The Dilution Reserve Comes Off and the Multiple Comes Down: Maintain Outperform on onsemi at 15x FY27
Key Takeaways
- onsemi chose AI over automotive, and the quarter shows the trade worked. Revenue of $1,722.3 million grew 11.1% and finished $22.3 million above the guidance midpoint. Non-GAAP gross margin rose 190 basis points to 41.2%, the fifth consecutive increase, and adjusted EPS of $0.90 grew 42.9%. Automotive revenue fell 1.3% year over year because management reallocated constrained supply, not because automotive demand failed.
- AI data center became a reported segment and a much larger number in the same quarter. It reached $180.2 million, up 174.7%, and the 2026 outlook rose to about $620 million from the "greater than $500 million" shown at the September Investor Day. Management reiterated a doubling in 2027 off that higher base, and the Ferric acquisition lifts the 2030 target to more than $3 billion.
- The Synaptics deal is now $5.7 billion in cash, which retires the dilution risk and replaces it with leverage. Our prior target carried a reserve for roughly 53 million new shares. That reserve is gone. In its place sit a $2.45 billion committed term loan, a buyback that fell to $19.9 million from $344.8 million, and delevering as the stated capital priority after closing.
- Margin now has a published bridge, and most of it is mechanical. About 450 basis points of under-absorption still sits in the Q3 cost base against a 77% utilization starting point, with Fab Right worth 225 basis points and new-product mix 400 basis points on the path to a 53% 2030 target. Utilization held flat at 83% and rises in 2027.
- Rating: Maintaining Outperform, with the twelve-month target held at approximately $95. Removing the share reserve lifts our FY2027 valuation EPS to about $4.73 from $4.31, while we cut the multiple to 20 times from 22 for added leverage and a higher-beta revenue mix. At the $71.24 close the stock trades at 15.1 times that figure, 46.8% below its June high.
Results vs. Consensus
Quarter ended October 2, 2026. Dollar amounts are in millions unless shown per share. Street figures cluster at $1,706 million of revenue and $0.88 of adjusted EPS across four of the five estimate sources we reviewed; one outlier carried a $1,739 million revenue figure that matches what other outlets published as the fourth-quarter consensus, so we treat it as a period mislabel.
| Metric | Q3 2026 actual | Consensus | Beat/Miss | Magnitude |
|---|---|---|---|---|
| Revenue | $1,722.3 | $1,706 | Beat | +$16.3 / +0.96% |
| EPS, non-GAAP | $0.90 | $0.88 | Beat | +$0.02 / +2.3% |
| Gross margin, non-GAAP | 41.2% | n/a | Beat vs. guide | +20 bp vs. 41.0% midpoint |
| Operating income, non-GAAP | $399.8 | n/a | n/a | +34.2% YoY |
| Operating expenses, non-GAAP | $309.5 | n/a | Favorable vs. guide | $1.0 below the $310.5 midpoint |
| EPS, GAAP | $0.79 | n/a | n/a | +25.4% YoY |
| Free cash flow | $488.6 | n/a | n/a | 28.4% of revenue |
Against management's own third-quarter guidance, every P&L line landed at or better than the midpoint and inside the range. Revenue beat the $1,700 million midpoint by $22.3 million, gross margin beat the 41.0% midpoint by 20 basis points, adjusted operating expense came in $1.0 million below the $310.5 million midpoint, and adjusted EPS beat the $0.87 midpoint by three cents. The adjusted share count of 392.2 million was 2.8 million below the roughly 395 million guided, and the non-GAAP tax rate held at the guided 15%.
Year-over-Year Comparison
| Metric | Q3 2025 | Q3 2026 | YoY change |
|---|---|---|---|
| Revenue | $1,550.9 | $1,722.3 | +11.1% |
| Gross margin, GAAP | 37.9% | 41.4% | +350 bp |
| Gross margin, non-GAAP | 38.0% | 41.2% | +320 bp |
| Operating expenses, non-GAAP | $291.2 | $309.5 | +6.3% |
| Operating margin, non-GAAP | 19.2% | 23.2% | +400 bp |
| Operating income, non-GAAP | $297.8 | $399.8 | +34.2% |
| Net income attributable to ON, GAAP | $255.0 | $313.9 | +23.1% |
| Net income attributable to ON, non-GAAP | $256.3 | $354.0 | +38.1% |
| EPS, GAAP | $0.63 | $0.79 | +25.4% |
| EPS, non-GAAP | $0.63 | $0.90 | +42.9% |
| Cash from operations | $418.7 | $516.4 | +23.3% |
| Free cash flow | $372.4 | $488.6 | +31.2% |
| Adjusted diluted shares | 408.0 | 392.2 | −3.9% |
Quarter-over-Quarter Comparison
| Metric | Q2 2026 | Q3 2026 | QoQ change |
|---|---|---|---|
| Revenue | $1,603.5 | $1,722.3 | +7.4% |
| Gross margin, GAAP | 38.4% | 41.4% | +300 bp |
| Gross margin, non-GAAP | 39.3% | 41.2% | +190 bp |
| Operating expenses, non-GAAP | $296.8 | $309.5 | +4.3% |
| Operating margin, non-GAAP | 20.8% | 23.2% | +240 bp |
| Operating income, non-GAAP | $334.1 | $399.8 | +19.7% |
| Restructuring and asset impairments, net | $41.2 | $4.6 | −88.8% |
| EPS, GAAP | $0.56 | $0.79 | +41.1% |
| EPS, non-GAAP | $0.74 | $0.90 | +21.6% |
| Free cash flow | $425.4 | $488.6 | +14.9% |
| Share repurchases | $344.8 | $19.9 | −94.2% |
Revenue. The 0.96% consensus beat understates what happened inside the quarter. Revenue grew 7.4% sequentially while automotive, which is 45.1% of the business, went backwards. Industrial grew 10.4% sequentially and AI data center grew about 39%, and management was explicit that it chose to send constrained supply to those two markets rather than to automotive replenishment orders. That is a materially different quarter from one in which all three end markets lifted together. It validates the intelligent-power franchise argument and leaves the breadth of the cyclical recovery still unproven.
Margins. The 190-basis-point adjusted step is the largest of the five consecutive increases, and it is the most important number in the print. We verified the series rather than take the company's word for it: adjusted gross margin ran 37.6% in Q2 2025, then 38.0%, 38.2%, 38.5%, 39.3% and 41.2%, so the fifth consecutive increase claim holds and the cadence is accelerating. GAAP gross margin of 41.4% sat 20 basis points above the adjusted figure this quarter, a reversal of the usual direction, because restructuring-related inventory charges swung to a $4.9 million credit. Two drivers were named: utilization improvements earned two quarters earlier, and favorable mix toward new products. Neither is a one-time item, but the mix component is the one that can reverse if automotive shipments resume at the same time AI growth moderates.
EPS. Adjusted net income grew 38.1% while adjusted EPS grew 42.9%; the gap is the 3.9% reduction in the share count from buybacks completed earlier in the year. The $40.1 million spread between GAAP and adjusted net income deserves a closer look than usual, because $22.3 million of it is an income-tax adjustment: the GAAP tax rate ran 21.3% this quarter against the 15% non-GAAP convention. Taken alone, that would make the quarter's adjusted premium more than half tax-driven. Over the nine months the adjustment runs the other way, a $42.7 million drag against an 18.6% GAAP rate, so the nine-month adjusted premium is entirely operating and restructuring in nature. The honest reading is that Q3's higher GAAP tax rate widened the quarter's bridge without changing the year's.
Segment Performance
Business Units
| Revenue | Q3 2025 | Q2 2026 | Q3 2026 | YoY | QoQ |
|---|---|---|---|---|---|
| Power Solutions Group (PSG) | $737.6 | $829.0 | $889.9 | +20.6% | +7.3% |
| Analog & Mixed-Signal Group (AMG) | $583.3 | $545.7 | $597.3 | +2.4% | +9.5% |
| Intelligent Sensing Group (ISG) | $230.0 | $228.8 | $235.1 | +2.2% | +2.8% |
| Total | $1,550.9 | $1,603.5 | $1,722.3 | +11.1% | +7.4% |
Power Solutions Group
PSG supplied $60.9 million of the company's $118.8 million sequential increase and $152.3 million of the $171.4 million year-over-year increase. Put differently, power accounted for roughly 89% of the year-over-year growth in a quarter when total revenue grew 11.1%. The AI power tree is the proximate cause: silicon carbide and silicon MOSFETs for power supply units and for 54-volt to 12-volt conversion, silicon carbide JFETs for solid-state disconnects, and multi-phase controllers closer to the processor.
