Teradyne at Goldman Sachs Communacopia + Technology Conference: ai push widens

Teradyne outlined a multi-year AI-driven growth strategy, citing 59%-61% gross margins, OpEx growth below half of revenue growth, and substantial operating leverage as semiconductor test, robotics and system-level test scale. Management expects meaningful merchant-compute share gains in 2027 and potentially about 30% of a major account over three to five years, while hyperscaler custom-silicon opportunities each represent hundreds of millions of dollars to $1 billion in annual TAM. Semiconductor CapEx is projected to rise from about $120 billion in 2025 to $250 billion-$300 billion in 2028-30, supporting higher test intensity; Teradyne also expects its marquee e-commerce robotics customer's revenue contribution to triple again in 2027.

Analysis

TER’s incremental earnings setup is more sensitive to qualification timing than to the headline semiconductor-capex cycle. The two 2027 second-source ramps create asymmetric upside if production releases occur on schedule, but the market is likely already capitalizing a material portion of that outcome after the stock’s sharp rerating. With gross margin structurally bounded near 60%, upside must come from revenue mix and operating leverage; therefore, investors should focus on backlog conversion, tester utilization and incremental operating margin rather than broad AI commentary.

The non-obvious exposure is that multi-die packaging raises the economic cost of a false pass, making test spend less discretionary even if unit demand weakens. That supports TER versus more wafer-fab-equipment-cyclical names during a modest semiconductor correction. Conversely, the merchant-compute opportunity is a slow share-conversion process against Advantest (ATEYY), not a winner-take-all displacement; any qualification slippage or customer platform-standardization decision could push the anticipated 2027 earnings inflection out by several quarters and compress TER’s premium multiple.

Robotics is strategically valuable but too small to underwrite near-term valuation: a concentrated e-commerce program can generate outsized growth rates from a low base while also creating customer-concentration risk. The more investable second-order beneficiary is electronics-manufacturing automation, where data-center buildouts pull demand for Universal Robots-type collaborative automation; however, this remains vulnerable to hyperscaler capex pauses. CPO is similarly an option value, not a 2026-27 earnings pillar, and technological substitution toward copper or near-packaged optics could delay its addressable market.

Contrarian view: the better relative trade may be long ATEYY versus TER if investors expect the CPO and leading-edge compute test cycle to broaden. ATEYY retains strong positioning in early photonics insertion and benefits if TER’s expected GPU-share gains take longer than management’s multiyear conversion framing implies. TER becomes more attractive on a pullback tied to macro risk or a near-term digestion scare, provided 2027 production qualifications remain intact.

Key Decisions for Investors

• Do not chase TER at current momentum-driven levels; establish a 3-6 month watch position only after confirmation that follow-on merchant-compute devices enter production. Add on evidence of 2027 revenue guidance moving higher, not merely qualification announcements.
• Run a 6-12 month relative-value pair: long ATEYY / short TER in equal semiconductor-test beta. Thesis: TER’s valuation embeds successful share conversion while ATEYY has underappreciated exposure to leading-edge test and photonics. Cover if TER discloses a faster-than-expected production-share ramp or raises forward operating-margin targets.
• For a bullish TER position, use defined-risk upside exposure via 6-9 month call spreads rather than outright shares; the catalyst window is 2027 qualification and customer-volume visibility. Risk/reward is favorable only if the premium paid assumes no more than a modest upside revision, given elevated execution expectations.
• Track hyperscaler capex guidance, AI-accelerator test capacity additions, and memory-test orders over the next two earnings cycles. A reduction in 2027 tester-demand commentary, delayed production release at the second hyperscaler, or gross margin below the stated operating band would falsify the long thesis.
• Avoid treating robotics as a standalone near-term TER catalyst. Reassess only if robotics becomes material enough to change consolidated growth or if customer concentration disclosures indicate the marquee logistics program is diversifying into repeatable deployments.