Teradyne’s Q2 was not the beat, it was the trough signal the market is underpricing

The print says the debate should shift away from whether Q2 cleared estimates and toward whether the June quarter marked the earnings trough. What was priced in was a modest beat against a reset bar: EPS of $0.57 versus the Street’s $0.54, a +5.2% surprise, and revenue of $651.8 million versus $650.6 million, only a +0.2% surprise. What actually surprised was not the magnitude of the beat, because revenue was essentially on the number, but the shape of the forward setup: management guided Q3 sales to $710 million to $770 million and Q3 non-GAAP EPS to $0.69 to $0.87. That range matters because it reframes Q2’s weakness as a mix and timing issue rather than a demand break, even though the reported Q2 revenue line was down -10.7% year-over-year and gross margin fell to 57.2%. The variant perception is that investors focused on the thin revenue beat and the margin giveback are underweighting the evidence that Teradyne is spending into a customer-led test upcycle before that upcycle is visible in trailing gross margin.

The reason that distinction matters is that Q2 itself was not an especially clean quarter. On the Street-comparison basis, revenue missed the spirit of a beat even while technically beating by +0.2%, and EPS did the real work with a +5.2% surprise. On the company’s own account, Sanjay Mehta’s wording was deliberately narrow: “Second quarter sales were $652 million and non-GAAP EPS was $0.57, both above the midpoint of our guidance ranges.” The quote earns attention because it commits only to clearing the midpoint, not to broad-based end-market strength. That restraint is consistent with the reported trajectory: revenue has moved from a late-2024 high into a Q2 FY2025 low, and gross margin gave back the benefit seen earlier in FY2025. The market can read that as deterioration. The better read is that Teradyne is absorbing a transition quarter while keeping R&D and customer-specific capacity in place for the next leg of semiconductor test demand.

That financial trajectory explains why the Q3 guide is the center of the call. The company is asking investors to accept a near-term gross margin range of 56.5% to 57.5% while sales move to $710 million to $770 million, so the operating leverage is arriving first through revenue rather than through immediate margin expansion. That is not cost cutting dressed up as a cycle story. Non-GAAP operating expenses were $275 million, and management said those costs were up year-over-year because of R&D investment and targeted growth opportunities. At the midpoint of the Q3 guide, the non-GAAP operating profit rate is 19.5%, which is acceptable if the revenue ramp is early, but not if Q3 is a one-quarter air pocket recovery. The thesis therefore rests on a specific claim: this quarter supports buying the forward test cycle only if Teradyne’s guided rebound is tied to compute and VIP ASIC content, not to temporary product shipments.

The segment evidence supports that cycle interpretation, with one caveat. Semi test revenue was $492 million, and within that SOC revenue contributed $397 million. That concentration is the point, not a flaw, because the compute test debate is about SOC content and new device complexity. Memory was $61 million and IST was $34 million, so the quarter was not carried by a broad memory recovery. The caveat is that the non-semi businesses are not yet large enough to offset a semi test pause. Product test revenue was $85 million, up 7% year-over-year, and robotics revenue was $75 million, up quarter-over-quarter but down year-over-year. Those numbers keep Teradyne from being a pure compute test instrument, but they do not change the equity argument: if SOC-driven semi test accelerates, the earnings sensitivity sits in the largest revenue pool.

That same mix is why the ASIC comments matter more than the headline beat. The questions on the call anchored investor attention to the VIP ASIC compute opportunity, including a prior TAM framing of “$300 million going to $800 million” and a later reference to “$2.3 billion for the compute market, $400 million for VIP.” Those are not management guidance figures in this data pack, so they should not be treated as forecasts. They do, however, reveal what the Street is trying to underwrite: whether Teradyne can attach to custom compute silicon test at scale and whether a potential “50% share” of that VIP SAM is plausible. Against that debate, Q2’s SOC revenue of $397 million is the better anchor than the total company revenue beat of +0.2%. The market was priced for a small beat and a cautious guide; what it got was a quarter where the largest semi test category is already the dominant revenue component and management guided sales materially above the just-reported quarter.

The margin debate is more nuanced, because the company did not guide a clean snapback in profitability. Non-GAAP gross margins were 57.3%, and Q3 gross margins are expected at 56.5% to 57.5%, so mix and investment are still weighing on the model. A bearish reading is that revenue growth is coming with lower gross margin quality. The more useful reading is that management is choosing not to starve the model ahead of a demand upturn. Q3 OpEx is expected at 36.5% to 38.5% of sales, and the Q3 non-GAAP operating profit rate at the midpoint is 19.5%. That combination says operating leverage is real but not yet normalized. The stock should not be paid for peak margins on this print; it should be paid for the probability that revenue scale returns before investors regain confidence in gross margin.

Cash return strengthens that argument because it reduces the risk that R&D spending is masking weak cash conversion. Free cash flow was $132 million in the quarter, and Teradyne repurchased $117 million of shares while paying $19 million in dividends. Management also returned $316 million in the first half, or 138% of free cash flow, and ended the quarter with $489 million in cash and marketable securities. The capital return is not the core thesis, but it matters because it shows management is funding both R&D and shareholder returns through the trough. If the company were stretching the balance sheet to defend EPS, the EPS beat would deserve a lower multiple. Instead, the free cash flow and cash balance support the view that Q2 was a trough setup rather than an earnings-quality red flag.

