Teradyne (TER) Q2 2025: AI Compute Now 20% of SOC, Shifting Revenue Mix and Visibility

July 30, 2025 | By EarningsIQ Research

AI compute demand has decisively shifted Teradyne’s revenue mix, with SOC and memory orders now driven by hyperscaler and accelerator ramps rather than consumer seasonality. Robotics restructuring and U.S. manufacturing add long-term leverage, but near-term headwinds persist outside of AI. Management’s confidence is up, underpinned by new program wins and structural market shifts in test utilization.

Summary

• AI Compute Reshapes Demand Patterns: SOC and memory sales are now governed by hyperscaler and accelerator program launches, not legacy mobile cycles.
• Robotics Pivot Gains Traction: Organizational consolidation and a major U.S. customer win set up a 2026 inflection, though 2025 remains muted.
• Visibility Improves for Second Half: Management signals stronger conviction in H2 growth, led by AI compute and memory ramps.

Performance Analysis

Teradyne’s Q2 results highlight a business in transition, with AI compute now the central growth engine. Semi-test revenue, representing the bulk of the company, was propelled by SOC demand tied to AI compute and networking, while traditional mobile and industrial segments stabilized at lower levels. Memory revenue dipped as expected due to shipment timing, but management anticipates a “snapback” in the second half as HBM4, high bandwidth memory for AI, ramps. Notably, compute represented roughly 20% of SOC revenue in Q2, a structural step-change from historical patterns.

Robotics delivered 9% sequential growth following a structural reorganization, but remains below prior year levels, with break-even unlikely this year. Product test and IST, integrated system test, both posted year-on-year gains, buoyed by mobile SLT, system-level test, and HDD, hard disk drive, testers. Operating expenses rose with increased R&D investment, but discipline on cost controls kept sequential growth flat. Free cash flow improved, enabling $117 million in buybacks and $19 million in dividends, returning 138% of first-half FCF to shareholders.

• AI Compute Drives Outperformance: SOC and memory upside was directly linked to AI compute program ramps, not legacy mobile cycles.
• Memory Timing Volatility: Memory revenue fell QoQ but is expected to rebound as HBM4 test insertions expand in H2.
• Robotics Still Subdued: Despite operational gains, robotics remains loss-making in 2025, with leverage from U.S. manufacturing expected in 2026.

Overall, Teradyne’s performance demonstrates the growing weight of AI in its business model, with traditional end-markets now secondary to compute-driven cycles.

Executive Commentary

Strategic Positioning

1. AI Compute as Primary Growth Engine

AI compute has overtaken mobile as the chief determinant of Teradyne’s revenue cadence. SOC and memory sales are now governed by hyperscaler, large-scale cloud provider, and accelerator program launches, with customer ramps driving lumpy but larger order patterns. The Ultraflex Plus platform is gaining share in new compute domains, and networking is an increasingly important part of compute revenue. Management expects AI compute and memory to represent the majority of semi-test revenue in H2 2025.

2. Memory Market Transformation via HBM4

HBM4 test insertions are expanding Teradyne’s memory addressable market. The company secured a key post-stack singulated die win in Q2, signaling a structural shift as suppliers add more test steps to improve AI device quality. While HBM4 is not yet pervasive, management sees it as an emerging TAM, total addressable market, driver, with most 2025 memory demand tied to HBM4 capacity adds for AI accelerators.

3. Robotics Realignment and U.S. Manufacturing

Robotics is being repositioned for large customer wins and operational leverage. The consolidation of sales, marketing, and service across Universal Robots (UR) and MIR is designed to serve global customers more effectively, with a U.S. manufacturing facility planned to enhance supply chain resilience. A “plan of record” win with a major U.S. customer is expected to be a significant growth driver in 2026, though not material in 2025. Internal manufacturing will provide margin leverage as volumes recover.

