The Inflection Arrives: First Year-Over-Year Growth Guide in Three Years, Margins Expand Through Seasonality, and AI Rack Content Doubles — Upgrading to Outperform

Results vs. Consensus

Q4 2025 Scorecard

Full-Year 2025 Summary

Quarter-Over-Quarter Comparison

Quality of Beat

Revenue: Q4 revenue of $1,530M landed exactly at the guide midpoint and a hair above consensus, a 1.3% sequential decline that management correctly framed as normal Q4 seasonality. The composition was healthier than the headline: automotive grew 1% sequentially to $798M with inventory digestion "largely behind us," industrial grew 4% to $442M (and 6% year-over-year, the first year-over-year growth in eight quarters), and the year-over-year decline narrowed. The soft spot was the "Other" bucket (down 14% sequentially), but that was normal Q4 seasonality plus ~$40M of non-core exits, with AI data center actually growing within it.

Margins: The 38.2% non-GAAP gross margin (above the 38.0% midpoint) is the most encouraging margin print of our coverage, not for its level but for its mechanism: it expanded even as utilization stepped down to 68% (a planned, seasonal reduction), because FabRite cost actions are beginning to flow through. Under-absorption charges fell to ~700bps from ~900bps. Critically, the Q1 2026 guide expands margin a further 30bps in a seasonally down quarter, confirming that the structural self-help lever is real and additive to the eventual utilization recovery.

EPS and cash: Non-GAAP EPS of $0.64 beat consensus by a nickel on operating leverage (operating margin 19.8%), a ~16% tax rate, and a reduced share count (402M after a $450M Q4 buyback). The headline cash figure is the full-year story: $1.4B of free cash flow at a record 24% margin, up 17% year-over-year, on capex that fell to 4.5% of revenue as the heavy capacity investments concluded. onsemi returned essentially all of it and authorized a new $6B buyback. This is a structurally cash-generative business throwing off cash at the bottom of its cycle.

Segment Performance

End-Market Revenue — Q4 2025

Business-Unit Revenue — Q4 2025

Industrial — The Inflection Point: First Year-Over-Year Growth in Eight Quarters

Industrial grew 4% sequentially to $442M and, more importantly, 6% year-over-year, the first year-over-year growth after eight consecutive quarters of decline. Management believes the segment has "seen the bottom," with global PMI trends pointing to early signs of expansion. Growth was led by traditional industrial and factory automation, with aerospace, defense, and security revenue up 70% year-over-year (a strategic high-barrier win was a solid-state circuit breaker using onsemi's SiC JFET). A meaningful slice of the AI buildout (energy storage, solar inverters, UPS, microgrids) is also reported within Industrial.

"Following eight quarters of year-over-year declines, Q4 marked the first quarter of year-over-year growth in our industrial revenue, increasing 6% over 2024." — Thad Trent, CFO

Assessment: This is the most important single data point in the quarter for the cyclical thesis. Industrial is the broadest, most macro-sensitive of onsemi's markets, and its return to year-over-year growth, corroborated by expanding PMIs, is the clearest evidence yet that the down-cycle has ended. The AI-power overlay within Industrial (ESS leadership at >50% global share, Field Stop 7 IGBT and SiC hybrid modules) gives the segment a secular growth layer on top of the cyclical recovery.

Automotive — Digestion Behind, Content Ahead

Automotive grew 1% sequentially to $798M, with inventory digestion "largely behind us." The forward story is content expansion as the industry moves to zonal architecture for software-defined vehicles: onsemi has built a $400M-plus design funnel for SmartFETs, eFuses, and T1S Ethernet transceivers, on top of its existing silicon and silicon-carbide leadership in electrified powertrains. Industry estimates suggest ~40% of new vehicles will feature zonal architecture within five to eight years. The eight-megapixel image sensor for front-facing vision and Treo-based ultrasonic ADAS sensors are also proliferating.

Assessment: Auto has transitioned from a destocking headwind to a content-growth story. The $400M zonal design funnel is incremental to the existing electrification content, and the SiC franchise continues to broaden into plug-in hybrids and beyond the top-10 OEMs. With digestion complete, auto should track end-demand plus content gains, with restocking as additional upside that has not yet begun.

AI Data Center — Scaling, With a Bigger Prize

onsemi delivered more than $250M of AI data center revenue in 2025 (from almost nothing), and guided Q1 2026 AI revenue up high-teens sequentially. The company raised its 2030 content-per-rack opportunity to roughly $105,000 (from ~$50,000 for 2027) as it introduces more products mapping onto rack content. The Q4 win list spanned the PowerTree: a SiC MOSFET hybrid module at SunGro (430kW, ~99.5% efficiency), a UPS design that cuts system footprint ~50%, rack-level BBU/PSU wins with Delta, Lite-On, and Great Wall, and XPU-board wins for multiphase controllers, smart power stages, and the dual 5x5 Vcore solution.

