Arcuri ran the math on the call: the Q1 guide implies the business outside Data Center shrinks 3-4% sequentially.
Thesis: Execution is real: 5 straight beats, GM +140bps while absorbing a 200bps gold headwind and annual price declines, OCF $36M against $32M net income, DSO and inventory days both down, $60M voluntary debt paydown. But the growth engine is 14% of revenue decelerating from 40% to 20-25% sequential inside the guide, and the 71% auto base prints flat with thin inventory and zero restocking cushion. Content per rack scaling $150 to $425 is gated on 800V adoption timing that management admits is correlated to hyperscaler schedules they do not control.
Verdict: HOLD — Conviction: MEDIUM
Catalyst: Q1 print showing DC at 16-17% with auto up a couple points and GM at 50-51% would confirm the mix shift is additive rather than substitutive. Gold-to-copper conversions landing through customer qualifications convert the identified 200bps tailwind into printed margin.
Key Risk: A hyperscaler CapEx pause removes the only growth engine while SAAR softness hits the 71% base simultaneously. Management itself frames 800V ramp speed as correlated to deployment schedules outside their control.
The Tell: D'Antilio volunteered unprompted: "we have a little bit of delinquency that we're starting to build in certain pockets in both auto and in Industrial." He raised it while citing multi-year high backlog as a positive. Customers paying late against record orders is the one crack in an otherwise clean narrative.
Friction Level: MODERATE_FRICTION — Both sides accept the numbers: DC at 14%, auto flat, margins expanding, backlog at a multi-year high. Disagreement is on implication. Bull reads the DC mix shift as repricing fuel; bear reads it as a masking of a decelerating core where 71% of revenue has no restocking catalyst.
2026Q4 · May 7, 2026
3 of 5 readings agreed3 HOLD · 2 LONGRead the full call
Data Center Decelerating in the Guide, Auto Flat at 71% of Sales
Execution is real: 5 straight beats, GM +140bps while absorbing a 200bps gold headwind and annual price declines, OCF $36M against $32M net income, DSO and inventory days both down, $60M voluntary debt paydown. But the growth engine is 14% of revenue decelerating from 40% to 20-25% sequential inside the guide, and the 71% auto base prints flat with thin inventory and zero restocking cushion. Content per rack scaling $150 to $425 is gated on 800V adoption timing that management admits is correlated to hyperscaler schedules they do not control.
D'Antilio volunteered unprompted: "we have a little bit of delinquency that we're starting to build in certain pockets in both auto and in Industrial." He raised it while citing multi-year high backlog as a positive. Customers paying late against record orders is the one crack in an otherwise clean narrative.
Fifth consecutive sales growth quarter. Q4 $243M +26% YoY, EPS $0.17 up 183%. Q1 guides $245-255M and $0.19-0.23 EPS, sequential EPS growth through the seasonally soft March quarter with 70% drop-through and opex down to $80M.
Backlog at a multi-year high, design wins up more than 30% YoY led by Focus Auto and Data Center, and D'Antilio adding back-end capacity in the Philippines specifically to clear delinquencies. Revenue is being pulled forward by real orders, not channel loading.
Multi-year high backlog with no cancellation terms, mix breakdown, or duration disclosed. D'Antilio volunteered that delinquency is building in pockets of both auto and Industrial. Record backlog and customers paying late in the same paragraph.
FY26 GM +140bps despite annual price declines and a 200bps gold headwind. Q1 guide 50-51% with ~70% drop-through already locked. Bridge to 55% is itemized: operating leverage, gold-to-copper conversion, factory efficiencies, select price increases starting end of Q1.
ASIL D passive TMR angle sensor for steer-by-wire carries 2x-3x content uplift vs conventional steering. TMR is 30% of FY26 sensor releases. Current sensors selected over local alternatives on performance and package size in Chinese robot wins. Design cycles lock them in.
Data Center quadrupled off a 2-3% base to 14% in two quarters and management guides 16-17% next quarter. D'Antilio's own math shows DC growth decelerating 40% to 20-25% sequential inside the guide while the rest of the business declines.