"In the third quarter, we began volume shipments into multiple marquee platforms, including Google with both silicon carbide and silicon MOSFETs for PSUs and 54-volt to 12-volt applications." — Hassane El-Khoury, President and CEO
Assessment: PSG is doing the work that the standing franchise pillar predicted, and doing more of it than our model assumed. The concentration cuts both ways: a business where one unit supplies nearly nine-tenths of growth is a franchise story when that unit is winning and a single-point-of-failure story when it is not. We treat PSG's momentum as evidence for duration rather than as a reason to raise the multiple.
Analog & Mixed-Signal Group
AMG grew 9.5% sequentially, its best sequential result of the coverage period, but remains only 2.4% above last year and still below where it sat in the third quarter of 2025 on a nine-month basis. The quarter's strength came partly from wireless seasonality inside the Other end market rather than from a broad analog recovery. The Treo platform is now described as contributing revenue, with the 10BASE-T1S Ethernet ramp underway in North American automotive programs and the 2026 sampling cohort including 48-volt SmartFET high-side drivers, isolated silicon carbide gate drivers and inductive position sensors.
"So you can think about the base for 2027 as new products that we've had in '25 and '26 that will start contributing meaningfully, while the new products, we've introduced like EPP and vertical GaN to start in '27 and continue forward, all leading up to a more mature mix of new products towards that 2030-time line depending on auto or industrial." — Hassane El-Khoury, President and CEO
Assessment: This is the first quarter in which the new-product pipeline reads as a 2027 revenue line rather than a design-win narrative. We still do not credit analog with a sharp 2027 acceleration, because the sequential improvement carries a seasonal component management itself called out and will not repeat. Treo matters to mix and therefore to gross margin more than it matters to revenue growth in the next four quarters.
Intelligent Sensing Group
ISG grew 2.8% sequentially and 2.2% year over year, the smallest contribution of the three units. The forward story is robotics and physical AI rather than automotive imaging: shipments of 2-megapixel and 8-megapixel Hyperlux image sensors into first-person-view drones, engagement with more than ten drone manufacturers, design wins with United States humanoid makers, and indirect time-of-flight depth resolution claimed at one millimeter.
"We continue to diversify across the robotics market and have secured multiple new design wins with leading U.S. humanoid makers across power, connectivity and sensing." — Hassane El-Khoury, President and CEO
Assessment: Sensing remains portfolio breadth rather than the marginal driver of the earnings case. The September Investor Day sized physical AI at a $6 billion 2030 market with roughly $900 of content per humanoid, which is a credible option but not a 2027 revenue line. We keep a measured recovery assumption here.
End Markets
The company reported AI data center as a separate fourth end market for the first time this quarter, and recast the prior year on the same basis in its quarterly filing. Sequential changes are exact for automotive, industrial and the combined other-plus-AI bucket; the split of the prior quarter between AI data center and the remaining other markets is derived from management's stated sequential growth rates and is marked accordingly.
| End market | Q3 2025 | Q2 2026 | Q3 2026 | YoY | QoQ | Q4 2026 management outlook |
|---|---|---|---|---|---|---|
| Automotive | $787.3 | $781.3 | $776.7 | −1.3% | −0.6% | Down high single digits |
| Industrial | $426.3 | $422.7 | $466.8 | +9.5% | +10.4% | Up low double digits |
| AI data center | $65.6 | ~$130 (derived) | $180.2 | +174.7% | +39% (as stated) | Up mid teens |
| Other, excluding AI data center | $271.7 | ~$270 (derived) | $298.6 | +9.9% | +11% (as stated) | Up high teens |
| Total | $1,550.9 | $1,603.5 | $1,722.3 | +11.1% | +7.4% | $1,715–1,815 |
Automotive
Automotive fell 0.6% sequentially and 1.3% year over year, and management attributed the decline to a deliberate choice rather than to demand. Supply was redirected to AI data center and industrial, where order strength exceeded plan. End demand was fully served; replenishment orders were not. Content, not unit volume, remains the growth mechanism: roughly $2,000 of content per vehicle as architectures move to 48-volt zonal designs and 800-volt-plus electric platforms, with new 1,200-volt silicon carbide traction inverter wins at NIO and Geely Zeekr in China and onboard charger platform wins in North America.
"So when we talk about resuming to our long-term model, which is high single digit over SAAR, that's the one I'm referring to. That's what the OEMs place orders for. It's -- they see their content growth in their new platforms that they are introducing, not necessarily a SAAR or unit. So that's what I want to be very clear. Inventory for us is content in the car." — Hassane El-Khoury, President and CEO
Assessment: The content argument is the right frame, and the nine-month figures support it: automotive revenue grew 3.2% over nine months while the quarter declined. But the fourth-quarter guide takes automotive down another high single digits, which means the segment will have declined sequentially in each of the last three quarters of the year. Management has put the recovery on a specific date, the first quarter of 2027, and attached it to a specific mechanism, supply coming online that is already paid for. That is the single clearest falsifiable commitment from this call, and we will grade it in February.
Industrial
Industrial grew 10.4% sequentially, well above the flat outlook management gave three months ago, and 9.5% year over year. The driver named was the AI halo rather than a broad industrial cycle: energy infrastructure from the grid to the data center, with hybrid EliteSiC and IGBT module wins at Sungrow and Sineng on the 99.3% efficiency platform introduced last quarter. Management was also explicit that none of the strength is channel refill, pointing to visibility through its customers to their end customers and to the reported results of the large grid-to-data-center equipment makers.
"And the broader industrial is about 40% of the total industrial bucket, and we're seeing strength there as well. I think just as you're seeing PMIs improve." — Thad Trent, Executive Vice President and CFO
Assessment: Industrial grew 10.4% sequentially against a guide for relatively flat revenue, a beat of roughly ten percentage points, and it supplied $44.1 million of the quarter's $118.8 million sequential increase. It also narrows the gap between the AI thesis and the cyclical thesis, because the September Investor Day sized the AI halo at $9 billion of industrial demand built to support data centers. We raise our industrial assumption, while noting that management puts the broader industrial economy at only about 40% of the segment, which makes the majority of industrial a partial second expression of the same AI capital cycle rather than an independent diversifier.
AI Data Center
At $180.2 million, AI data center was 10.5% of revenue this quarter and grew 174.7% year over year. The 2026 outlook moved to approximately $620 million, a 140% annual increase and more than 20% above the "greater than $500 million" figure shown at the Investor Day seven weeks earlier. Revenue now spans the complete power tree, from power supply units through low-voltage smart power stages near the processor, and silicon carbide JFETs are already contributing to data center revenue through solid-state disconnect applications.
"In the third quarter, our AI data center revenue grew 39% sequentially to $180 million, and we now expect to deliver approximately $620 million of revenue in 2026, an increase of 140% year-over-year. For 2027, we still expect our AI data center revenue to double over 2026 as we continue to ship with new product ramps." — Hassane El-Khoury, President and CEO
Assessment: The quarter resolves a disclosure complaint we have raised repeatedly, and the resolution is favorable. It also changes the shape of the investment case. On our estimates AI data center goes from 9.4% of 2026 revenue to roughly 16% of 2027 revenue and supplies about 57% of the incremental dollars. A business where the majority of next year's growth comes from one high-beta end market is not the same risk as a diversified industrial compounder, and we reflect that in a lower multiple rather than in a lower growth estimate.
Other, excluding AI Data Center
The residual other bucket, covering computing, consumer, networking and communications, grew about 11% sequentially and 9.9% year over year on wireless seasonality, and is guided up high teens in the fourth quarter on a single large customer program. Management flagged two things about that program: it is externally manufactured, so it does not consume constrained internal capacity, and it is lumpy rather than a new run rate.
Assessment: We take the fourth-quarter strength at face value and do not extend it. Management guided this bucket back toward "normal historical rates" in 2027, which the Investor Day framed as a 1% long-term growth rate. Nine-month other revenue is down 0.9% against last year, but the guided fourth-quarter program alone lifts the full year to growth of more than 10% on our estimates, which is why we treat it as a one-off. We model it as a non-contributor.
