Teradyne (TER) Q2 2025 Earnings Call Transcript

Teradyne reported Q2 revenue of $652 million and non-GAAP EPS of $0.57, both above the midpoint of guidance, with AI compute driving strength in Semi Test and management signaling improving visibility into the second half. Q3 guidance calls for $710 million to $770 million in sales and EPS of $0.69 to $0.87, while AI compute is expected to become the dominant driver of Semi Test revenue and memory demand is expected to snap back later in 2025. The company also returned $316 million to shareholders in the first half through buybacks and dividends, and management described new robotics and GPU opportunities as meaningful 2026 growth drivers.

Analysis

The key read-through is that Teradyne’s mix is shifting from cyclical utilization recovery to an AI-driven product-cycle expansion, which is materially better for durability but also more timing-sensitive. The first-order beneficiary is TER itself, but the second-order winner is any supplier enabling higher-throughput test and packaging complexity; if AI accelerators keep adding test insertion steps, the industry is moving from a capacity-disciplined market to a content-intensity market, which should support pricing and utilization for best-in-class ATE vendors while pressuring laggards with weaker throughput or software integration.

What the market may be underestimating is that the bigger near-term upside is not just more AI units, but more test steps per unit across SOC, HBM, and silicon photonics. That creates an asymmetric setup: revenue can compound faster than unit demand if customers keep pushing reliability upstream, while competitors with legacy footprints may struggle to win sockets where supply-chain resilience and dual sourcing matter more than raw installed base. The Robotics reset is more interesting as a 2026 operating leverage story than a 2025 revenue story; if management is right, the segment could swing from a drag to an earnings lever just as semi-test momentum broadens.

The main risk is not demand disappearing; it is quarterly phasing slipping into Q1, which could make the second-half setup look weaker than the full-year trajectory and create a temporary multiple reset. A second risk is that the GPU opportunity remains aspirational until a production win is locked, so any disappointment in qualification timing could hit sentiment even if the long-term TAM is real. The concentration in one customer also means TER is increasingly dependent on a small number of hyperscaler-like decision makers, so a procurement delay could matter more than macro does over the next 1-2 quarters.

Key Decisions for Investors

• Long TER into Q3/Q4 on dips: the setup is a 2-3 quarter re-rating trade if AI compute phasing converts into orders; use a 6-9 month horizon and expect volatility around Q4/Q1 shipment timing.
• Pair trade: long TER / short a more mature industrial automation name with weaker AI exposure and slower operating leverage; the relative thesis is that TER’s mix shift should drive higher growth and margin expansion into 2026.
• Buy upside call structures on TER rather than outright stock for the next earnings window: the catalyst is sequential AI compute acceleration, but the main risk is timing noise, making defined-risk convexity preferable.
• If seeking a second-order winner, screen suppliers to semiconductor packaging/test and photonics exposure; the thesis is higher test-content intensity, so names leveraged to HBM4 and electro-optical test should outperform on every incremental AI design win.
• Do not chase Robotics for 2025 revenue: any long-based on the large customer win should be sized as a 2026 operating leverage call, not a current-year sales story.