Diodes’ beat is not the story: the market is underpricing the 35% gross-margin bridge

The print says Diodes is past the demand trough, but not yet being paid for the earnings power management is now asking investors to underwrite. What was priced in was a modest recovery quarter: Street revenue was $381.3 million and Street EPS was $0.26, low enough to allow a clean beat if orders simply stabilized. What actually surprised was the mix of a +2.7% revenue beat, with actual revenue of $391.6 million, and a much larger +30.8% EPS beat, with actual EPS of $0.34, even though gross margin was still only 31.1%. That combination matters because it tells us the market was looking at Diodes as a revenue-rebound stock, while the company used the call to frame the next three years around margin and profit dollars. The variant perception is that Q4 should not be valued as a one-quarter beat; it should be valued as the first quarter in which the recovery became credible enough for management to publish a path to $2 billion in annual revenue, approximately $700 million in gross profit, and 35% plus gross margin.

The distinction between the Street-comparison basis and the company’s own accounts is important because the earnings beat was not a clean GAAP inflection. On the Street basis, EPS was $0.34 versus $0.26, a +30.8% surprise. In the company’s own reported framework, Brett Whitmire said non-GAAP adjusted net income in Q4 was $15.7 million, or $0.34 per diluted share, while GAAP net income was $10.2 million, or $0.22 per diluted share. The gap was not hidden: non-GAAP adjusted net income excluded, net of tax, $3.9 million of acquisition-related intangible asset costs and $1.6 million of loss on investment. That is why the thesis cannot simply be “EPS beat, buy the stock.” The better argument is that Q4 gave enough evidence of revenue stabilization to make the margin bridge investable, while the current GAAP earnings base still shows how much leverage has not yet arrived.

The revenue trajectory supports that interpretation more than the headline beat alone. Diodes’ quarterly revenue moved from $332.1 million in Q1 FY2025 to $366.2 million in Q2 FY2025, then $392.2 million in Q3 FY2025, before holding at $391.6 million in Q4 FY2025. The Street expected $381.3 million, so the market had priced a step down from Q3, not a hold near $392.2 million. The company’s own wording carried the right nuance: per CFO Brett Whitmire, “Revenue for the fourth quarter of 2025 was $391.6 million, an increase of 15.4% over $339.3 million in the fourth quarter 2024 and essentially flat compared to $392.2 million in the third quarter 2025.” That “essentially flat” matters because Q4 seasonality did not erase the Q2 and Q3 recovery. The next quarter guide of approximately $395 million, plus or minus 3%, is therefore not a heroic acceleration; it is management saying the business can sustain roughly the recovered revenue run-rate into Q1.

The problem, and the opportunity, is that gross margin has not yet validated the revenue recovery. Gross margin was 31.1% in Q4 FY2025, up from 30.7% in Q3 FY2025, but still below 32.7% in Q4 FY2024, 33.7% in Q3 FY2024, and far from the 41.6% and 41.8% levels Diodes printed in Q1 FY2023 and Q2 FY2023. The market can look at that sequence and conclude the recovery is low quality. I think that is too static. The same revenue base that held at $391.6 million in Q4 produced gross profit of $121.9 million, compared with $120.5 million in Q3 on $392.2 million of revenue. That means the incremental signal in the quarter was not volume, it was a 31.1% margin that finally moved off 30.7% despite essentially flat revenue. The move is small, but it is the first piece of evidence needed for the company’s medium-term margin claim.

The expense line explains why the EPS surprise was larger than the revenue surprise, but also why investors should not over-credit one quarter. GAAP operating expenses were $108.7 million, or 27.8% of revenue, while non-GAAP operating expenses were $104 million, or 26.6% of revenue. In the prior quarter, GAAP operating expenses were $108.9 million, or 27.8% of revenue, and non-GAAP operating expenses were $103.1 million, or 26.3% of revenue. In other words, Diodes did not manufacture the beat through a large sequential cost reset; opex dollars were essentially held near Q3 levels while gross profit improved to $121.9 million from $120.5 million. That is the right kind of leverage for a distributor-sensitive, broadline discrete and analog supplier: when revenue stabilizes, cost growth does not have to accelerate for EPS to recover. The bear pushback is valid that non-GAAP adjusted net income fell to $15.7 million from $17.2 million in the prior quarter, and non-GAAP EPS fell to $0.34 from $0.37; the bull response is that the Street was at $0.26 because investors had not credited stability in revenue and gross margin simultaneously.

That brings the debate to management’s new long-range targets, which are unusually explicit for a company still below its prior-cycle margin structure. Gary Yu did not just point to vague recovery; he said, “In order to help our investors track our progress towards these goals, today, I'm introducing 3-year interim financial targets, which include achieving $2 billion in annual revenue with approximately $700 million in gross profit or 35% plus in gross margin.” The commitment is notable because Q4’s gross margin was 31.1%, full year 2025 GAAP gross profit was $462.4 million, and full year 2025 revenue was $1.5 billion. Management is asking the market to believe in a three-year bridge from $1.5 billion revenue and 31.3% full year GAAP gross margin to $2 billion revenue and 35% plus gross margin. That is not a quarterly guide; it is a re-rating framework. If investors accept even part of it, a stock being judged on $0.34 of quarterly non-GAAP EPS has to be reconsidered against the company’s stated $4 EPS objective embedded in the three-year discussion.

