TER · Q3 2025 Earnings

Bullish

Teradyne

Reported October 29, 2025

30-second summary

Revenue of $769M landed at the top of the prior guide ($710M–$770M), up 17.9% QoQ and 4.3% YoY, with non-GAAP gross margin of 58.5% beating the 56.5%–57.5% guide. The signal is the Q4 guide: $920M–$1,000M (midpoint $960M), +25% sequential and +27% YoY, with non-GAAP operating margin guided to 25.5% at the midpoint — a meaningful step-up in profitability driven by AI compute and memory. Management explicitly upgraded the 2026 setup ("stronger today than it did six months ago") and reframed AI as the singular engine, not one of several.

Headline numbers

EPS

Q3 FY2025

$0.85

Revenue

Q3 FY2025

$0.77B

+4.3% YoY

Gross margin

Q3 FY2025

58.4%

Free cash flow

Q3 FY2025

$0.00B

Operating margin

Q3 FY2025

18.9%

Key financials

Q3 FY2025

Metric	Q3 FY2025	YoY	Q2 FY2025	QoQ
Revenue	$0.77B	+4.3%	$0.65B	+17.9%
EPS	$0.85	—	$0.57	+49.1%
Gross margin	58.4%	—	57.2%	+120bps
Operating margin	18.9%	—	13.9%	+500bps
Free cash flow	$0.00B	—	$0.13B	-98.2%

Guidance

Q3 results beat revenue and gross margin guidance; Q4 forward guidance reflects robust 27% YoY growth and strong AI-driven demand momentum with material operating leverage expected.

Guidance is issued for both next quarter and the full year. Both may appear below.

Actuals vs prior guidance

Metric	Period	Prior guide	Actual	Δ	Result
Revenue	Q3 FY2025	$710M to $770M	$769M	Within guide range (midpoint $74M was $740M, actual at high end)	Beat
Non-GAAP EPS	Q3 FY2025	$0.69 to $0.87	$0.85	Within range, near high end	Beat
Gross Margin	Q3 FY2025	56.5% to 57.5%	58.4%	+0.9 to +1.9 pts above guide	Beat

New guidance

Metric	Period	Guide	YoY
Revenue	Q4 FY2025	$920M to $1,000M	27% YoY
Non-GAAP EPS	Q4 FY2025	$1.20 to $1.46	—
Gross Margin	Q4 FY2025	57% to 58%	—
Operating Margin (non-GAAP)	Q4 FY2025	25.5% (midpoint)	—
Operating Expense as % of Sales	Q4 FY2025	31% to 33%	—

Segment performance

Q3 FY2025

Segment	Q3 FY2025
Semiconductor Test revenue	$606M

Capacity & utilization

Q3 FY2025

Segment	Q3 FY2025
Q4 2025 sequential revenue growth guidance	25% (midpoint at $960M)
Q4 2025 YoY revenue growth guidance	27% (vs Q4 2024)
AI-related SOC test demand	Strong across compute, networking, memory

Profitability

Q3 FY2025

Segment	Q3 FY2025
Operating margin - non-GAAP	20.4%
Gross margin - GAAP	58.4%

Management tone

Narrative arc: Q2 "AI compute is the dominant H2 driver" → Q3 "AI is the singular engine; consumer/mobile explicitly weakening."

The AI framing escalated from "majority of H2 semi-test" to "primary engine of growth," with consumer/mobile actively deprioritized in the same breath. Last quarter management was still hedging — AI compute as "majority" of H2 but mobile turning a corner. This quarter the language hardened: "our semi-test business has evolved to where the largest demand driver is AI data center investments rather than consumer end markets." That is a structural reclassification, not a mix update — Teradyne is now positioning itself as an AI infrastructure name with a mobile option, rather than a balanced test vendor.

Project timing pulled forward versus the July call. In Q2, management described projects "straddling Q3/Q4 or Q4/Q1." This quarter Sanjay confirmed Q3 captured the early ramp and Q4 demand is "ramping significantly," requiring expedited supply chain and accelerated production capacity additions across multiple geographies. That is operational urgency — capex and one-time costs being absorbed to chase demand — and it is the strongest possible read-through on near-term order velocity.

The 2026 setup was explicitly upgraded. Management stated "2026 looks stronger today than it did six months ago, and all indications suggest solid growth from 2025." In Q2 the company was "significantly more confident than 90 days ago" on H2; this quarter the confidence window has extended into next year. The January update will reweight the 2028 model materially toward compute and AI-driven memory.