+62
CONFIDENT"Direct answers with exact figures: D'Antilio gave the DC math himself (14% to 16-17%, implying 20-25% growth). Doogue volunteered Asia trip specifics, current sensor mix (0% to 10% to 20% of DC in one year), and design win counts. One unprompted hedge on delinquency."
Full transcript, segment-level percentages, exact guide math, and an analyst who reconstructed the underlying trend on the call. Cash flow reconciles to net income. Gaps: no backlog cancellation terms, no DC backlog duration.
Q1 print showing DC at 16-17% with auto up a couple points and GM at 50-51% would confirm the mix shift is additive rather than substitutive. Gold-to-copper conversions landing through customer qualifications convert the identified 200bps tailwind into printed margin.
A hyperscaler CapEx pause removes the only growth engine while SAAR softness hits the 71% base simultaneously. Management itself frames 800V ramp speed as correlated to deployment schedules outside their control.
“Starting with our fourth quarter results. Sales were $243 million, and non-GAAP earnings per share were $0.17.”
“We expect first-quarter sales to be in the range of $245 million-$255 million.”
Figures are extracted from the transcripts and checked in code: every quote appears word for word. Descriptive only: none of this feeds the verdict.
This is operating leverage printing in real time. Q4 $243M +26% YoY, op margin 15.6% up 660bps, and FY26 margins expanded 140bps while eating 200bps of gold and annual price declines. Q1 guides 70% drop-through with opex falling to $80M, so the margin step is locked before the quarter starts. Data Center hit 14% of sales growing 41% sequentially, went from zero to 20% of the DC business in current sensors in one year, and management guides it well above 20% growth for FY27 after Doogue visited the actual customers in Taiwan and Vietnam. Backlog at a multi-year high, design wins +30% YoY, DSO down to 35 from 40, inventory days 128 from 133, record FCF of $125M funding $60M of debt paydown. Content per rack scales $150 to $425 as racks move to 1MW, and each voltage transition adds Hall, TMR, and eventually isolated gate drivers with 2x-3x uplift. Switching costs rise with ASIL D and proprietary TMR. The Street still models this as a China auto sensor company.
71% of sales is automotive printing sequentially flat with thin inventory, no restocking catalyst, and 25% China exposure. Focus Auto flatness got pinned on Chinese New Year rather than structural China EV weakness. Arcuri's math on the call: if DC grows as guided, the rest of the business falls 3-4% sequentially and ex-DC Industrial is down mid-teens. DC itself decelerates from 40% to 20-25% sequential inside the guide. The $150-to-$425 content story depends on 800V adoption whose speed management admits is "somewhat correlated" to hyperscaler deployment schedules they do not control. Isolated gate drivers are 18-24 months from material revenue. Chinese robot design wins shipping 2026 with volumes in 2027 are pre-revenue narrative. Backlog is at a record with no cancellation terms disclosed while delinquencies build in both auto and Industrial. Two single-input risks stack: hyperscaler CapEx pause kills the growth engine, SAAR decline hits the base.
Backlog Quality
Multi-year high backlog with no cancellation terms, mix, or duration disclosed. D'Antilio simultaneously volunteered delinquency building in pockets of auto and Industrial. Record orders and late-paying customers in the same quarter.
Unsustainable Trend Confidence
Data Center went 2-3% to 14% of sales in two quarters and the guide already embeds a drop from 40% to 20-25% sequential growth. Management frames deceleration as "larger base" rather than demand cooling, and no restocking exists to cushion auto.
HOLD. Execution earns the benefit of the doubt: cash conversion checks out, working capital is tightening, and the margin bridge is itemized and progressing. I cannot size a long when the only growth engine is 14% of revenue decelerating inside the guide and gated on hyperscaler schedules management does not control, while 71% of the base has no restocking cushion and 25% sits in China. I cannot short a business printing beat-and-raise with OCF above net income. Wait for the Q1 print to see whether DC growth is additive or substitutive. Full position only if auto grows while DC hits 16-17%.