The CEO volunteered a $4 EPS target while the company printed $1.22, and told analysts the margin story is underloading charges going away.
Thesis: Diodes is a clean balance sheet with genuine cash flow recovery but a widening gap between what the CEO is promising and what the income statement is delivering. The 45% pull-through story hinges entirely on fab utilization recovering, which is a cyclical variable, not a structural one. Hold until Q1 proves sequential margin expansion above 31.5%.
Verdict: HOLD — Conviction: MEDIUM
Catalyst: Q1 2026 gross margin printing above guided 31.5% would confirm the 45% pull-through math and validate the underloading charge dissipation narrative.
Key Risk: The 45% pull-through depends on fab utilization rising. If revenue stalls or underloading persists, the $4 EPS target becomes a bag-holder thesis for anyone buying on the 2028 story.
The Tell: William Stein directly asked whether the 45% pull-through was underutilization charges going away rather than normal operating leverage. Yu answered 'definitely, under loading charge is going to be the key factor.' He confirmed the margin story is fab utilization recovery, a cyclical variable, not structural pricing power or mix improvement.
Friction Level: MODERATE_FRICTION — Both sides agree FY2025 showed revenue growth with declining adjusted EPS. The bull reads the 45% incremental pull-through as operating leverage waiting to unlock. The bear reads current negative leverage as the baseline that contradicts the target.
2025Q4 · Feb 10, 2026
5 of 5 readings agreed5 HOLDRead the full call
Revenue Up 13%, Adjusted EPS Down 7%, and Management Just Promised 50% Growth
Diodes is a clean balance sheet with genuine cash flow recovery but a widening gap between what the CEO is promising and what the income statement is delivering. The 45% pull-through story hinges entirely on fab utilization recovering, which is a cyclical variable, not a structural one. Hold until Q1 proves sequential margin expansion above 31.5%.
William Stein directly asked whether the 45% pull-through was underutilization charges going away rather than normal operating leverage. Yu answered 'definitely, under loading charge is going to be the key factor.' He confirmed the margin story is fab utilization recovery, a cyclical variable, not structural pricing power or mix improvement.
FY2025 non-GAAP EPS fell 7% to $1.22 on 13% revenue growth. Management just set a target of $4 EPS in three years implying 50% annual growth while currently posting negative operating leverage.
Fourth consecutive quarter of double-digit YoY growth. Q1 guide of $395M at +19% YoY in a seasonally down quarter, above typical seasonal declines. Channel inventory normalized, bookings strong.
Automotive credited for margin improvement at 20% of revenue while consumer at 17% fell 5% sequentially. GM at 31.1% sits 160bps below year-ago. The mix improvement is real but small relative to the gap to 35%.
Computing grew 25% and AI server content reached $103/unit, but computing is 28% of revenue on a base including notebooks and desktop. The AI attribution overstates the AI-weighted portion.
Emily Yang cited 'really strong book-to-bill' without a ratio. Q1 guidance described as plugged with seasonality rather than anchored to non-cancellable orders. No B2B number disclosed.
Q1 at +19% YoY combines Nexperia disruption revenue, depressed prior-year comparison, and auto/industrial recovery. Each tailwind is time-limited. Management's hedge-free language sits on top of cyclical support.
+72
CONFIDENT"Yu volunteered three-year interim targets unprompted, committed to a specific timeline by 2028, said he went to the board first, and confirmed 'definitely' when Stein probed the pull-through math. Direct, specific, forward commitments."
Specific revenue, margin, content, and cash flow numbers throughout. No ratios given on book-to-bill. MSA and Nexperia impacts explicitly undisclosed.
Q1 2026 gross margin printing above guided 31.5% would confirm the 45% pull-through math and validate the underloading charge dissipation narrative.
The 45% pull-through depends on fab utilization rising. If revenue stalls or underloading persists, the $4 EPS target becomes a bag-holder thesis for anyone buying on the 2028 story.
“Revenue for the fourth quarter of 2025 was $391.6 million, an increase of 15.4% over $339.3 million in the fourth quarter of 2024, and essentially flat compared to $392.2 million in the third quarter of 2025.”
“Now turning to our outlook, for the first quarter of 2026, we expect revenue to be approximately $395 million, plus or minus 3%.”
Figures are extracted from the transcripts and checked in code: every quote appears word for word. Descriptive only: none of this feeds the verdict.
Q4 revenue $391.6M at +15.4% YoY, fourth straight double-digit quarter. Q1 guide $395M at +19% YoY in a seasonally down quarter. Channel inventory at 11-14 week normal range after years of destocking. Auto content per vehicle $239 vs $160 two years ago. AI server content $90 to $103. Design wins with 650 new part numbers, 40% auto. FCF $137.2M, triple from prior year, balance sheet clean with $56M debt. Yu committed publicly to $2B revenue and 35% GM by 2028 with specific interim milestones. Nexperia disruption is unquantified upside the street has zero in. Underloading charges are the known cost lever, and revenue scale activates it.
FY2025 revenue up 13%, non-GAAP EPS down 7% to $1.22. That is negative operating leverage, the exact opposite of the 45% pull-through narrative. Gross margin compressed 190bps year-over-year to 31.3%. Q4 GM at 31.1% is 160bps below year-ago despite auto mix. MSA expiring in 2026 is an undisclosed binary event Yu refuses to size. Nexperia windfall is lower margin per Yang and temporary. Book-to-bill cited as 'strong' with no ratio. Consumer segment fell 5% sequentially. The $4 EPS target requires a step-change in leverage that the last four quarters of data do not support.
Say/Do Gap
Revenue grew 13% while adjusted EPS fell 7%. Management set a 50% EPS CAGR target from a base where operating leverage is currently negative. The numbers contradict the narrative.
Mix Shift Spin
Auto at 20% of revenue credited for margin improvement, yet GM is 160bps below year-ago. Consumer at 17% declining sequentially. Mix story outruns mix math by a wide margin.
Backlog Quality
Yang claims 'really strong book-to-bill' but provides no ratio. Q1 guidance is described as plugging in seasonality, not driven by firm orders. No B2B number to validate.
AI Label Without Substance
Computing at 28% of revenue includes notebooks and desktop. AI server content of $103/unit is a small absolute number. The AI premium in the multiple is doing heavy lifting on a diversified franchise.
Unsustainable Trend Confidence
Q1 above-seasonal combines Nexperia disruption, depressed base, and cyclical recovery. Management's hedge-free posture sits on trends that each have a clock.
Hold. Cash flow is real, balance sheet is clean, inventory discipline is evident, and there is no channel stuffing. But the earnings target versus current execution gap is too wide to size into. Wait for Q1 gross margin above 31.5% to confirm pull-through, or a book-to-bill number above 1.0 to confirm backlog quality. The fab utilization story is cyclical. Do not buy the 2028 target until the intermediate quarters deliver.