One discipline note worth flagging. Management explicitly excluded new design wins (merchant GPU, VIP customers, SLP penetration) from current numbers: "our Q3 results and our Q4 guidance do not include any revenue from these types of new opportunities." That is a deliberate signal — guidance is conservative on the cyclical, with the strategic optionality kept off-balance-sheet for now.

Caveat that didn't go away: memory and AI shipments remain "lumpy." Management is bullish on 2026 but reiterated that quarterly results can swing on individual project timing in a concentrated customer base.

Recurring themes management leaned on this quarter:

AI data center investments as singular demand driver for semiconductor testMemory strength in HBM/DRAM/AI-related LPDDR offsetting overall memory TAM weaknessProject acceleration into Q3/Q4 creating supply chain and manufacturing capacity pressureRobotics stabilizing but indirect distribution channel remaining challengedSignificant R&D and go-to-market investment for 2026 AI growth with OpEx discipline2026 outlook strengthened versus six months prior

Risks management surfaced:

Timing of individual projects can swing quarterly results significantly due to highly concentrated and dynamic AI marketMobile and auto-industrial conditions 'somewhat weak' with uncertain timing and intensity of recoverySupply chain and manufacturing capacity constraints requiring expedited actions and one-time costsWorking capital increases in inventory tied to compute/memory ramp reducing free cash flowRobotics core indirect distribution channel experiencing persistent weakness

Q&A highlights

CJ Muse · Cantor Fitzgerald

Breakdown of Q4 upside versus consensus: what portion is from HBM, VIP, networking, SLT versus other segments? Long-term question on compute intensity, test insertions, and test time for high-performance compute.

Q4 upside is all in compute and memory: two-thirds compute, one-third memory, with HBM strongly represented in memory. Management is bullish on test intensity growth due to larger die sizes, chiplet-based designs increasing scrap costs (driving shift-left testing), and data center tolerance for defects being extremely low. Dual sourcing strategies becoming important in compute, benefiting Teradyne from lower market share position.

Q4 upside: two-thirds compute, one-third memoryHBM strongly represented in memory upsideTest intensity expected to grow as die sizes increase and performance requirements escalateChiplet-based designs escalating cost of scrap, driving upstream test insertion

Timothy Arcuri · UBS

Confirms the $200M upside in Semitest guidance composition and asks about revenue shaping in 2026, particularly Q1 seasonality given lumpy memory demand.

Confirms composition is two-thirds compute/networking, one-third memory. Revenue expected to be up in 2026 relative to 2025 driven by compute and memory. Projects accelerating from Q4 into Q1. Business seasonality fundamentally changed from historical mobile-driven (Q2/Q3 peaks) to compute-driven and lumpy based on customer launches. Will provide detailed update in January.

Q4 upside composition: two-thirds compute/networking, one-third memory2026 revenue expected up versus 2025Projects straddling Q4/Q1 with acceleration into Q4Business no longer mobile-driven; now compute-project-driven with different seasonality

Samik Chatterjee · JP Morgan

How much of memory increase is from final test vs. wafer sort? Update on 2028 composition model (previously third/third/third split across mobility, compute, auto-industrial)?

Memory split into 2x2: DRAM vs. flash, wafer sort vs. final test. Teradyne has high share in DRAM final test and HBM performance test, but significantly lower share in wafer sort (ex-HBM). 2026 will see expansion in wafer sort TAM, so revenue growth expected but not necessarily share gains. 2028 model composition will be updated in January given AI/compute becoming much heavier weighted.

High share in DRAM final test and HBM performance testSignificantly lower share in wafer sort (excluding HBM performance)2026 memory expansion expected; wafer sort TAM expanding significantlyQ3 AI-driven revenue: 50% of total; Q4: 60% of total

Medhi Hosini · SIG

How will design wins in wafer-level test (SLP) manifest into penetration increases? Will burning at wafer level be part of design wins?

SLP (singulated die test) is critical for ensuring no latent defects into data centers. New technologies like COWOP enable complex modules requiring extensive system test and burn-in. Management sees burn-in and SLP as contiguous market. Design wins have not been embedded in current guidance; long-term model will be updated in January showing different composition of business.