Product Technologies
Quarter ended October 2, 2026
| Technology | Q3 2025 | Q3 2026 | YoY |
|---|---|---|---|
| Intelligent power | $792.8 | $918.8 | +15.9% |
| Intelligent sensing | $284.0 | $296.2 | +4.3% |
| Other | $474.1 | $507.3 | +7.0% |
| Total | $1,550.9 | $1,722.3 | +11.1% |
Nine months ended October 2, 2026
| Technology | 9M 2025 | 9M 2026 | YoY |
|---|---|---|---|
| Intelligent power | $2,243.1 | $2,531.1 | +12.8% |
| Intelligent sensing | $859.8 | $873.5 | +1.6% |
| Other | $1,362.4 | $1,434.5 | +5.3% |
| Total | $4,465.3 | $4,839.1 | +8.4% |
Assessment: The technology cut is the cleanest view of what is actually compounding. Intelligent power grew 15.9% in the quarter and 12.8% over nine months, about 4.8 and 4.4 percentage points faster than the company; everything else grew below the company average. That gap is the entire margin story in one line, because intelligent power is where the accretive new products sit.
Key KPIs
| Operating indicator | Q2 2026 | Q3 2026 | Trend | Investment implication |
|---|---|---|---|---|
| Factory utilization | 83% | 83% | Flat | Q3 cost of revenue still reflects the 77% Q1 base; Q4 similar, rising in 2027 |
| Under-absorption still in cost of revenue | n/a | ~450 bp | Disclosed | Implies about 92% to 93% as the practical ceiling at 25–30 bp per point |
| Inventory days | 192 | 186 | Down 6 days | Dollars up $22.2M; days down on higher shipments |
| Strategic inventory | n/a | Down 7 days | Draining | Base inventory up 1 day to support demand |
| Distribution weeks of inventory | 10.1 | 9.9 | Down | Sell-through still exceeds sell-in |
| Capital expenditures | $34.3 | $27.8 | Down | 1.6% of revenue; about 2% for 2026, about 5% guided for 2027 |
| Share repurchases | $344.8 | $19.9 | Sharply down | Deal-related trading restrictions |
| Free cash flow margin | 26.5% | 28.4% | Up | Trailing twelve months 25.4% on $1,616.6M |
| Cash and short-term investments | $3,864.5 | $4,367.7 | Up | Against $4,463.0M of debt; net debt about $95M |
| Long-term supply agreement obligations | ~$6.1B | ~$5.5B | Down | 39% expected within twelve months, up from 37% |
| AI data center share of revenue | ~8% | 10.5% | Up | About 9.4% for 2026 on our estimate |
Key Topics & Management Commentary
Overall Management Tone: Management was the most forward-leaning of the coverage period, framing a sequential automotive decline as a deliberate allocation choice rather than defending it, and treating two acquisitions and a restructured $5.7 billion transaction as routine capital deployment. Analyst pushback was narrow and concentrated on whether withholding automotive supply costs share; the response was specific and dated rather than framework-level. The notable absence of pressure was on the balance sheet, where a near-complete halt in buybacks and a step in capital intensity from about 2% to about 5% of revenue drew only two questions.
1. Automotive Supply Was Withheld on Purpose, and That Is a New Kind of Risk
The most consequential disclosure of the quarter is that onsemi is short of capacity in specific product lines and chose to serve AI data center and industrial customers ahead of automotive replenishment orders. End demand in automotive was fully supported. The inventory rebuild that onsemi's customers wanted was not.
"We remain supply limited in certain technology segments of our business, and we made the short-term decision in the third quarter and into the fourth quarter to allocate to markets where we see strong end demand like AI data center, industrial, which came in much stronger than we originally planned. In automotive, we are supporting all deliveries for end demand, but we are not supporting inventory replenishment backlog." — Hassane El-Khoury, President and CEO
The fourth-quarter guide extends the decision, with automotive down high single digits on continued reallocation plus timing and phasing of customer deliveries. Management placed the recovery in the first quarter of 2027 and attached it to capacity that is already installed and paid for, arriving through normal fab and back-end cycle times.
Assessment: A company that can choose which customers to serve has pricing power, which is the favorable reading. The unfavorable reading is that onsemi has now spent two quarters, and will spend a third, telling automotive customers to wait, in a market where the standing thesis says content gains are won platform by platform. We accept the no-share-loss claim for now because it is checkable, and we have put the first quarter of 2027 automotive number at the top of our watch list.
2. AI Data Center Becomes a Reported Segment in the Same Quarter It Becomes Material
We have repeatedly flagged the absence of a reported AI data center revenue line as a gap in onsemi's reporting. That gap closed this quarter. AI data center is now a fourth reported end market in the quarterly filing and in the company's supplemental materials, with the prior year recast on the same basis at $65.6 million.
"Please note, beginning this quarter, we are enhancing our disclosures by separately reporting AI data center revenue as a fourth market segment to provide greater transparency into this important growth driver for onsemi." — Thad Trent, Executive Vice President and CFO
The disclosure arrives with the number already large. Nine-month AI data center revenue of $413.2 million compares with $179.0 million a year earlier, and the fourth quarter is implied at roughly $207 million by the approximately $620 million annual figure.
Assessment: Companies break out a segment when it helps them, and this one helps. But the disclosure is genuinely useful: it converts a narrative into a line we can forecast and grade, and it removes the ambiguity that let the prior "other" bucket carry both an AI ramp and a declining consumer business. We resolve that standing complaint and replace it with a narrower one, which is the absence of segment profitability.
3. Ferric Buys the Last Few Millimetres of the Power Tree
onsemi agreed to acquire Ferric for $600 million in cash plus up to $175 million tied to revenue milestones running through 2029, funded from the balance sheet and expected to close in the fourth quarter. Ferric builds integrated voltage regulators, which place the controller, switches, capacitors and magnetics in one package adjacent to the processor.
"As power requirements of AI processors continue to increase, IVR incorporates the complete power regulator, including controller, power switches, capacitors and magnetics in one thin package and places it closer to the GPU, reducing power losses and solving the power density problem at the core." — Hassane El-Khoury, President and CEO
The stated financial effect is deliberately back-loaded. Management raised its 2030 AI data center target by $500 million to more than $3 billion, attributed the entire increase to the acquisition, and declined to size 2027 or 2028 revenue. The deal is described as accretive to long-term growth, gross margin and EPS targets "at scale," with no meaningful near-term financial impact.
"We expect Ferric to be accretive to our growth and gross margin targets with EPS accretion at scale and no meaningful near-term financial impact." — Thad Trent, Executive Vice President and CFO
Assessment: Paying $600 million up front for a 50-person company with no disclosed revenue, against a target that sits four years out, is a venture-style allocation inside an industrial balance sheet. The strategic logic is sound, since the Vcore stage, at $24 billion, is among the largest application lines on onsemi's own addressable-market slide. We put nothing in our estimates for Ferric and treat the 2030 target increase as an option rather than a forecast.
4. Synaptics Became a Cash Deal, Which Swaps Dilution for Leverage
The transaction we reserved against three months ago no longer exists in its original form. On October 1 the parties replaced the all-stock agreement with an amended and restated merger agreement at $123.00 per share in cash, an aggregate value of approximately $5.7 billion against approximately $7 billion previously. The change followed an unsolicited competing proposal for Synaptics from an unnamed third party. The registration statement has been withdrawn, the transaction is no longer structured as a tax-free reorganization, the requirement to place a Synaptics director on onsemi's board has been removed, and four closing conditions tied to issuing onsemi stock have been eliminated.
Financing comes from cash on hand plus a commitment letter for up to $2.45 billion of senior secured term loan, with no financing condition on closing. Management now describes the transaction as immediately accretive to non-GAAP EPS rather than accretive within eighteen months, and says it has identified further value beyond the previously announced $200 million of run-rate synergies from revenue synergies and from insourcing part of Synaptics' production, to be realised after the first eighteen months. The timetable is unchanged at a close by mid-2027, and the United States Federal Trade Commission has approved the deal.
"We repurchased $20 million of shares in Q3 with activity limited by trading restrictions related to the Synaptics acquisition and renegotiations." — Thad Trent, Executive Vice President and CFO
For shareholders the arithmetic improves substantially. Our prior target reserved for roughly 53 million new shares, which cost about $0.29 of FY2027 valuation EPS on the estimates we published three months ago. That reserve is retired. Against it, onsemi takes on debt priced at something close to 6% when its existing $4.46 billion of borrowings cost about 1.25% annualized, because almost all of it is low-coupon convertible paper.
Assessment: Paying 19% less, in cash, after a competing bidder appeared is a good outcome for onsemi shareholders, and we credit management for the discipline. The risk did not disappear; it changed character. Dilution was a known, quantifiable haircut. Leverage is a smaller headline cost with a longer tail: a secured term loan carries covenants and a fixed amortisation obligation through a cycle that onsemi does not control, and it arrives alongside $802.8 million of convertible notes that come current in 2027. We keep this pillar at the same risk level with a different label.
