They raised the full-year outlook while guiding Q3 non-GAAP EPS to nearly a dollar below Q2's print.
Thesis: Test is the one part of the AI capital cycle where every wafer fabbed converts to revenue with no supply ceiling on Teradyne's own output, and the second-hyperscaler correlation completion converts a single-threaded compute book into multi-threaded share gains starting H1 2027. The Q3 guide-down is program timing (merchant GPU shipped, next surge in H1 2027), not demand loss, and memory book-to-bill above 2 with customers planning capacity into 2027 confirms backlog quality.
Verdict: LONG — Conviction: MEDIUM
Catalyst: Production ramp at the second hyperscaler beginning H1 2027 after the nine-to-twelve-month dual-vendor timeline from correlation, plus memory manufacturers executing 2027 capacity add plans.
Key Risk: Test's share of semi-cap reverts from 8% to 6-7% in 2027 on the WFE-to-ATE time lag, which the CEO explicitly flagged, while memory mix drags gross margin through 2027.
The Tell: Shane Brett showed that management's own slide has memory test growing at roughly half of bit growth, contradicting the 70% memory growth narrative. Greg Smith answered: 'I'm going to need to take that question offline. You're interpreting a lot of information from that chart that I'm not sure we were intending to communicate.' He then conceded part of the market does follow bit growth, undercutting the extrapolation.
Friction Level: MODERATE_FRICTION — Both sides accept the numbers: record $1.3B revenue, $378M FCF, memory B2B over 2. Disagreement is whether Q3's compute decline and 130bp margin headwind are a cycle peak or a timing gap before the H1 2027 second-hyperscaler production ramp.
Report not found
The report data is no longer available. Please return to the archive.