backlog, and gross-margin expansion to accelerate together.

Key Decisions for Investors

• Do not chase TER after the recent rerating; retain or initiate only on a 10-15% pullback or after Q3 results demonstrate bookings/backlog growth and gross-margin resilience. Target a 6-12 month upside case only if AI-related mix translates into estimate revisions; exit on revenue below the low end of guidance or a material sequential decline in Semi Test orders.
• Express a relative-quality AI-test view via long KLAC / short TER over 3-6 months, sized beta-neutral. KLAC offers exposure to HBM and advanced-packaging complexity with less dependence on a high-growth end-market narrative; cover the TER short if TER reports order conversion in optical testing sufficient to raise forward EPS expectations by more than 5%.
• Place COHU on an event-driven watchlist rather than buy now: initiate only after disclosed follow-on Diamond X orders or a meaningful data-center power-test backlog validates repeatability. The relevant catalyst window is the next two earnings reports; lack of follow-on orders would indicate that isolated platform wins are not yet a revenue inflection.
• Monitor TER's book-to-bill, backlog duration, compute-test mix, and gross margin at the next earnings release. A combination of book-to-bill below 1.0x and unchanged earnings estimates would support reducing exposure, as the premium multiple leaves limited cushion for a digestion cycle.