5. The Margin Bridge Is Now Published, and Most of It Is Mechanical
The 190-basis-point sequential step to 41.2% is the fifth consecutive increase and the largest. More useful than the step itself is that management put numbers on the remaining path to the 53% target it set at the September Investor Day: about 450 basis points from utilization, 225 basis points from Fab Right, 400 basis points from mix and new products, and an unquantified tailwind from pricing already taken but not yet in the income statement.
"Non-GAAP gross margin expanded by 190 basis points quarter-over-quarter to 41.2% and non-GAAP earnings per share was $0.90, growing 4x faster than revenue growth year-over-year." — Thad Trent, Executive Vice President and CFO
The part of the utilization component already run is the one an investor can underwrite without a demand forecast: the fabs have run at 83% since the second quarter, while Q3 cost of revenue still reflects 77%. The rest needs higher loadings, which management ties to improving demand. Management said roughly 450 basis points of under-absorption sits in the Q3 result against a 77% starting point, and framed the practical ceiling at 92% to 93%, which reconciles to the published 25 to 30 basis points of margin per utilization point.
Assessment: This is the most investable disclosure of the quarter, and it is the reason we hold a 44.5% FY2027 gross margin rather than cut it when the sector de-rated. The fourth-quarter guide takes the midpoint to 42.5%, a further 130 basis points, with utilization flat. The vulnerability is the 400-basis-point mix component, which assumes the accretive new products keep taking share of revenue; if automotive volume returns in 2027 at the same time AI growth normalises, mix works against the bridge rather than for it.
6. Utilization Held Flat, and the Lag Is Doing the Work
Utilization was 83% in both the second and third quarters and is expected at similar levels in the fourth. That stability sits awkwardly beside a supply shortage, and management explained the apparent contradiction directly: the constraint is in specific lines rather than across the fab base, and what shipped this quarter reflects production started two quarters ago at 77%.
"Utilization remained flat at 83%. We expect similar levels in Q4 with utilization increasing as we move into 2027 in response to improving demand signals." — Thad Trent, Executive Vice President and CFO
Assessment: The two-quarter lag that worked against the margin in 2025 is now working for it, and it will keep working for two more quarters regardless of what demand does next. That makes the near-term margin forecast unusually low-risk. It also means a utilization increase taken in 2027 does not reach the income statement until the second half of that year, which is why we hold a 44.5% full-year average rather than extrapolating the recent quarterly cadence.
7. Capital Return Pauses While Capital Intensity Is Set to More Than Double
Share repurchases fell to $19.9 million from $344.8 million in the second quarter and $325.0 million a year ago, with nine-month buybacks of $710.4 million. Management attributed the near-halt to trading restrictions related to the Synaptics transaction and its renegotiation, and set out the forward priority plainly: use balance-sheet cash for Ferric, then focus on delevering after Synaptics closes, with buybacks opportunistic.
Capital intensity moves the same way. Third-quarter capital expenditure of $27.8 million was 1.6% of revenue and came in $12.2 million below the low end, and $17.2 million below the midpoint, of the $40 to $50 million guided three months ago. Full-year 2026 intensity is now expected at about 2% of revenue, against the "below 5%" framing from the second-quarter call. For 2027 management guided back to about 5%, in line with its long-term model.
"And the capacity that we talk about that will contribute to catching up in Q1, you can think about this as the capacity was -- is already behind us and the investment is behind us. It will just come online and deliver in -- think about it in 2027." — Hassane El-Khoury, President and CEO
Assessment: This is the pillar that moved against us. The capital-return discipline we have carried as on track since the fourth quarter of 2025 is now subordinated to two acquisitions and a delevering programme, and the 2027 capital-expenditure step from roughly 2% to 5% of revenue removes about $250 million of free cash flow at our revenue estimate. Neither change damages the earnings case, and the 2026 underspend is a timing benefit that must reverse. But an investor who owned onsemi partly for the $6 billion authorisation and the roughly 105% of first-half free cash flow returned should understand that the next eighteen months look different.
8. Replenishment Is Being Negotiated Now Rather Than Waited For
Customer restocking has been undated upside in this thesis since the initiation. This quarter the framing inverted. Rather than waiting for customers to decide to rebuild inventory, onsemi is sitting down with strategic automotive customers on business continuity plans, building the replenishment into a 2027 capacity model, and declining to ship it in the meantime.
Management's reasoning was that customers can see constraint spreading across power, analog, compute and memory, and are placing orders early to avoid being caught short. The demand, in management's phrase, is not perishable: the backlog is on the books and will be serviced from the first quarter of 2027.
Assessment: Converting optional upside into a scheduled, contracted catch-up is a material improvement in visibility, and it is the first time in six quarters that replenishment has had any shape at all. We still do not put a restocking bulge into our base case, because the orders are not quantified and because a book of early orders placed to avoid shortage is exactly the kind of demand that gets trimmed if the shortage eases. We treat it as the main source of upside to our 2027 automotive estimate.
9. Industrial's Strength Is the AI Cycle Wearing a Different Hat
Industrial beat a flat guide with 10.4% sequential growth, and management was precise about the source. The beat was not the broad industrial economy, although purchasing indices are improving and the broader industrial portion is now described as about 40% of the segment. The beat was the AI halo: energy infrastructure, grid equipment and storage built to feed data centers.
"Think about it from the grid to the AI data center. So everything outside the walls of the AI data center has been very strong. It came in better than we expected." — Hassane El-Khoury, President and CEO
Management also addressed whether any of it was channel refill, and gave an unusually concrete answer: it can see sell-through from its customers to their end customers, those customers have reported strong quarters, and the equipment is being installed on sites rather than stored.
Assessment: The datapoint is strong and the explanation is credible. The investment consequence is uncomfortable: the broader industrial economy is only about 40% of the segment by management's own description, so the majority sits in the specific verticals it named, led by energy infrastructure and the AI halo; industrial is therefore partly a second expression of the capital cycle that drives the AI data center line rather than a counterweight to it. On our 2027 estimates the two together are about 42% of revenue and the large majority of growth. That correlation is a reason to hold a lower multiple, not a reason to lower the estimates.
10. Fab Right Delivers Its First Closed Divestiture
The Philippines back-end facility sale closed on October 2 for $24.2 million of net cash proceeds against $24.3 million of net assets disposed, producing a $3.1 million gain recorded in other income and excluded from adjusted results. The Mountain Top, Pennsylvania fab remains on track for a close by the end of 2027, with onsemi continuing to operate it through 2027 to transition production; as of quarter end it did not meet held-for-sale criteria.
"In Q3, we closed the previously announced divestiture of a back-end facility in the Philippines, and we remain on track to close our divestiture of the Mountain Top fab exiting 2027. We are operating with a more focused and efficient footprint, creating additional opportunity for gross margin expansion over time." — Thad Trent, Executive Vice President and CFO
Assessment: The programme is executing to its stated schedule, which is more than can be said for most multi-year manufacturing restructurings. Quarterly restructuring charges fell to $4.6 million from $41.2 million, so the cash and reported-earnings drag from this programme is largely behind the company. The 225 basis points of Fab Right margin benefit still sits mostly in 2028, since the larger of the two sites will not transfer until the end of 2027.
11. Cash Generation Is Excellent, and Part of It Is Borrowed Time
Third-quarter cash from operations of $516.4 million grew 23.3% and free cash flow of $488.6 million grew 31.2% to 28.4% of revenue. Trailing-twelve-month free cash flow is $1,616.6 million on $6,369.2 million of revenue, a 25.4% margin. Cash and short-term investments reached $4,367.7 million against $4,463.0 million of total debt, leaving net debt of roughly $95 million.
"We believe the recovery is gaining momentum, and we are in a fundamentally stronger position than in prior quarters. We have more content, a broader market opportunity and substantially greater earnings power." — Thad Trent, Executive Vice President and CFO
Two qualifiers belong on that free cash flow figure. The first is the capital-expenditure underspend already discussed, worth roughly $17 million against the quarter's own guidance midpoint and considerably more against the long-term model. The second is working capital, which contributed $38.5 million in the quarter but drained $118.2 million over nine months, so the quarterly figure is flattered relative to the year.
Assessment: Cash conversion is genuinely strong and is the reason a $5.7 billion acquisition is financeable without equity. We model free cash flow down in 2027 on the capital-expenditure normalisation and the interest cost, not because the business deteriorates. At 25% of revenue on a trailing basis, onsemi can service the term loan and delever on the schedule management described, which is the test that matters for the leverage pillar.
