onsemi’s miss is not the story: the buyback-and-capex pivot is absorbing a shallow revenue trough

The clean separation between what was priced in and what actually surprised matters because this was not a demand beat masquerading as a restructuring story. The Street had $1,535.8 million of revenue and $0.62 of EPS embedded; onsemi delivered $1,530.1 million of revenue, a -0.4% surprise, and $0.64 of EPS, a +2.6% surprise. That combination says investors were broadly right on the revenue run-rate but too pessimistic on below-the-line and cost absorption. The defensible thesis from the print is therefore not that demand has snapped back, because Q4 revenue was still down -1.3% sequentially and -11.2% year over year. The thesis is that the market may be mispricing the trough mechanics: at $1.53 billion of revenue, the company still beat EPS by +2.6%, repurchased $450 million of shares in the fourth quarter, deployed approximately 100% of free cash flow to repurchase $1.4 billion of shares in 2025, and guided Q1 revenue to $1.44 billion to $1.54 billion with gross margin of 37.5% to 39.5%. That is not a revenue acceleration thesis; it is an earnings and cash conversion thesis at a still-depressed sales base.

The reason the revenue miss should not dominate the read is that the absolute trajectory shows the business has already absorbed the steepest part of the correction. Revenue fell from $2,180.8 million in Q3 FY2023 to $1,445.7 million in Q1 FY2025, then moved to $1,468.7 million in Q2 FY2025, $1,550.9 million in Q3 FY2025, and $1,530.1 million in Q4 FY2025. The Q4 slip of -1.3% sequentially is not attractive, but it is materially different from the Q1 FY2025 reset of -16.1% sequentially. The actual surprise versus consensus was only -0.4%, so the market was already pricing a flat-to-down seasonal quarter rather than a clean inflection. What it did not fully price was the coexistence of that revenue level with $485 million of free cash flow in the quarter, cash from operations of $555 million, capital expenditures of $69 million or 4.5% of revenue, and cash and short-term investments of approximately $2.5 billion. If one expected a semiconductor downcycle to force a choice between protecting the balance sheet and buying back stock, this quarter says onsemi is not yet facing that choice.

The margin line is where the debate gets more nuanced, because reported history and management’s non-GAAP framing point in different directions without being contradictory. The quarterly history shows gross margin at 34.9% in Q4 FY2025, down from 37.9% in Q3 FY2025 and far below the 46.7% level in Q4 FY2023. On the call, Hassane El-Khoury framed the company’s own operating basis differently: “Turning to the demand environment, we are seeing seasonal patterns and are encouraged by improving order trends across our core markets, contributing to fourth quarter revenue of $1.53 billion, non-GAAP gross margin of 38.2%, and earnings per share of $0.64, both exceeding the midpoint of our guidance.” The wording matters because he did not claim a broad demand recovery; he tied the quarter to “seasonal patterns” and “improving order trends,” which is a narrower statement. The conflict is not in the direction of the business, but in the basis of measurement: the print table shows Q4 gross margin at 34.9%, while management’s call language points to non-GAAP gross margin of 38.2%. For a portfolio manager, that distinction is central. The GAAP margin history still shows structural damage versus 2023, but the non-GAAP guide of 37.5% to 39.5% for Q1 says the company expects the operating margin base to stay near the management-defined Q4 level even with midpoint revenue below Q4.

That bridge from margin to cash is the core of the variant perception, because onsemi’s capacity cycle is now reversing from drag to support. Thad Trent’s most important sentence was not the Q1 revenue guide; it was the explicit capex and depreciation handoff: “These actions together will lower our 2026 depreciation by approximately $45 to $50 million, and we expect to see the gross margin impact in the second half of the year.” This is a measurable commitment, not a generic productivity promise. Q4 capital expenditures were $69 million or 4.5% of revenue, and Q1 capital expenditures are expected in the range of $35 to $45 million. Free cash flow increased 17% year over year to $1.4 billion in 2025, and 2025 free cash flow was 24% of revenue. The market’s normal instinct is to wait for auto and industrial revenue to turn before underwriting margin expansion, but onsemi has a second lever from lower depreciation that begins in the second half of the year. That does not rescue a demand collapse, but it changes the earnings sensitivity if revenue simply holds near the guided Q1 range.

The segment detail supports that “stabilization, not surge” interpretation and argues against paying for a broad cyclical reacceleration today. Automotive revenue was $798 million, up approximately 1% quarter over quarter, and industrial revenue was $442 million, up approximately 4% quarter over quarter. Those are the two end-market figures that matter most for the stock narrative, and neither says the customer base has returned to 2023 behavior. PSG revenue was $724 million, down 2% quarter over quarter and down 11% year over year; AMG revenue was $556 million, down 5% quarter over quarter and down 9% year over year; ISG revenue was $250 million, up 9% quarter over quarter but down 17% year over year. The second-order read is that the recovery is uneven: industrial helped ISG sequentially, while the larger power and analog businesses still declined. That matters for suppliers because Tower Semiconductor, a specialty foundry for image sensors and power, gets a mixed signal from a $250 million ISG quarter that rose 9% sequentially but remained down 17% year over year, while SICC and Wolfspeed, tied to SiC substrates and 150mm SiC wafers, should read the automotive data as stabilization rather than a wafer-order acceleration, given automotive revenue rose only approximately 1% quarter over quarter. The magnitude is the point: these are not numbers that justify a sudden restock call across the supply chain.

