JPMorgan upgrades Teradyne to Buy after 20% pullback, shares jump

JPMorgan upgraded Teradyne to Overweight and kept a $400 price target, implying about 31% upside from Wednesday’s close of $306.33. The bank highlighted multiple AI-driven growth vectors, including merchant GPU testing, custom AI chips, memory test demand from HBM complexity and new fab builds, and emerging silicon photonics/co-packaged optics. JPM raised 2026 adjusted EPS to $7.00 from $6.65 and 2027 EPS to $9.80 from $9.50, though it noted execution and market-share risks.

Analysis

The key implication is that the market is still treating this as a cyclical test-equipment story, while the earnings power is increasingly being re-rated as an AI infrastructure tollbooth. If merchant GPU qualification and HBM-related memory test both scale simultaneously, TER’s mix shifts toward higher-margin, less price-elastic programs, which can expand operating leverage faster than consensus models that extrapolate lumpy book-to-bill behavior.

The second-order beneficiary set is broader than the headline suggests: every incremental AI server platform needs validation across multiple nodes, so the real winners are not just the dominant chip vendors but also the foundry and OSAT ecosystems that drive test intensity per wafer. That also means TER becomes a picks-and-shovels expression on AI capex that is less exposed to any single hyperscaler’s custom silicon roadmap than the market assumes.

The main risk is timing mismatch. The bullish case is multi-year, but the stock is likely to remain sensitive to quarterly order digestion and any hint that customers are deferring test spend after front-loading AI capex elsewhere; that can create sharp but temporary multiple compression even if the end-market thesis remains intact. A failure to hit share gains in merchant GPU would matter less than feared if memory and photonics ramp, but consensus may still punish any near-term disappointment as if the entire bull case is binary.

Contrarian setup: the move may still be under-owned rather than over-owned because the market is likely underestimating how much of TER’s upside can come from adjacent AI test categories before the merchant GPU opportunity fully matures. The cleaner expression is not a straight momentum chase, but a staged entry around volatility, because the valuation can re-rate quickly once investors see evidence that non-GPU AI test revenue is filling the gap between current bookings and the longer-dated target.

Key Decisions for Investors

• Go long TER on pullbacks over the next 2-6 weeks; use weakness tied to quarterly order noise as an entry point, with a 6-12 month target for re-rating as AI test mix expands. Risk/reward favors a 20-30% upside path if share gains broaden beyond merchant GPU.
• Buy TER Jan-2027 call spreads to express the multi-year earnings step-up while limiting the risk of near-term lumpy demand. This is the better structure than stock if you expect the market to underwrite the thesis only after additional proof points.
• Pair trade: long TER / short a basket of slower-growth industrial semi-cap names over 3-9 months. The thesis is that AI-related test intensity will decouple TER from the broader equipment cycle, while legacy capex names remain vulnerable to digestion.
• Sell upside in the near term if TER spikes on the downgrade reversal; use covered calls or trim 20-30% into strength. The stock likely remains headline-sensitive until the market sees a cleaner quarterly cadence, so chasing every gap higher is poor risk-adjusted execution.