Guidance & Outlook
| Metric | Q3 2026 actual | Q4 2026 low | Q4 2026 high | Midpoint | Change |
|---|---|---|---|---|---|
| Revenue | $1,722.3 | $1,715 | $1,815 | $1,765 | +2.5% QoQ, above seasonal |
| Gross margin, GAAP | 41.4% | 41.4% | 43.4% | 42.4% | +100 bp |
| Gross margin, non-GAAP | 41.2% | 41.5% | 43.5% | 42.5% | +130 bp |
| Operating expenses, non-GAAP | $309.5 | $303 | $318 | $310.5 | Flat; includes about $34 of stock compensation |
| Other income, non-GAAP | $17.3 | n/a | n/a | +$27 | Favorable |
| EPS, GAAP | $0.79 | $0.88 | $1.00 | $0.94 | +19.0% |
| EPS, non-GAAP | $0.90 | $0.95 | $1.07 | $1.01 | +12.2% QoQ |
| Diluted shares, non-GAAP | 392.2 | n/a | n/a | 392 | Flat |
| Tax rate, non-GAAP | 15% | n/a | n/a | 15% | Maintained |
| Capital expenditures | $27.8 | $40 | $50 | $45 | About 2% of 2026 revenue |
The guide is above the Street on both lines. The $1,765 million revenue midpoint sits 1.5% above the $1,739 million consensus, and the $1.01 EPS midpoint is 5.2% above the $0.96 figure that was polled. The composition is unusual: revenue grows 2.5% sequentially while the largest end market declines high single digits, so the entire increment comes from industrial, AI data center and a lumpy externally manufactured programme in the other bucket. Gross margin does the rest of the work, adding 130 basis points at the midpoint with utilization flat.
"The fourth quarter outlook reflects continued demand momentum and another step forward in our model. Based on our current visibility, we expect this momentum to continue into 2027 with continued revenue growth and gross margin expansion as we scale our differentiated products and realize the benefits of our more efficient operating model." — Thad Trent, Executive Vice President and CFO
Implied quarter-over-quarter ramp: Applying the four stated end-market qualifiers to the third-quarter bases produces roughly $1,780 to $1,795 million of revenue, which sits above the $1,765 million midpoint and inside the upper half of the guided range. The arithmetic holds even at the steep end of a high-single-digit automotive decline. That is a mild tell that management expects to finish above the midpoint, and it is consistent with the pattern of the last three quarters, each of which landed above the midpoint.
Street at: Consensus for the fourth quarter was $1,739 million and $0.96 before the print, so the guide moves estimates up on both lines. For the full year, nine-month revenue of $4,839.1 million plus the midpoint gives $6,604 million, 10.2% above the $5,995.4 million we calculate for 2025 from the release's nine-month and trailing-twelve-month tables. Full-year adjusted EPS at the midpoint is $3.29 against the nine-month figure of $2.28.
Guidance style: Conservative at the midpoint and reliable at the range. We checked the EPS arithmetic against the stated inputs: $1,765 million at 42.5% gives $750 million of gross profit, less $310.5 million of operating expense and plus $27 million of other income gives $467 million pre-tax, which at 15% tax over 392 million shares is $1.01. The guide is internally consistent, and the only soft element is the breadth of the margin range at 200 basis points.
Analyst Q&A Highlights
Whether Withholding Automotive Supply Costs Share
The opening question, and the dominant line of questioning on the call, challenged the characterisation of automotive as strategic while the company guided it down for the fourth quarter after two sequential declines. The exchange produced the most specific forward commitment management made: the recovery is dated to the first quarter, the mechanism is identified as supply already installed, and the claim that no share is being lost rests on customers agreeing to the deferral rather than on management's assertion alone.
Q: "Hassane, you have been sort of, I guess, rightly deemphasizing automotive in Q2 and Q3, and I think even in Q4 with this signaling of, I think, high single digit from what I heard. But then you suggested that next year, auto would go back to growing per your long-term target, which I think is up whatever, 9% or 10%. So I'm curious, does underserving that market have any market share implications longer term? And then where does the extra supply come from to start regrowing in that market from next year?"
— Vivek Arya, Bank of America Securities
A: "What you're seeing today or this quarter and next quarter, what we've hinted at is short term in nature and it's supply-driven, not strategically driven. What I mean by that is we're not losing share because we're working closely with the OEMs, as I mentioned in my prepared remarks, we are supporting all of their end demand needs. What we are holding back, again, with conversations with them and alignment with them is their replenishment that they would have liked to see this -- during the quarters. We will resume shipment in Q1 as the supply comes online."
— Hassane El-Khoury, President and CEO
Assessment: Management answered directly and corrected the premise rather than deflecting, which is the right posture when the premise is wrong. The answer is also falsifiable in a single quarter, which is unusual and welcome. The residual risk is that a customer which has been told to wait for three quarters has had three quarters to qualify a second source.
Where the Supply Constraint Actually Sits If Utilization Is Flat
The sharpest technical challenge of the call noted that a company claiming to be supply-limited would normally be running its factories harder, and asked whether the constraint was therefore external. The answer separated installed capacity from finished-goods timing and conceded the geography of the problem, which is narrow rather than company-wide.
Q: "And then Thad, second one for you is just you talked about the automotive revenue just being a bit lighter because of supply relocation, timing, phasing of customer deliveries. But it looks like the utilization is flat. So are you short external components? Because I would expect if you're not able to get more supply, like the utilization would continue to move up. You're saying kind of that's more next year. Maybe just describe the dynamic, like where are you seeing the constraints?"
— Thomas O'Malley, Barclays
A: "Keep in mind that what you're seeing come out of the line today was utilization from 2 quarters ago when we were running at 77%. We ran at 83% last quarter. We're running at 83% going into Q4. So this is us catching up to that supply as you get cycle time through the fab in the back end. So that's where we have supply coming online. The supply constraints are in specific lines in certain areas. So it's not across the board or obviously, we'd be taking utilization up further."
— Thad Trent, Executive Vice President and CFO
Assessment: The answer is coherent and reconciles the two facts, but it also narrows the story. A constraint in specific lines is a mix problem, not a capacity problem, which supports the low-capital-intensity case and simultaneously means the automotive catch-up depends on a handful of processes rather than on general fab loading. It also explains why a 5% capital-expenditure year in 2027 is described as debottlenecking rather than expansion.
The Walk From 41% to a 53% Gross Margin Target
A recurring line of questioning sought to convert the long-term margin target into something an investor could model. The response produced the quarter's most useful disclosure, quantifying the under-absorption still embedded in reported cost of revenue and attaching figures to each remaining component of the bridge.
Q: "So I just wanted to ask specifically about the utilization lever in the gross margin bridge. About half of the path to that 53% gross margin target you guys set out comes from utilization. So just given the restructuring and also the changes to depreciation, how should we sort of think about the incremental gross margin on each dollar of revenue from here on out?"
— Gaurav Saini (on behalf of Gary Mobley), StoneX
A: "Well, look, the way you should think about the Q3 results is there's about 450 basis points of underutilization charges. Now again, you got to go back to the Q1 utilization of 77% for that. So if you take that, getting up to what is kind of peaking out for us, 92%, 93% somewhere in there, you can do that math on that 25 basis points to 30 basis points and you get to that 450 basis points. So to do the walk for everybody, getting from here to the 53%, you've got about 450 basis points of utilization improvement. You've got 225 basis points of Fab Right. I talked about the activities we're taking there. You've got another 400 basis points of mix and the new products kicking in, and I've already kind of talked about we're starting to see that hit the P&L, and we'll start to see a lot more of it next year."
— Thad Trent, Executive Vice President and CFO
Assessment: This is the exchange that justifies holding a 44.5% full-year gross margin for 2027 while the sector multiple compresses. Fab Right is demand-independent and already underway, and so is the part of the utilization lever already run, the step from 77% to 83%; the rest of the utilization lever needs higher loadings. The third, 400 basis points of mix, is the one that requires the AI and new-product revenue to keep outgrowing the base, and it is the component we would cut first if automotive volume returns faster than AI grows.
Funding Two Acquisitions While Maintaining the Return Policy
The only question of the call on capital returns asked how two cash transactions and new borrowing would coexist with capital returns. The answer established the post-close hierarchy explicitly and in order, and it is the clearest signal in the quarter that the buyback is not coming back soon.