The same data also pushes back on a bearish conclusion that onsemi is simply losing mix or pricing faster than it can cut costs. Inventory decreased by $58 million to 192 days from 194 days in Q3, which is still elevated but moving in the right direction. Non-GAAP operating expenses declined 3% sequentially to $282 million at the lower end of guidance, while GAAP operating expenses were $351 million, including $59 million in restructuring expenses. For Q1, non-GAAP operating expenses are expected to be between $285 and $300 million, including share-based compensation of $29 million. This is not a clean cost-down story, because Q1 non-GAAP operating expenses are guided above Q4’s $282 million. But the company is pairing that with lower capital intensity and a share count expectation of approximately 397 million shares, after $450 million of repurchases in Q4 and a new $6 billion share repurchase program announced in November following $2.6 billion repurchased under the prior program that expired at the end of 2025. If the top line is range-bound, the buyback becomes part of the EPS bridge rather than excess capital decoration.

The competitive read is also more subtle than the headline miss suggests, because onsemi is not keeping pace with the best analog peers on revenue growth or margin, yet it is no longer the outlier at the bottom of the IDM margin stack. TXN reported $4,825.0 million of revenue, 58.0% gross margin, and +18.6% revenue YoY, while NXPI reported $3,181.0 million of revenue, 56.2% gross margin, and +12.2% revenue YoY. Those figures make onsemi’s $1,530.1 million of revenue and 34.9% gross margin look late-cycle and still impaired. But STM reported $3,095.0 million of revenue with 33.8% gross margin and +22.8% revenue YoY, while IFNNY reported €3,874.7 million of revenue with 38.7% gross margin and +7.9% revenue YoY. The peer point is not that onsemi deserves a premium multiple to analog leaders with 56.2% to 58.0% gross margin. It is that onsemi’s equity debate should be framed against its own depressed margin base and capital returns, not against a fantasy of immediate convergence with TXN or NXPI. Joe Moore’s call reference to the long-term targets, the 53% gross margin and 40% operating margin, is useful precisely because the current 34.9% gross margin history shows how much execution remains before those targets matter to valuation.

The tone of the call reinforces the “measured stabilization” view, although it also warns against over-reading management confidence. The tone history shows Q4 FY2025 sentiment at 0.24, guidance_tone at 0.35, prepared_sentiment at 0.41, qa_sentiment at 0.14, ai_optimism at 0.47, uncertainty at 54.1, and qa_evasiveness at -33.2. The next recorded call, Q1 FY2026, moved sentiment to 0.31 and guidance_tone to 0.51, with qa_sentiment rising to 0.35 and uncertainty falling to 41.2, but prepared_sentiment dropped to 0.01, ai_optimism fell to 0.35, and qa_evasiveness rose to 41.4. The conflict in those figures matters: guidance delivery improved by +0.17 and uncertainty fell by -12.9, yet ai_optimism declined by -0.12 and qa_evasiveness increased by +74.6. In plain English, management sounded more constructive on the forward framework but less promotional in prepared remarks and less clean in Q&A. That is consistent with a company willing to underwrite seasonality and cost actions, not one ready to declare a demand breakout.

That delivery profile matches the Q1 guide, which is deliberately narrow in message even if the range is wide in dollars. Trent said, “We anticipate Q1 revenue will be in the range of $1.44 billion to $1.54 billion in line with normal seasonality at the midpoint.” The phrase “normal seasonality” is the part investors should underwrite, because it does not require a restocking cycle or a rapid auto production recovery. Q1 FY2026 in the quarterly history shows $1,513.3 million of revenue, 38.5% gross margin, -1.1% revenue QoQ, +4.7% revenue YoY, and -$0.08 diluted EPS. Again, basis matters: the history includes a negative diluted EPS figure, while the call guide is non-GAAP and includes non-GAAP gross margin of 37.5% to 39.5%, a non-GAAP tax rate of approximately 15%, and a non-GAAP diluted share count of approximately 397 million shares. The thesis does not need GAAP EPS to turn immediately. It needs revenue to avoid another Q1 FY2025-style -16.1% sequential break and non-GAAP gross margin to stay inside the 37.5% to 39.5% range until depreciation relief arrives in the second half of the year.

The investment conclusion is therefore intentionally narrower than a generic semiconductor recovery call. What was priced in was a roughly in-line revenue quarter, with the Street at $1,535.8 million and actual at $1,530.1 million, plus some skepticism after revenue remained down -11.2% year over year and gross margin in the history sat at 34.9%. What surprised was the $0.64 EPS versus $0.62 expected, the $485 million of Q4 free cash flow, the $1.4 billion of 2025 free cash flow, the repurchase intensity of approximately 100% of free cash flow, and the willingness to guide Q1 non-GAAP gross margin to 37.5% to 39.5% despite Q1 revenue of $1.44 billion to $1.54 billion. The market may be missing that onsemi can generate an acceptable equity outcome without a V-shaped revenue recovery, because capex is guided to $35 to $45 million, depreciation is set to fall by approximately $45 to $50 million in 2026, and the new $6 billion repurchase authorization gives management a direct mechanism to convert free cash flow into per-share support.

What to watch next is precise. The thesis is confirmed if Q1 revenue stays within the $1.44 billion to $1.54 billion range, non-GAAP gross margin lands within 37.5% to 39.5%, capital expenditures are within $35 to $45 million, and inventory days move below the Q4 level of 192 days rather than back toward 194 days. It is strengthened if automotive grows more than the Q4 approximately 1% quarter-over-quarter pace and industrial holds or improves on the Q4 approximately 4% quarter-over-quarter pace, because that would turn the current stabilization case into a demand-led case. It breaks if revenue falls below $1.44 billion, if non-GAAP gross margin misses 37.5%, if non-GAAP operating expenses exceed the $285 to $300 million guide, or if management backs away from the second-half gross-margin benefit tied to approximately $45 to $50 million lower 2026 depreciation. The next quarter does not need to prove the 53% gross margin and 40% operating margin targets are near; it needs to prove that the $1.5 billion revenue floor can fund cash returns while the depreciation tailwind begins to show up.