4. Diversification Across Product Test and IST

Growth in product test and IST is supported by new applications in HDD, mobile SLT, and AI accelerator test. The Quantify Photonics acquisition accelerates Teradyne’s leadership in silicon photonics, optical data transmission, a critical enabler for next-gen AI compute. These segments provide balance against the cyclicality of semi-test and robotics.

5. Supply Chain Resilience as a Differentiator

Dual-sourcing and geographic diversification are now prerequisites for major customer wins, especially in robotics and compute. Teradyne’s investments in U.S. manufacturing and supply chain flexibility are directly aligned with hyperscaler and industrial customer requirements for operational continuity.

Key Considerations

This quarter marks a clear pivot in Teradyne’s business model and strategic priorities, with AI compute now the dominant force shaping both near- and long-term outcomes. Investors should track:

• AI Compute Program Ramps: Revenue timing is now dictated by hyperscaler and accelerator launches, introducing lumpiness but higher average order values.
• Memory Test Insertions: The spread of HBM4 and new test steps could expand Teradyne’s memory TAM and drive incremental growth in 2025 and beyond.
• Robotics Leverage Timeline: U.S. manufacturing and large-customer wins will not materially impact results until 2026, but set up operating leverage as volumes recover.
• Mobile Segment Inflection: Two nanometer and advanced packaging may drive a mobile rebound in 2026, but mobile will be a smaller share of the mix as compute grows.
• Margin Structure and Mix: Gross margins are holding, but product mix and utilization rates will be key as AI compute and memory become even larger contributors.

Risks

Visibility remains limited around the precise timing of AI compute ramps, with order lumpiness tied to customer program launches and test capacity cycles. Robotics recovery is contingent on large-customer adoption and successful U.S. manufacturing ramp. Competitive intensity in China, especially in industrial and robotics, poses ongoing share risk. New tax legislation will raise the effective tax rate in Q3, impacting net earnings.

Forward Outlook

For Q3 2025, Teradyne guided to:

• Sales of $710 million to $770 million
• Non-GAAP gross margins of 56.5% to 57.5%
• Non-GAAP operating profit rate at 19.5% (at midpoint)
• Non-GAAP EPS of $0.69 to $0.87

Full-year guidance remains one quarter at a time, but management signaled:

• Stronger H2 driven by AI compute and memory ramps
• Robotics to remain below break-even in 2025, with leverage expected in 2026

Management cited improved visibility and demand materializing into orders, but flagged ongoing uncertainty around the timing of customer shipments between Q3, Q4, and Q1 2026.

Takeaways

Investors should recognize that Teradyne’s long-term thesis now rests squarely on AI compute and memory, with legacy seasonality and mobile cycles relegated to secondary roles.

• AI Compute as Core Value Driver: The business is now levered to hyperscaler and accelerator test cycles, with program wins and test insertions expanding the addressable market.
• Robotics and Product Test Provide Optionality: While near-term contribution is muted, operational realignment and U.S. manufacturing set up 2026 for a margin and revenue inflection.
• Timing Volatility Remains: Investors should expect continued lumpiness in quarterly results, but the multi-year growth trajectory is increasingly underpinned by structural AI adoption.

Conclusion

Teradyne’s Q2 marked a decisive shift toward AI compute as the anchor of its business model, with program-driven revenue cycles replacing legacy seasonality. The setup for H2 and 2026 is stronger, but execution on new customer wins and robotics leverage will be critical for sustained outperformance.

Industry Read-Through

Teradyne’s results signal a broader industry pivot, as AI compute and HBM memory drive outsized capital spending in test and validation equipment. The shift away from consumer-driven cycles toward hyperscaler and accelerator launches is likely to reverberate across the semiconductor capital equipment sector. Robotics and manufacturing automation remain long-term secular opportunities, but near-term growth is gated by large-customer adoption and supply chain resilience. Competitors in test, robotics, and photonics should expect heightened demand volatility and increased customer requirements for geographic and operational flexibility.

TER Technology Semiconductor Equipment Artificial Intelligence Robotics Memory Test Equipment Electrification