Q: "And then as a follow-up, as we think about just the cash outlay that's kind of in front of you in terms of this acquisition and then also Synaptics and you're raising some debt for that as well. But how do we think about just the use of free cash flow and capital returns in that vein?"
— Joseph Quatrochi, Wells Fargo
A: "So we'll be using cash off the balance sheet from the Ferric acquisition. As we close the Synaptics acquisition, we will be focused on delevering. That will be our top priority. That's obviously after investing in the business as well. But we'll focus there, and we'll be opportunistic on buying back shares over time. But the intent here is to delever quickly."
— Thad Trent, Executive Vice President and CFO
Assessment: The answer is candid and the priority order is the right one for a levered acquirer, but it is a change of policy stated without being framed as one. Capital return goes from first call on free cash flow to third, behind reinvestment and debt reduction, and the questioner did not press on the duration. We move the capital-return pillar to at risk on the strength of this answer rather than on the quarter's $19.9 million of repurchases, which has a transaction-restriction explanation.
Whether Customers Are Finally Willing to Carry Inventory
The answer to this question on each of the last four calls was that restocking had not begun. This quarter the answer changed in substance, describing formal planning conversations with strategic customers and a contracted catch-up rather than a hoped-for one.
Q: "Hassane, I was wondering if you could maybe characterize the OEM customers' behavior in terms of their willingness or interest in restocking. I think you referred in your prepared comments to automotive customers are not supporting restocking but only supporting increases in end demand. Can you maybe sort of broadly describe what's happening with those customers and then the broader OEM customer set with respect to taking on more inventory in the balance sheet?"
— James Schneider, Goldman Sachs
A: "And they're looking at it as they see the writing on the wall. AI and industrial is recovering, and they're trying to place demand, so they're not caught in traffic as far as capacity and capacity constraint because they see it across the board, not just on power semis that we see, but across analog, compute and of course, the memory. So they're looking at their 2027 plans, and they want to think about it as have a what I would call a business continuity plan or BCP as known in the industry. Those are strategic plans that we do with our strategic OEMs where we sit down and figure out kind of their capacity and their pull rate, and we build ahead of that."
— Hassane El-Khoury, President and CEO
Assessment: The exchange converts a six-quarter open question into a scheduled event, and the stated motivation, fear of being shut out of constrained capacity across several component categories, is consistent with what other suppliers are reporting. It is also the kind of demand that evaporates if the constraint eases, so we keep it as upside rather than base case. The thesis pillar on the cyclical trough stays contained rather than resolved on this basis.
Capacity and Capital Intensity Heading Into 2027
A follow-on line of questioning tested whether the current shortfall reflects an investment gap that would require a capital-expenditure step up. The answer confirmed the catch-up framing, stated that capacity is adequate even above the current outlook, and put a number on next year's intensity.
Q: "And correct me if I'm wrong on this one, but it sounds like some of the supply constraints you're facing now is just a function of not having started the wafers earlier in the year. And so now you're catching up. So I guess if that's the case, where do you stand with regard to capacity as you go into next year? Do you think you have the capacity to be able to meet what you see for the market next year? And then from that, what sort of CapEx should we be expecting as you head into next year?"
— Christopher Caso, Wolfe Research
A: "I think that's a fair characteristic, that this is a catch-up, right? And that's why Hassane talked about this as being temporary. We'll catch up in Q1. So we do have the capacity that we need based on the outlook that we have today and even some upside to that outlook. So I don't think we're going to be anywhere near capping out our capacity as we go into 2027. This year, our CapEx intensity is about 2% of revenue. I expect it to be about 5%, in line with our long-term model next year."
— Thad Trent, Executive Vice President and CFO
Assessment: Confirming that 2027 growth does not need new capacity is the strongest free-cash-flow statement on the call, and it supports the low-capital-intensity leg of the franchise argument. The 2% to 5% intensity step still removes roughly $250 million of free cash flow at our 2027 revenue estimate, and it arrives in the same year as the Synaptics cash outflow. That is a timing problem rather than a quality problem, but it is a real one.
What Makes Ferric the Right Integrated-Regulator Asset
The closing question challenged the premise that any one integrated voltage regulator design wins, given how many companies are positioning in the category. The answer was unusually technical and specific, which is informative in itself: management defended the purchase on measurable engineering attributes rather than on market size.
Q: "And as my follow-up on Ferric, so I think there's a lot of companies positioning in IVR right now. And I was just hoping you could share with us a little bit what's the uniqueness that Ferric brings? I know in your presentation, you talked about better magnetics, smaller footprint and so on and so forth. But yes, anything else that you could share with us why their technology is sort of the winning formula versus what we've seen from other companies?"
— Tore Svanberg, Stifel
A: "From an electrical perspective, we know Ferric has a superior transient performance than anything else in the market. So that covers the electrical. That's important because of the dynamic load that high-end processors are driving, where you got a lot of transient and the voltages have to regulate or respond to it. So that provides a superior electrical performance versus peers or competitors. Then you have the mechanical side of it, which is the power density, where there are other solutions that are not necessarily magnetic. There are other innovative solutions out there, but the solution from Ferric with electro and magnetic components of it with the controller, that gives you the most dense from an area and the most compact from a z-height. The -- you can think about everybody's z-height is in the 2.5 millimeter. Ferric blows that away."
— Hassane El-Khoury, President and CEO
Assessment: A defence built on transient response, area density and package height is a better answer than a market-size slide, and the claim that Ferric is engaged with all the hyperscalers is checkable over time through design-win disclosure. What the exchange did not produce is any revenue, timeline or margin figure, and management declined to provide one when asked directly earlier in the call. We therefore carry the acquisition at zero in our estimates.
What They're NOT Saying
- No Ferric revenue, no 2027 contribution, no multiple paid. Management explicitly declined to guide Ferric revenue for 2027 or 2028 and described the financial effect as accretive "at scale" with no meaningful near-term impact. A $600 million base price plus up to $175 million of milestone payments through 2029 is being justified entirely by a 2030 target increase. Without a revenue base there is no way to assess whether the price was reasonable, and the only measurable commitment is the $500 million addition to a target four years out.
- No combined accretion bridge for Synaptics, and the measurement basis matters more than usual. onsemi expenses share-based compensation inside its non-GAAP results, and the fourth-quarter guide names about $42 million of it. Synaptics excludes share-based compensation from its non-GAAP results, roughly $0.90 per share per quarter by its own disclosure. An "immediately accretive" claim is comfortable on Synaptics' reported basis, where approximately $186 million of fiscal 2026 adjusted net income roughly offsets our estimate of the financing cost. On onsemi's stricter basis the acquired earnings are closer to $62 million, and accretion then requires most of the $200 million of run-rate synergies. onsemi has not published which basis it used.
- No explanation of why the seller accepted 19% less. The amended agreement followed an unsolicited competing proposal for Synaptics from a party identified only as "Party A" in the withdrawn registration statement. The counterparty settled for approximately $5.7 billion in cash against approximately $7 billion in stock, and the call contained no discussion of the competing bid, the negotiation or the valuation reset. The company is entitled to decline, and the outcome favours onsemi shareholders. But nobody asked the obvious question of what the competing proposal revealed about the asset.
- No comment on a long-term supply agreement backlog that fell $600 million. Committed obligations under long-term supply agreements declined from approximately $6.1 billion at the second-quarter end to approximately $5.5 billion at the third-quarter end, with the share expected within twelve months rising from 37% to 39%. The decline is consistent with shipping against existing agreements faster than signing new ones, which is benign at this point in a ramp. It is nonetheless the one forward-commitment metric that moved the wrong way in a quarter that was otherwise about order strength, and it was not mentioned on the call or raised in the question period.
- No quantification of the deferred automotive shipments. Three quarters of withheld replenishment are described qualitatively and dated to a first-quarter catch-up, but the dollar amount is undisclosed. That makes it impossible to separate how much of any first-quarter automotive recovery is genuine demand growth from how much is a one-time catch-up that will not repeat in the second quarter.
- No 2027 guide, only direction, and no buyback commitment attached to it. Management offered revenue growth and gross margin expansion for 2027 without ranges, reiterated the AI doubling, and put automotive back on the long-term model. It did not quantify any of it, and it did not say when or whether repurchases resume beyond "opportunistic" and a stated focus on delevering after the Synaptics close. For a company that returned roughly 105% of its first-half free cash flow, that is a meaningful silence.