The cash-flow evidence makes that target easier to entertain, although it does not solve the margin question by itself. Full year cash flow provided by operations was $215.5 million, up by $96.1 million from $119 million last year, while full year net cash flow was positive $57.6 million, including $33.8 million for the stock buyback program. Q4 net cash flow was negative $9.7 million, including $23.8 million returned to shareholders under the previously announced $100 million stock buyback program, so the quarter itself was not cash generative after repurchase activity. Inventory is the more important tell for next quarter: total inventory dollars increased $600,000 from the prior quarter to $471.5 million, with work in process up $2.1 million, raw materials up $1.2 million, and finished goods down $2.7 million. The composition is consistent with management preparing to support demand rather than sitting on finished goods that customers do not want. That statement is only as strong as the next quarter allows: if finished goods rebuild while revenue only lands near the low end of the guide, the margin bridge weakens.

The call tone supports a more constructive read, but it also shows where management was less convincing. The tone history improved in the places that matter for guidance delivery: guidance_tone rose to 0.54 in Q1 FY2026 from 0.35 in Q4 FY2025, prepared_sentiment rose to 0.60 from 0.54, uncertainty fell to 59.9 from 80.4, and qa_evasiveness fell to 22.3 from 40.1. Those are not market fundamentals, but they are useful when a management team introduces multi-year financial targets; lower uncertainty and higher guidance_tone make the target sound less like aspiration and more like an operating plan. The conflict is that qa_sentiment fell to 0.18 from 0.29 and ai_optimism fell to 0.49 from 0.91, so the Q&A did not validate the prepared-script confidence at the same intensity. I would not ignore that split. It says management was clearer about the framework than it was expansive under questioning, which is precisely why the next two quarters must prove the 35% plus gross-margin path with reported numbers rather than slogans.

The absence of disclosed named customers or suppliers in the data pack limits company-specific supply-chain read-throughs, and that limitation itself matters for portfolio work. There are no customers of Diodes listed and no suppliers to Diodes listed, so this print cannot be used to make a defensible named read-through to a particular buyer or upstream vendor. What it can say for the broader power-discrete chain is that Q4 demand was sufficient for Diodes to beat Street revenue by +2.7% and guide to approximately $395 million, plus or minus 3%, while inventory composition improved through a $2.7 million decrease in finished goods. For competitors and component distributors, that is a tighter signal than “demand is better”: it is a signal that replenishment is not yet forcing a gross-margin reset, because Diodes’ gross margin was only 31.1%, but sell-through was healthy enough to prevent the Street-modeled revenue step down.

The peer comparison reinforces the variant view that Diodes is not merely riding a sector rebound; it is trying to exit the trough with peer-leading growth while still carrying margin upside. In the Power_Discrete peer set, DIOD’s latest reported quarter shows $405.5 million of revenue, 31.8% gross margin, and +22.1% revenue YoY. That compares with VSH at $839.2 million of revenue, 21.0% gross margin, and +17.3% revenue YoY. It also compares with 6503.T at 32.2% gross margin and +14.3% revenue YoY, 6504.T at 31.0% gross margin and +13.3% revenue YoY, and ROHCY at 26.7% gross margin and +9.5% revenue YoY. The point is not that Diodes has already reached a premium margin position; 31.8% is only slightly below 6503.T’s 32.2% and above 6504.T’s 31.0%. The point is that Diodes’ +22.1% revenue YoY in the peer table gives management more room to argue for utilization-driven margin recovery than slower-growth peers with similar or lower gross margins.

The principal risk to the thesis is that the three-year framework gets capitalized before gross margin moves enough to validate it. Q4 gross margin of 31.1% was up from 30.7% in Q3 FY2025, but still down from 32.7% in Q4 FY2024. Full year GAAP gross margin was 31.3%, down from 33.2% in 2024, even though full year revenue increased 13% to $1.5 billion from $1.3 billion. That is the factual tension in the story: revenue is recovering faster than margin. Bulls need that to reverse; bears only need it to persist. Capital expenditures do not point to reckless capacity expansion, with cash-basis capex of $25.7 million in Q4, or 6.6% of revenue, and $78.4 million for the full year, or 5.3% of revenue, within the targeted annualized range of 5% to 9% of revenue. Still, if demand softness reappears, Diodes has not yet earned the right to be valued on 35% plus gross margin.

What to watch next is therefore simple and numerical. For Q1 FY2026, the company guided revenue to approximately $395 million, plus or minus 3%; confirmation starts with revenue holding near that level rather than falling back toward the Q4 Street estimate of $381.3 million. The gross-margin line is more important than the top line: Q4 was 31.1%, Q3 was 30.7%, and the interim target is 35% plus, so the thesis needs another sequential step, not just revenue stability. Inventory should also be checked against the Q4 baseline of $471.5 million, especially finished goods after the $2.7 million decrease in Q4; a rebuild in finished goods without revenue above the guided range would argue the channel is not as clean as the beat implied. Finally, listen on the next call for whether management repeats the three-year targets of $2 billion annual revenue, approximately $700 million gross profit, and 35% plus gross margin, and whether guidance_tone stays near Q1 FY2026’s 0.54 while uncertainty remains closer to 59.9 than Q4 FY2025’s 80.4. If those numbers line up, the market is still too focused on the $0.34 beat and not focused enough on the margin bridge; if they do not, the Q4 print was only a cyclical bounce with aspirational targets attached.