Teradyne Stock Plunges 15% in a Month: Should You Buy the Dip?

Teradyne shares fell 15.4% over the past month despite benefiting from AI-driven data-center demand, with management noting that >60% of Q2 2026 revenue was AI-driven. The article attributes the stock’s underperformance to customer concentration, program timing, muted mobile demand, memory margin dilution, higher operating spend, and currency exposure. Still, Teradyne initiated positive Q3 2026 guidance of $1.20–$1.30B revenue (consensus $1.27B) and non-GAAP EPS of $1.85–$2.15 (consensus $2.03), alongside premium valuation (forward P/S 9.35x vs 6.03x for the sector).

Analysis

TER’s setup is a classic “good business, expensive stock” trade: the AI test content is real, but the market is already paying for a multi-year growth story while the near-term earnings path still depends on a narrow set of ramps. The key second-order issue is mix: memory/HBM and high-power digital testing can lift revenue faster than gross profit if utilization and product mix stay unfavorable, so EPS can lag headline demand. That makes the next 1-2 quarters more about margin conversion and backlog quality than revenue growth alone.

Relative winners are the companies with the most direct AI infrastructure capex pass-through and the least cyclicality in end markets. TER should keep taking share in advanced test, but VRT still looks like the cleaner “picks-and-shovels” beneficiary if datacenter buildout stays hot; KLAC is more levered to sustained WFE digestion, while COHU remains the most vulnerable to mobile and customer-concentration noise. A useful read-through is that AI is becoming a test-intensity story: more transistors, more power rails, more lanes, more validation steps — which supports TER structurally, but also raises the bar for execution because any slip in program timing can hit consensus hard.

The contrarian view is that the recent drawdown may have partially washed out the obvious negatives, but it has not washed out valuation risk. With the stock still screening expensive versus the broader tech complex, upside likely needs either an upward guide reset or clear evidence that AI-driven revenue is translating into gross margin recovery. Falsifiers are straightforward: if the next update shows AI mix flattening, memory revenue normalizing, or operating expenses staying elevated while margins stay pinned, the multiple can compress further even if revenue remains healthy.

Key Decisions for Investors

• Initiate a small relative-value long TER / short COHU position over the next 2-6 weeks: TER has the cleaner AI test mix, while COHU remains more exposed to legacy end-market weakness and customer timing risk. Target 8-12% relative outperformance if TER confirms margin stabilization; cut if TER gross margin or guide disappoints.
• Avoid chasing TER outright at current valuation until the next earnings print or channel check confirms that AI-driven revenue is converting into operating leverage. If the stock re-rates above ~10x forward sales without a margin inflection, treat that as a trim signal rather than a buy.
• For event-driven accounts, consider a TER call spread only into the next catalyst if implied volatility is not already elevated; the trade works only if management raises confidence on HBM/advanced test and power-test ramps. Falsifier: guide lands at the low end and EPS is again diluted by memory mix.
• Prefer VRT over TER for a cleaner 1-3 month AI infrastructure expression if the goal is datacenter capex beta rather than semiconductor-test-specific execution risk. TER has more idiosyncratic program timing and concentration risk, so use it as a more selective, lower-conviction AI expression.
• Set a watch item on KLAC and WFE order commentary: if wafer fab spend rolls over, TER’s revenue multiple can compress faster than consensus expects because the market is pricing a sustained transistor-growth cycle. Any sign of capex digestion is a reason to reduce TER exposure.