Market Reaction
- Pre-print setup: ON closed November 2 at $70.11 after falling 2.2% that session. The stock entered the print up 29.5% year to date and 40.0% over twelve months, but down 17.4% over the prior thirty days from $84.89 on October 2. Its preceding 52-week closing range was $44.90 to $133.93, and the pre-print close sat 47.7% below the 52-week high. The S&P 500 was up 10.2% year to date over the same period.
- After-hours move: Shares fell roughly 4% in the extended session on November 2 despite the beat and the above-consensus guide, a move we attribute to the automotive guidance and the supply-reallocation framing.
- Reaction session, November 3: Shares opened at $71.99, a 2.7% gap higher, traded between $70.06 and $74.00, and closed at $72.60, up 3.55% or $2.49. Volume was 15.5 million against a thirty-day average of 10.6 million, or 1.5 times normal.
- Market comparison: The S&P 500 gained 0.4% during the reaction session, so ON outperformed by about 3.2 percentage points.
- Sessions since: Shares gave back most of the gain on November 4, closing at $71.24, down 1.87%, on in-line volume. That $71.24 close is the last completed session and the price our valuation anchors to. A broad semiconductor drawdown is under way on the publication session, with the sector index and every analog and power peer down sharply; those levels are intraday rather than closes, and the move is sector-wide rather than specific to this print.
The two-stage reaction is the informative part. The immediate after-hours verdict was negative, which is what the headline end-market table supports: the largest segment shrank and is guided to shrink again. By the next morning the market had re-read the quarter, and the open gapped 2.7% higher on the above-consensus guide, the raised AI figure and the Ferric announcement. Volume at 1.5 times normal on a session that closed up 3.55% is consistent with genuine repositioning rather than a short-covering bounce.
What the session does not settle is the valuation question. The stock has fallen 46.8% from its 52-week high while our FY2027 earnings estimate has gone up, which means the entire decline is multiple compression rather than estimate revision. It also means a beat-and-raise quarter produced a 3.55% move rather than a re-rating. Two readings are available: either the market is discounting an AI capital cycle it does not trust, or the June price was wrong and this is the correction. We lean toward the first, because the operating evidence has improved in each of the five quarters since the margin inflection began, and because the stock is now 11.8% below the $80.78 level at which we upgraded three months ago, despite a stronger set of numbers.
Street Perspective
Debate: Is the Automotive Air Pocket a Choice or a Symptom?
Bull view: A supplier that turns away replenishment orders to serve better-priced demand is exercising pricing power, and the fact that end demand is fully served means no vehicle programme is short. Content per vehicle keeps rising through flat unit volumes, and the catch-up is contracted for the first quarter.
Bear view: Three consecutive quarters of no automotive growth, in a business that is still 45% automotive, is a long time to ask customers to wait. The explanation has conveniently shifted from weak automotive demand to deliberate automotive deferral without the revenue line changing, and a deferral that keeps being extended starts to look like a forecast.
Our take: The bull case has the better evidence, because nine-month automotive revenue grew 3.2% and the deferral has a named mechanism and a date. The bear case has the better framing of the risk, which is that this claim is unverifiable until February. We accept the characterisation and treat the first-quarter automotive print as the single test that could break the pillar.
Debate: Does a 53% Margin Target Deserve a Growth Multiple?
Bull view: Two of the three components of the published bridge, utilization recovery and the manufacturing programme, do not require demand growth, and the charge is already visible in reported cost of revenue. Five consecutive quarterly increases with an accelerating cadence is a track record rather than a projection, and gross profit is targeted to compound 1.6 times faster than revenue.
Bear view: The 400-basis-point mix component assumes the favourable product shift continues indefinitely, which requires AI to keep outgrowing the base. The bridge also runs to 2030, and investors have been offered multi-year semiconductor margin targets before. Paying a growth multiple for a target four years out, in a business that just cut its largest end market, is the error the last five months of multiple compression already corrected.
Our take: The near-term portion of the bridge is credible enough to underwrite and the long-term portion is not something we pay for. We carry 44.5% for 2027, which captures the mechanical components and roughly half of the mix benefit, and we cut the multiple rather than the margin. A business whose margin expansion is mostly mechanical deserves a reliability premium; it does not deserve a growth multiple when the growth is concentrated in one capital cycle.
Debate: Was the All-Cash Synaptics Pivot a Win or a Stretch?
Bull view: Paying $1.3 billion less, in cash, with a competing bidder in the room, is as good an outcome as an acquirer gets. Shareholders avoid issuing about 53 million shares near a cyclical low in the share price, the transaction becomes immediately accretive, and the financing is committed with no closing condition attached to it.
Bear view: A company with roughly $95 million of net debt is taking on up to $2.45 billion of secured term debt to buy a business with about $1.2 billion of revenue, while simultaneously closing two smaller acquisitions, pausing its buyback and more than doubling capital intensity. The accretion claim rests on an unpublished bridge and on a definition of adjusted earnings that the target and the acquirer apply differently.
Our take: The renegotiation is a clear win on price and structure, and we remove the share reserve from our target without hesitation. We do not remove the pillar. Three concurrent integrations, a step change in funding cost from roughly 1.25% to something near 6%, and a convertible maturity in the same year is a lot of simultaneous execution for a management team that is also ramping a new end market. The risk is smaller in dollars and longer in duration than the dilution it replaced.
Our Estimates & Valuation
| Estimate | Prior Q2 recap | Current view | Reason |
|---|---|---|---|
| FY26 revenue | About $6.62B | About $6.60B | Q3 beat offset by a Q4 midpoint below our $1.80B assumption |
| FY26 non-GAAP gross margin | About 40.6% | About 40.5% | Q3 at 41.2%; Q4 guided to 42.5% |
| FY26 non-GAAP EPS | About $3.30 | About $3.29 | Nine months $2.28 plus the $1.01 midpoint |
| FY27 revenue | $7.60B | $7.70B | AI doubles off a $620M base rather than a "greater than $500M" base |
| FY27 non-GAAP gross margin | 44.5% | 44.5% | Unchanged; faster near-term cadence, same 2027 exit assumption |
| FY27 non-GAAP operating expense | $1,320M | $1,340M | Ferric and Silver Atena added to the cost base |
| FY27 adjusted shares | 391M | 393M | Buyback subordinated to delevering; no Synaptics issuance |
| FY27 EPS used in valuation | About $4.31 | About $4.73 | Share reserve retired; transaction now cash and debt funded |
| Valuation multiple | 22x | 20x | Leverage added; AI-concentrated growth mix; group multiple compressed |
| 12-month base value | About $95 | About $95 | Same answer, different composition |
FY2026 bridge
Nine-month revenue of $4,839.1 million plus the $1,765 million fourth-quarter midpoint gives $6,604 million, 10.2% above the $5,995.4 million we calculate for 2025. Weighting the nine-month adjusted gross profit of $1,923.5 million with the fourth quarter at 42.5% gives a full-year adjusted gross margin of about 40.5%, marginally ahead of the 40.3% shown for 2026 at the September Investor Day. Adjusted EPS is the nine-month $2.28 plus the $1.01 midpoint, or $3.29. Our forecast from three months ago was $6.62 billion and $3.30; the quarter moved the composition without moving the year.
FY2027 revenue bridge
| End market | FY26 estimate | Growth applied | FY27 estimate | Basis for the growth rate |
|---|---|---|---|---|
| Automotive | About $3,060 | +9% | About $3,340 | Management expects 2027 automotive growth "in line with our long-term target"; the Investor Day set that at 9%, high single digits above production |
| Industrial | About $1,820 | +10% | About $2,005 | Investor Day long-term rate; AI halo and energy infrastructure |
| AI data center | $620 | +100% | $1,240 | "We still expect our AI data center revenue to double over 2026" |
| Other, excluding AI | About $1,115 | +1% | About $1,125 | Management guided this bucket back to normal historical rates; Investor Day long-term rate 1% |
| Total | About $6,615 | +16.5% | About $7,700 |
End-market columns are rounded and the FY2026 column sums about $11 million above our $6,604 million revenue estimate, which is built from nine-month actuals plus the fourth-quarter guidance midpoint. The 16.5% growth rate sits above the 12% to 14% long-term range because the AI doubling lands in a single year off a base that just grew 140%. AI data center supplies about 57% of the incremental revenue and moves from 9.4% of 2026 revenue to about 16% of 2027 revenue, which is the single most important change to the shape of this business since we initiated.
