inTEST’s EPS Beat Is Not the Story, the Order Mix Is

The cleanest read on the quarter is that the stock should not trade simply on the revenue miss. What was priced in was a weak top line: the Street sat at $29.8 million, and inTEST delivered $28.1 million for a -5.7% surprise. What was not priced in was the profit outcome: EPS came in at $0.03 against -$0.04, a swing from an expected loss to a profit. The variant perception is that Q2 did not show a business missing a cyclical recovery; it showed a company with enough cost takeout and mix resilience to earn through a still-subscale revenue base. That matters because the revenue history has been volatile without establishing a new trough, while gross margin has already stopped deteriorating at the point where orders are being rebuilt outside the weakest semi buckets.

That distinction matters because the quarter’s revenue was not a disguised recovery in the core headline number. Revenue of $28.1 million was still below the Street’s $29.8 million, and the company’s own reported basis was framed plainly by CFO Duncan Gilmour: “As Nick noted, revenue for the second quarter was $28.1 million.” The surprise was therefore not demand magnitude; it was operating leverage at a revenue level the market thought would lose money. Gross margin at 42.6% improved 110 basis points sequentially, and operating expenses fell by $1 million sequentially. Those two figures are the entire earnings argument: inTEST did not need a revenue beat to clear the EPS bar, which makes the next quarter less binary on top-line upside than the estimate miss implies.

The financial trajectory supports that interpretation because the company is no longer in the Q4 FY2024 spike-and-collapse pattern. Revenue fell from the $36.6 million high to $26.6 million in Q1 FY2025, then recovered to $28.1 million in Q2 FY2025; this is still a reset, not acceleration. The more investable fact is that gross margin moved from 41.5% to 42.6% while revenue was only up +5.6% sequentially. That is not enough to call a demand inflection, but it is enough to say cost actions are visible in the P&L. If a business can move margins higher on $28.1 million of sales after missing revenue by -5.7%, the downside case needs to shift from “loss-making at current volume” to “revenue shortfall can be absorbed if mix and opex hold.”

The order book is the reason not to dismiss the EPS beat as cost-only. Management said orders were nearly $28 million and grew 10% sequentially, with backlog at $37.9 million and essentially flat over the last 2 quarters. That is not a breakout, but it is stabilization at a level above the current quarterly revenue run-rate. The year-over-year backlog comparison is less flattering, down $9.8 million from the prior-year period, and management tied that comparison to the “large backlog we acquired with Alfamation.” The implication is important: the post-acquisition backlog air pocket is not getting worse, but the company has not yet proven it can rebuild the acquired book fast enough to regain prior revenue levels.

The mix inside orders is the real variant perception because it argues that the next leg of demand is not dependent on a clean semiconductor capex turn. Auto/EV demand increased 40% to $7.1 million, and management attributed a $2 million increase to Alfamation wins with key Tier 1 suppliers for OEM 2027 model year program starts. Those are unnamed customers, so there is no disclosed named customer read-through to quantify; the second-order point is still actionable. The Alfamation asset is being pulled into model-year programs rather than living off inherited backlog, and the magnitude is already visible in the quarter’s demand bridge. For auto test-equipment suppliers serving Tier 1 programs, inTEST’s $7.1 million auto/EV demand says procurement for 2027 platforms is active even while reported auto/EV sales remain depressed year over year.

The same order bridge also warns against over-reading semi strength. Sales to semi increased 13% to $10.2 million sequentially, but orders in semi declined $3.7 million. That conflict is the central hedge in the print: revenue recognized in semi improved, while forward demand in semi deteriorated. Investors looking for inTEST to be a pure beneficiary of the test-and-assembly upcycle should therefore be careful. The quarter’s better-than-expected earnings came with semi sales up, but the order momentum came from auto/EV, life sciences, industrial, and safety/security. The company’s internal demand mix is telling PMs to underwrite diversified test exposure, not a clean semi recovery trade.

