Oct 28, 2026 earnings call — AI-driven wafer-to-data-center growth
Total revenue topped $1.3 billion this quarter, with non GAAP EPS of $2.29.
Revenue $1.3B (est $1.3B) · EPS $2.29 (est $2.09) · Next-quarter revenue outlook $1.3B–$1.4B
Key quotes
"AI was the driver again this quarter. Total company revenue topped $1.3 billion up 70% year over year and consistent quarter over quarter with non GAAP EPS of $2.29 up over 150% year over year."
"The first is the UltraPin 5 thousandXM. The eM stands for extended memory."
"In the past 2 quarters, it has become clear that the constraints on memory supply are so severe that AI accelerators are turning to lower stack heights of HBM memory to get a larger number of stacks out of the same volume of DRAM wafers."
Analyst Q&A 7
memory demand and gross margin
Q (CJ Muse · Cantor Fitzgerald): How should we think about memory demand into Q4 and 2027, and which memory segment will drive growth? Also, is product test driving gross margin changes and what is normalized gross margin next year?
A: Management said memory should stay very strong in Q4 and they expect significant year-on-year growth in 2027 driven mainly by HBM. They also said overall gross margins improved year over year and that product test is accretive to the portfolio, while opposing headwinds and tailwinds will affect 2027 margins.
"So as we look at second half of next year, we are continuing to see robust demand signals to be extent that we are continuing to build on our capacity." — Michelle L. Turner (CFO)
inventory build and timing of next year
Q (Timothy Arcuri · UBS): Why did inventory rise about $200 million in 6 months and does that imply next year will be back-end weighted or ratable?
A: Management said two things drove inventory: preparing for robust 2027 growth and longer lead times for components so they must order earlier. They said ASIC programs look more back-half loaded, expect a strong year overall, and will give more phasing detail in Q4.
"1 is you are right, we are preparing for robust growth in 2027." — Gregory S. Smith (CEO)
updated growth model and demand drivers
Q (Mehdi Hosseini · SIG): Can you update the demand vectors and how your prior target model should change? Also, what needs to happen for robotics/physical AI to grow faster?
A: Management said they will update the model in January and that the semiconductor TAM has grown quickly, changing mix versus the old model; compute is growing faster and they are gaining modest share, mobile is underperforming, and product test and robotics track expectations. For physical AI, they said partnerships and working with labs and CMs are in progress and expect transactions in 2027 but broader end-market growth likely in 2028.
"compute is growing faster than we anticipated, and we are gaining share there at a modest rate." — Gregory S. Smith (CEO)
near-term visibility and robotics visibility
Q (James Schneider · Goldman Sachs): What is your visibility into the next several months and directionally for Q1? And how does robotics visibility compare?
A: Management said they now have a couple quarters of line of sight and exit 2026 with a strong backlog, so they expect Q1 to have some growth and expect the year to grow quarter-on-quarter in 2027. They added robotics visibility is short term except for their largest customer, where deployments are moving from engineering to broader installs and may drive back-half weighting.
"So where we are sitting as we exit 2026 is really strong backlog." — Michelle L. Turner (CFO)
quarterly growth pattern and HBM stack heights
Q (Vivek Arya · BofA): Could Q1 be below Q4 as it was some prior years, and what is the impact of lower HBM stack heights on testing opportunity and your share?
A: Management clarified they expect each quarter in 2027 to grow versus 2026 and noted increased visibility reduces reliance on old seasonal patterns; they cautioned there can still be quarter-to-quarter lumpiness. On HBM stack heights, they said the trend toward higher stacks has reversed in 2026 and lower stack heights free DRAM wafer supply to create more stacks to test, which is positive for their share, but they could not give hard numbers.
"So we expect the quarters next year to represent growth quarter on quarter when you look at next year." — Michelle L. Turner (CFO)
HBM pin speed and competition
Q (Krish Sankar · TD Cowen): Would lowering HBM signal or pin speed change test complexity? And how is competition from Korean players shaping up?
A: Management did not provide a direct answer to these specific technical and competitor questions in the transcript segment provided.
inventory build: demand vs constrained parts
Q (Krish Sankar · TD Cowen): How much of the inventory build is for true demand versus stocking constrained components?
A: Management said the inventory build is about 50-50 between preparing for demand and stocking constrained components, and that part of the increase normalizes unusually low inventory levels from Q1; they think days-inventory-out around 200 is more normal for them.
"I would argue the 189 days that we are at, I still think that we should probably be around 200 when you think about kind of normal levels in comparison to some of our peers." — Michelle L. Turner (CFO)