FY2027 earnings bridge
Revenue of $7,700 million at a 44.5% adjusted gross margin produces $3,426 million of gross profit. Subtracting $1,340 million of adjusted operating expense gives $2,086 million of adjusted operating income, a 27.1% margin against the 23.2% reported this quarter. Adding $100 million of net other income, applying a 15% tax rate and dividing by 393 million adjusted shares produces approximately $4.73 of adjusted EPS. A one-percentage-point change in the 2027 gross margin moves adjusted EPS by about $0.17, which is why mix remains the central sensitivity.
How we treat the transaction
Our prior target reserved for roughly 53 million shares issued at a mid-2027 close, which reduced the FY2027 valuation figure from about $4.60 to about $4.31. The amended agreement eliminates that issuance entirely, so the reserve comes off. In its place sits a financing cost we estimate at approximately $212 million annually before tax: interest on $2.45 billion of term debt at roughly 6%, plus forgone yield on approximately $3.25 billion of balance-sheet cash at roughly 2%. Set against that, Synaptics earned $185.9 million of adjusted net income in its fiscal year ended June 2026 on approximately $1.2 billion of revenue at a 53.7% adjusted gross margin.
Measured on the target's own reporting basis, acquired earnings and financing cost roughly offset before synergies, which is consistent with management's accretion claim. Measured on onsemi's stricter basis, where share-based compensation is an expense, acquired earnings fall to roughly $62 million and accretion requires most of the $200 million of run-rate synergies. Because the close is a half-year event in 2027, the net effect on that year is small under either reading, and we therefore value FY2027 on the standalone figure with no credit for acquired earnings and no charge for the financing. We capitalise none of the synergies, consistent with the treatment we have carried since the transaction was announced.
Multiple and horizon
We value FY2027 earnings at 20 times, down from the 22 times used three months ago, for a target of approximately $95 by November 2027. Three things justify the lower multiple. The balance sheet moves from roughly $95 million of net debt to something near $5 billion of pro forma net debt at close, which on our estimates is roughly two times forward adjusted operating income plus depreciation before any acquired contribution. The revenue mix becomes materially more concentrated in a single high-beta capital cycle. And the sector multiple has compressed, which is evidence about the cost of capital rather than noise. The target is unchanged in dollars because the removal of the share reserve offsets the lower multiple almost exactly, which is a coincidence worth stating plainly rather than dressing up.
| 12-month scenario | FY27 revenue | FY27 gross margin | FY27 adjusted EPS | P/E | Value | Return from $71.24 |
|---|---|---|---|---|---|---|
| Base: AI doubles, automotive catches up, margin bridge delivers | $7,700M | 44.5% | About $4.73 | 20x | About $95 | +33.4% |
| Bull: faster AI ramp, mix holds, automotive replenishment lands | $7,900M | 46.0% | About $5.15 | 23x | About $118 | +65.6% |
| Bear: AI digestion, automotive stays flat, mix reverses | $7,000M | 42.0% | About $3.57 | 15x | About $54 | −24.2% |
At the $71.24 close the stock trades at 21.7 times our 2026 estimate and 15.1 times our 2027 estimate. The bear case applies the multiple the market is paying today to materially lower earnings, which is the right construction of downside when a stock has already de-rated: the loss comes from the estimates, not from further compression. We assume no dividend, so price return equals modelled total return. The 33.4% base-case return compares with our 8% twelve-month planning assumption for the S&P 500, and the ratio of upside to downside is better than 1.3 to 1 on absolute magnitudes.
Model Update Needed
| Item | Current model | Suggested change | Reason |
|---|---|---|---|
| FY27 revenue growth | +14.9% | +16.5% | AI data center doubles off $620M, not a "greater than $500M" base |
| FY27 AI data center revenue | Not separately modelled | $1,240M, 16% of revenue | New reported segment; the only end market with a company doubling commitment |
| FY27 automotive revenue | Recovery assumption | About $3,340M, +9% | Management put 2027 automotive back on the long-term model with a Q1 catch-up |
| FY27 operating expense | $1,320M | $1,340M | Ferric and Silver Atena cost bases added |
| FY27 capital expenditure | Mid single digit percent of revenue | 5.0% of revenue, about $385M | Explicitly guided; up from about 2% in 2026 |
| FY27 adjusted share count | 391M standalone; about 418M weighted with a 53M issuance reserve | 393M, no issuance | All-cash amended agreement; buyback subordinated to delevering |
| FY27 net other income | +$60M | +$100M standalone | Q4 2026 guide of +$27M per quarter on the pre-transaction balance sheet |
| Pro forma net debt at close | Not modelled | About $5B, roughly 2x forward EBITDA | $5.7B cash consideration funded with up to $2.45B of term debt |
| Blended cost of debt | About 1.25% | Blended toward 3% | Low-coupon convertibles supplemented by a secured term loan near 6% |
| Ferric contribution | n/a | Zero in 2027 and 2028 | No revenue, timeline or margin disclosed; management declined to guide |
Valuation impact: The twelve-month target is unchanged at approximately $95, with the composition rebuilt. The FY2027 valuation figure rises $0.42 to about $4.73, of which retiring the share reserve is the larger part and the revenue and operating-expense revisions the remainder. At the prior 22 times multiple that increase would have added about $9 to the target; cutting the multiple to 20 times removes about $9. The forward return improves from 17.6% to 33.4% almost entirely because the share price fell 11.8% since the upgrade while the earnings estimate rose.
Thesis Scorecard Post-Earnings
| Standing thesis point | Status / movement | What this quarter showed |
|---|---|---|
| Bull 1: Intelligent-power franchise | ON TRACK → ON TRACK, strengthened | AI data center $180.2M and now a reported segment; 2026 outlook raised to about $620M; 2030 target raised to more than $3B; intelligent power revenue +15.9%, about 4.8 points faster than the company |
| Bull 2: Margin transformation | ON TRACK → ON TRACK, strengthened | 41.2% adjusted, +190 bp and the fifth consecutive increase; Q4 guided to 42.5%; the remaining path to 53% quantified at 450 bp utilization, 225 bp Fab Right, 400 bp mix |
| Bull 3: Capital-return discipline | ON TRACK → AT RISK | Repurchases $19.9M against $344.8M in Q2; delevering named the top post-close priority with buybacks "opportunistic"; 2027 capital intensity guided from about 2% to about 5% of revenue |
| Bear 1: Cyclical trough | CONTAINED → CONTAINED | Industrial +10.4% sequentially on end demand rather than refill; automotive down for a second quarter by choice and guided down again; replenishment contracted for Q1 2027 but unquantified |
| Bear 2: Margin pinned by under-absorption | CONTAINED → CONTAINED, improving | About 450 bp of under-absorption still in the Q3 cost base; utilization flat at 83% in Q3 and Q4, rising in 2027; the two-quarter lag now works in favour of reported margin |
| Bear 3: Self-inflicted shrinking base | CONTAINED → CONTAINED | No further exits discussed; Philippines divestiture closed October 2 for $24.2M net; Mountain Top on track for end-2027; quarterly restructuring charges down to $4.6M from $41.2M |
| Bear 4: Valuation / melt-up risk | CONTAINED → CONTAINED | 15.1x our FY27 estimate at $71.24, 46.8% below the 52-week high, with estimates higher than when the stock peaked; the risk is now that the market is right, not that the price is extended |
| Bear 5: Acquisition and leverage risk (previously dilution) | EMERGING → EMERGING, re-characterised | Share issuance eliminated at a $1.3B lower price; replaced by $5.7B of cash consideration, up to $2.45B of secured term debt, three concurrent integrations, a 2027 convertible maturity and an unpublished accretion bridge |
Overall: The thesis strengthened on operations and on price, and weakened in one place. Both bull pillars that carry the earnings case delivered more than they promised three months ago, and the valuation condition that triggered the Q1 downgrade is now comfortably behind us. The capital-return pillar moves to at risk, not because management did anything wrong, but because the money is going somewhere else for the next eighteen months. The fifth pillar changed character rather than size: dilution retired, leverage arrived.
Action: Maintain Outperform with a twelve-month target of approximately $95 and conviction of 6 out of 10. Conviction does not rise despite the better numbers, because two of the three things that would raise it are unresolved: the first-quarter 2027 automotive print that tests the no-share-loss claim, and a published combined earnings bridge for Synaptics on onsemi's own definition of adjusted earnings. The third test is whether the 400-basis-point mix component of the margin bridge survives automotive volume returning. A reversal in the sequential margin trajectory, an automotive first quarter that does not recover, or financing terms materially worse than our 6% assumption would each warrant reassessment. Faster Ferric commercialisation or a restocking bulge beyond the contracted catch-up would each be upside we do not currently carry.