Teradyne designs and sells automated test equipment for semiconductor manufacturers and collaborative robots for industrial customers. Semiconductor Test was 79% of 2025 revenue, Product Test 11%, Robotics 10%. [S1][S5]
Sources through 2026-09-11 · Report dated 2026-09-11
Business details
How it makes money
Revenue comes from selling test systems (83% product) and related services (17%), with service revenue recurring and tied to the installed base at 62.9% gross margin. Product revenue is transactional and tied to semiconductor capital spending cycles. [S5]
What drives results
Analysis: Semiconductor capital spending, particularly AI compute and memory test demand, drives the majority of revenue and gross profit. [S3][S5]
Analysis: Product mix between high-margin AI test systems and lower-margin memory testers determines gross margin, which expanded to 59.8% in Q2 2026. [S3]
Key dependencies
Customer concentration: the five largest customers accounted for 44% of 2025 revenue, and one customer accounted for 19%. [S5][S6]
Taiwan exposure: 36% of 2025 revenue came from Taiwan, and critical semiconductor manufacturing is concentrated there. [S5]
Teradyne's stock has risen ~220% on an AI test demand surge that lifted Q2 2026 revenue 104% year over year. The tension is whether this is a durable structural inflection or a cyclical peak that mean-reverts, with customer concentration amplifying both outcomes.
Customer concentration: top five customers were 44% of 2025 revenue, one at 19%. Loss of a major customer or a capital spending pause would materially cut revenue and margins, and concentrated customers have pricing leverage.
Teradyne designs and sells automated test equipment for semiconductor manufacturers and collaborative robots for industrial customers. Semiconductor Test was 79% of 2025 revenue, Product Test 11%, Robotics 10%. Revenue is transactional, tied to semiconductor capital spending cycles, with a smaller recurring service stream (17% of 2025 revenue) from the installed base.
At $370, the market prices sustained 20%+ revenue growth and 30% operating margins for several years, implying 2028 revenue near $6B and FCF of $1.2-1.5B. This assumes the AI test boom persists through 2027 without a cyclical downturn.
Base case intrinsic value is ~$410 per share versus $370 market price, implying ~11% upside. The stock is a hold: good business, narrow moat, but valuation offers inadequate margin of safety given cyclicality. Risk/reward is balanced with bear case downside to $180.
AI and memory semiconductor test demand: current annualized Semiconductor Test revenue is ~$4.5B, versus a historical $2.1-3.7B range. A 10% change affects EPS by ~$1.75. Leading indicator: SEMI equipment billings.
Gross margin: 59.8% in Q2 2026 versus historical 57-59%. Each 100bps is ~$0.22 EPS. Mix shift toward AI test systems drives expansion. Watch product mix and pricing.
Robotics profitability: segment lost $99.4M pre-tax in 2025. Breakeven adds ~$0.55 EPS. Fifth consecutive quarter of growth in Q2 2026; watch order trends and segment margin.
Present-value estimates at the analysis date; DCF forecast horizon 2026-2030 with terminal growth 3.0%, WACC 13.8%, 157.7M diluted shares, SBC expensed. Base case intrinsic value ~$410 per share versus $370 market price.
Scenarios are conditional estimates, not promised returns.
Bear · USD 180.0000
AI capex slows in 2027 as hyperscalers digest capacity; test orders fall 30-40%; revenue drops to $3.5B; operating margin compresses to 18%; EPS ~$4.50; multiple compresses to 25x trough earnings, value ~$180.
AI test demand robust through 2027, then moderates to 6-8%; revenue $5.0B in 2027, $5.5B in 2028; margins normalize at 26-28%; EPS $8-9 in 2028; 35x forward P/E discounted back implies ~$410.
AI becomes a sustained platform shift; revenue reaches $6.5B by 2028, $8B by 2030; margins sustain 30-32%; EPS $12-14 by 2029; FCF $1.8-2.0B; 40x forward P/E discounted back implies ~$650.
The exact revenue contribution of the NVIDIA SoC tester certification to 2026 results is not disclosed. The MultiLane JV revenue contribution is not separately reported. The sustainability of AI test demand beyond 2027 cannot be established from current data.