Teradyne heads into its third-quarter report with an unusual setup. Management told investors last quarter that revenue would step down from a record, and the stock has been punished anyway. Shares have fallen 11.7% since the July report while the S&P 500 gained 3.2%. That leaves the October 27 release, due after the close, to settle a narrower question than the headline numbers suggest. Is the dip really just compute order timing, as management insists, or the first sign that the AI test boom is cooling faster than promised?

The Street expects earnings of $2.04 per share on revenue of $1.27 billion. Those figures sit in the upper half of management's guidance of $1.85 to $2.15 per share and $1.20 billion to $1.30 billion in sales. When the guide was issued it landed well above the consensus of the time, $1.40 and $1.07 billion, and analysts have since moved up to meet it. The Earnings Whisper of $2.24 sits above the top of the company's range, which tells you the informal bar is higher than the published one. A result near the midpoint could therefore disappoint even if it technically meets guidance. On a year-over-year basis the comparisons are dramatic, with EPS expected to rise about 140% and revenue about 65%. But the more relevant yardstick is the sequential decline from the second quarter's $2.47 and $1.33 billion.

The prior call was among the most confident Teradyne has delivered. Management lifted its long-term view of the test equipment market to $20 billion or more by the end of the decade, well above the $12 billion to $14 billion framing it offered late last year. It also said 2026 memory demand now looks more than 40% larger than 2025, compared with the low-double-digit growth it had guided earlier. Memory revenue hit a record $212 million, its third straight quarter above $200 million, with book-to-bill above 2x. That backlog is the clearest thing to check this quarter. If memory holds above $200 million and orders stay healthy, the claim that growth is broadening beyond compute gains real support. A slip would undercut the story that other businesses can carry the company through the compute lull.

The second validation point is the shape of the year. Management narrowed its first-half revenue weighting to 50% to 52% of the annual total, down from roughly 60% earlier in the year. In plain terms, it expects the back half to be nearly as large as the front. Fourth-quarter commentary needs to back that up. The next compute surge is not expected until the first half of 2027, so any sign that it is slipping further would be damaging. So would evidence that the merchant GPU win and the second hyperscaler correlation are not turning into orders. Growth in integrated systems test, which jumped about 2.5x sequentially on hard drive and AI demand, along with robotics and product test, also needs to continue.

Margins are the quieter swing factor. Gross margin was guided to 58% to 59%, down from 59.8% in the second quarter, reflecting a lighter compute mix and new product launch costs. Landing at or above the top of that range would suggest mix is holding up better than feared. Investors should also listen for early framing of the new target earnings model management promised for the fourth-quarter call, and for any update on the risk that test spending as a share of chip equipment spending drifts back toward 7% in 2027.

Sentiment has swung from solidly bullish before the last report to slightly bearish now. The stock trades at $325, below its 200-day moving average near $344 and just above its post-earnings low of $319. Expectations in the share price look subdued even as the numbers bar remains high. The report hinges on whether Teradyne can show that memory, systems test and second-half bookings are robust enough to make the compute pause look like a timing gap rather than a turning point.