SLP critical for latent defect prevention in data centersCOWOP technology driving new burn-in and system test requirementsDesign wins not yet embedded in current guidanceLong-term 2028 target (7-9.50 EPS) destination unchanged but composition will differ materially

Jim Schneider · Goldman Sachs

Expectations for mobile SOC in 2026, including N2 complexity and WMCM packaging impacts. OPEX leverage expectations in 2026.

Mobile SOC expected to be bigger than 2025 but magnitude uncertain. Three TAM factors: (1) complexity—N2 and WMCM expected to drive double-digit test intensity increases; (2) yield expected strong per historical execution; (3) unit volume is the key X-factor. Mobile TAM highly dependent on handset refresh inflection. OPEX leverage target: ~$0.50 of OPEX growth per dollar of revenue growth in 2026.

Mobile SOC TAM uncertain but expected to be up YoYN2 and WMCM expected to drive double-digit test intensity increasesUnit volume is critical X-factor for mobile TAM size in 2026If handset sales inflect, strong year; if not, modest improvement

Answers to last quarter's watch list

Does Q3 land at or above the $740M midpoint, and does Q4 step up materially from Q3? Q3 revenue of $769M came in $29M above the midpoint at the top of the range. Q4 guide midpoint of $960M is a 25% sequential step-up — materially exceeding the implicit "modest step-up" framing from last quarter.

Resolved positively

Memory snapback magnitude in Q3. Memory contributed $128M (+110% QoQ) and is one-third of the Q4 upside, with HBM heavily represented. Memory revenue expected to sustain at 2024 levels despite TAM down low double-digits — share/mix gain confirmed.

Resolved positively

First evidence of merchant-GPU customer wins. Management referenced progress on new GPU/VIP opportunities and "cautious optimism," but explicitly excluded any of this revenue from Q3 actuals or Q4 guide. No named customer disclosure.

Continue monitoring

Mobile system sales mix. No quantified update on system sales versus idle-capacity upgrades; management instead characterized mobile/auto-industrial as "somewhat weak" with uncertain recovery timing. Mobile is now framed as a 2026 swing factor on handset unit volume, not a near-term driver.

Not resolved

Robotics revenue trajectory and large-customer concentration. Robotics held at $75M sequentially — flat off the Q1 trough. Service and AI products growing but core indirect channel "persistently weak." The "2026 growth driver" framing is intact but unproven; the flat print is neither validation nor refutation.

Continue monitoring

Gross margin durability at 57%+. Q3 non-GAAP gross margin of 58.5% beat the high end of the guide by 100bps. Q4 guide of 57%–58% holds the level.

Resolved positively

What to watch into next quarter

Does Q4 land at or above the $960M midpoint, and what does the January 2026 update imply for Q1 and full-year 2026? With management already signaling 2026 up versus 2025, anything other than a confident framework in January would undercut the conviction-upgrade narrative.

Q4 non-GAAP operating margin delivery against the 25.5% midpoint. This is the highest-conviction operating leverage signal Teradyne has put on the table — actual realization (versus mix or one-time cost slippage) validates the OpEx leverage thesis for 2026 ($0.50 OpEx growth per $1.00 revenue growth).

First disclosed merchant-GPU or new VIP customer revenue. Management has now explicitly excluded this twice. The first time a named win or quantified revenue contribution lands in guidance is a re-rating event for the AI TAM.

2028 model reweighting in January. The composition shift away from third/third/third toward compute/AI-memory weighting will reset how the long-term $7.00–$9.50 EPS target is underwritten — watch for the implied compute revenue share and memory mix.

Robotics — sequential trajectory above $75M and any named large-customer wins. Three quarters of flat-to-down revenue with "persistent weakness in indirect channels" is approaching the point where the 2026 turnaround framing needs a real data point.

Memory TAM read-through into 2026. Management expects memory revenue flat to 2024 despite a down-low-double-digit TAM — sustaining this in 2026 requires HBM/DRAM share and wafer sort TAM expansion to materialize as described.

Sources

Teradyne Q3 2025 earnings press release (SEC EDGAR): https://www.sec.gov/Archives/edgar/data/97210/000119312525254602/ter-ex99_1.htm
Teradyne Q3 2025 earnings call commentary (management prepared remarks and Q&A).

Get the next brief, free.

We publish analyst-grade earnings briefs the same day or morning after every call — headline numbers, segment KPIs, Q&A highlights, and tone analysis. Free during beta.

This is not investment advice.