That read is reinforced by the product-cycle evidence. New products generated sales of $6 million, just over 20% of total sales, which is a meaningful contribution for a company with $28.1 million of quarterly revenue. The qualitative claim here is not that new products are “working”; it is that they are already large enough to influence the quarterly P&L. When management gets asked about taking that to $10 million or $15 million in revenue, the right investor reaction is not to capitalize those numbers immediately, but to recognize the hurdle. The company has proven a $6 million quarterly contribution, while the next investor debate is whether that contribution can scale without the gross margin giving back the 110 basis points just regained.

The read-through to named customers and suppliers is unusually constrained because the data pack discloses no named customers of inTEST and no suppliers to inTEST. That absence itself limits second-order mapping: there is no basis to assign this quarter’s order gains to a specific OEM, semiconductor manufacturer, or supplier. The only customer-proxy detail is the unnamed Tier 1 supplier activity tied to OEM 2027 model year program starts, with a $2 million demand increase and auto/EV demand at $7.1 million. For portfolio work, that means the read-through is sectoral rather than company-specific: inTEST is showing measurable 2027 auto-program test demand, but the print does not justify attaching upside or downside to any named customer or supplier.

The peer context also argues against treating inTEST as a broad test-equipment recovery proxy. In the subsector table, several peers are showing positive top-line momentum, including ATEYY at +43.8% revenue YoY and 6871.T at +48.3% revenue YoY, while inTEST’s reported Q2 FY2025 revenue was down -17.2% year over year. That is not a small timing difference; it means inTEST is lagging the larger test-and-assembly recovery visible elsewhere. The offset is that inTEST’s 42.6% gross margin sits near 7729.T at 42.4%, not at the low end of the peer set. The investable conclusion is asymmetric: revenue recovery is not yet confirmed, but profitability is not broken relative to mid-margin peers.

The tone of the call was more constructive than the revenue miss, and this matters because management’s language matched the cost and backlog data rather than trying to sell a false acceleration. The tone history shows Q2 FY2025 sentiment at 0.26 and guidance_tone at 0.55, with uncertainty at 57.4. Those levels were materially cleaner than the immediately prior call, where sentiment was 0.09 and uncertainty was 92.7. This is the call-delivery version of the P&L: management sounded more confident because opex, gross margin, and backlog were no longer moving against them at the same time.

The guidance language is where the thesis becomes testable rather than impressionistic. Gilmour gave a tight frame for Q3: “For the third quarter, revenue is forecasted to be $28 million to $30 million with gross margins similar to Q2 2025 and operating expenses of $12.6 million to $13.1 million, excluding approximately $100,000 of restructuring expenses.” That quote earns the space because it commits management to three linked variables, not just a revenue range. The midpoint does not need to be spectacular to validate the thesis; what matters is whether gross margin can stay similar to 42.6% while opex remains around the guided range. If inTEST holds those together, the EPS beat should be treated as operating repair, not a one-quarter adjustment artifact.

The bear case is still real because the company has not shown sustained top-line growth. Compared with Q2 2024, revenue was down $5.9 million, driven by a $4.9 million decline in auto/EV sales, even as auto/EV demand rebounded in orders this quarter. That creates a timing gap: demand is improving in precisely the segment whose revenue decline explains most of the year-over-year weakness. The market may reasonably refuse to pay for that until orders convert. But the variant perception is that the market is too focused on the Q2 revenue miss and not focused enough on the fact that the auto/EV order recovery is tied to model-year programs, while backlog has stopped falling over the last 2 quarters.

What to watch next is therefore narrow and concrete. For Q3 FY2025, the thesis is confirmed if revenue lands inside the $28 million to $30 million guide, gross margins remain similar to Q2 2025, and operating expenses stay within $12.6 million to $13.1 million excluding approximately $100,000 of restructuring expenses. The order test is equally important: auto/EV demand needs to hold near the $7.1 million level or show that the $2 million Alfamation program-start contribution was not a one-quarter pull-forward. The thesis breaks if semi orders continue to decline after the $3.7 million drop while backlog falls below the $37.9 million level, because then the Q2 EPS beat would look like cost control against a shrinking opportunity rather than the first evidence of a repaired earnings base.