The supply-chain read-through is most direct for customers buying advanced test capacity and for suppliers tied to sockets and probes. For TSMC, Samsung, ASE Group, Intel, and Amkor, Teradyne’s $492 million of semi test revenue and $397 million of SOC revenue point to sustained demand for UltraFlex SoC and related ATE capacity even before memory becomes the main contributor. That is particularly relevant for outsourced assembly and test exposure at ASE Group and Amkor, because the Teradyne print suggests SOC test complexity is absorbing tester dollars while memory sits at $61 million. On the supplier side, Leeno Industrial and Yamaichi Electronics should see the better read-through from tester utilization and customer-specific fixtures than from Teradyne’s consolidated revenue beat, since Q2 revenue surprised by only +0.2% but the Q3 sales guide moved to $710 million to $770 million. The implication is not broad capex exuberance; it is a targeted recovery in SOC and custom-compute test sockets, probes, and handlers around the customers already in Teradyne’s installed base.

The peer context also argues against treating Teradyne as a generic laggard in test and assembly. In the latest peer table, the strongest revenue growth sits at +48.3% for 6871.T and +43.8% for ATEYY, while Yamaichi Electronics shows +43.0% revenue YoY and 38.0% gross margin. Teradyne’s Q2 revenue was down -10.7% year-over-year with 57.2% gross margin, so it is not winning the backward-looking growth screen. That is precisely where the mispricing can sit: peers with visible current growth have already shown the cycle, while Teradyne’s guide implies the inflection is still moving into the reported numbers. The comparative point is not that Teradyne’s Q2 was better than peers. It is that Teradyne’s gross margin remains structurally above socket-supplier economics, while its revenue line is closer to a trough than to a fully expressed recovery.

The call delivery supports the same interpretation, but it is not uniformly bullish. The tone history shows Q2 FY2025 guidance_tone at 0.43, up from 0.10 in Q1 FY2025, and uncertainty at 46.6 versus 71.5. That is the strongest evidence that management’s confidence improved at the same time the reported quarter looked optically soft. Prepared_sentiment also moved to 0.51, while qa_sentiment was only 0.20, which says the scripted message was cleaner than the investor interrogation. The conflict is worth taking seriously: management could articulate the guide, but the Q&A did not fully eliminate skepticism around customer timing and compute-market sizing. Still, the decline in uncertainty is the better signal for this event because it aligns with a numerical guide for higher Q3 revenue rather than with vague optimism.

That tone shift also changes how to interpret the R&D and OpEx burden. Gregory Stephen Smith said the new organization “delivered 9% quarter-on-quarter growth despite persistent difficult market conditions,” language that matters because it acknowledges the market is still difficult while pointing to organization-specific improvement. Sanjay Mehta’s OpEx framing is more important for the model: “Non-GAAP operating expenses were $275 million, up year-over-year as we have increased our R&D investments and targeted opportunities to drive longer-term growth.” The wording commits to spending for identifiable opportunities, not simply higher overhead. If the Q3 sales guide is achieved, that spending will look like capacity preparation. If Q3 lands near the low end and gross margin stays near the bottom of 56.5% to 57.5%, investors should treat the same OpEx as evidence that the compute ramp is slower than the company is funding for.

The risk to the thesis is not hidden in the EPS line; it is in the possibility that Q3 revenue growth arrives without the mix that investors want. Q2 GAAP diluted EPS in the history table was $0.49, while the Street-comparison and company non-GAAP EPS figure was $0.57, so basis discipline matters. The beat was on the non-GAAP street basis, and the margin discussion should stay anchored to non-GAAP gross margin commentary and the Q3 non-GAAP operating profit rate. If investors blur those bases, they may either overstate the quarter’s profitability or dismiss the EPS beat as accounting noise. The cleaner conclusion is that Q2’s reported profitability was not the buy case. The buy case is that the Q3 guide, semi test mix, and lower uncertainty together signal an inflection before trailing earnings make it obvious.

What to watch next is therefore narrow and measurable. For the thesis to hold, Q3 sales need to land inside the guided $710 million to $770 million range, and the market should require evidence that SOC remains the engine rather than revenue being filled by lower-quality mix. Gross margin should stay within the guided 56.5% to 57.5% band; a result below that range would break the argument that current spending is controlled ahead of scale. EPS should track the $0.69 to $0.87 non-GAAP guide on 158 million diluted shares, because failure there would mean OpEx is outrunning the recovery. On the next call, the confirmatory signal is not a bigger adjective from management, but guidance_tone staying near Q2 FY2025’s 0.43 while uncertainty does not revert toward Q1 FY2025’s 71.5. If those levels hold and semi test remains anchored by SOC revenue near the Q2 base of $397 million, Q2 will look like the trough investors were supposed to buy, not the warning they were tempted to sell.