"We're the only company or one of the very few companies that are able to do the high voltage, think 800 volts with our 1,200 volt devices all the way to the SPS type devices closer to the core… our architecture is moving forward are collapsing the conversion tree." — Hassane El-Khoury, President & CEO

Assessment: AI is now a scaling, quantified growth engine, no longer pure optionality. The doubling of the 2030 content-per-rack opportunity is the key tell: as architectures collapse the power-conversion tree and move to higher-voltage rails, onsemi's full-power-tree breadth (silicon, SiC, SiC JFET, GaN, vGaN, Vcore) maps onto more dollars per rack. At ~4% of sales today and growing high-teens sequentially, AI is on a trajectory toward a double-digit revenue share that would re-rate the company.

Key Topics & Management Commentary

Overall Management Tone: The most confident posture of our coverage, and for the first time grounded in forward growth rather than trough discipline. Management moved from "stabilization" (Q2) to "seasonality is the first step to recovery" (Q3) to, now, a Q1 guide that explicitly contains the first year-over-year revenue growth in over three years. The confidence is calibrated: management still says it has not seen restocking and continues to manage utilization to demand, but it is now describing improving order trends (book-to-bill up, more expedites, fewer turns needed) and a margin path it controls through FabRite. This is a team that believes the cycle has turned and is positioning to scale into it.

1. The First Year-Over-Year Growth Guide in Three Years

"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. This marks the first quarter with expected year-over-year growth since the downturn started over three years ago." — Thad Trent, CFO

The Q1 2026 guide is the headline of the entire report. After more than three years of year-over-year declines, onsemi is guiding to year-over-year growth, even while exiting ~$50M of non-core revenue in the quarter. Excluding the exits, management says revenue would be "above seasonal." The guide is "in line with normal seasonality at the midpoint" on the reported number, which understates the underlying momentum once the deliberate exits are netted out.

Assessment: This is the data point that changes the rating. A return to year-over-year growth, with the underlying (ex-exits) base above seasonal, is the cyclical confirmation we were waiting for at our Q3 recap. It converts "stabilization" into "recovery" with a number attached. The fact that it arrives despite the self-inflicted exit headwind makes it more, not less, convincing.

2. Margins Expand Through Seasonality: FabRite Self-Help Is Working

"The fact that we're actually expanding gross margins 30 basis points shows you that our FabRite initiatives of taking cost out are offsetting the headwind. So it's actually up quarter on quarter." — Thad Trent, CFO

This is the structural change in the thesis. At our initiation, onsemi's gross-margin recovery was entirely demand-gated: nothing moved without utilization, and utilization would not move without a recovery management would not forecast. Q4 breaks that dependency. The company cut fab capacity 12% in 2025, took additional Q4 footprint actions that lower 2026 depreciation by ~$45–50M (gross-margin benefit weighted to H2 2026), reduced under-absorption charges to ~700bps, and is now guiding Q1 gross margin up 30bps in a seasonally down quarter. The utilization-driven upside (each point worth 25–30bps, with ~700bps to recover toward a reset full-utilization level) is now additive on top of the self-help, not a substitute for it.

Assessment: The presence of a working, demand-independent margin lever materially de-risks the margin thesis. Even if the cyclical recovery is gradual, FabRite plus the depreciation step-down plus mix expand margins through 2026, with utilization recovery as upside. The bridge to 40%+ (and eventually the 53% long-term target) now has a structural floor under it that did not exist at our initiation.

3. Record Free Cash Flow and a New $6B Buyback

FY2025 free cash flow was a record $1.4B (24% margin, +17% year-over-year), on cash from operations of $1.8B and capex down to 4.5% of revenue as the large capacity investments concluded. onsemi returned ~100% of free cash flow via buyback in 2025 ($450M in Q4), repurchased $2.6B under the prior program, and authorized a new $6B program in November. The full-year free-cash-flow margin target is 25–30%, with management expecting to expand toward it in 2026.

"Our plan, as you can see, with our announced repurchase authorization of $6 billion is to return 100% of our free cash flow to shareholders." — Thad Trent, CFO

Assessment: Generating a record 24% free-cash-flow margin at a cyclical trough, with capex structurally lower, is the strongest evidence of the model's resilience. As revenue and margins recover, free cash flow has substantial operating leverage. The $6B authorization (roughly a fifth of the market capitalization) and the 100%-return policy provide a powerful per-share tailwind and a firm valuation floor.

4. AI: A Bigger Prize and the "Wall to Core" Moat

Management raised the 2030 content-per-rack opportunity to ~$105,000 (from ~$50,000 for 2027) as more products map onto rack content. The differentiation argument is that onsemi can convert power at every stage from the 800-volt input to the processor core, and that next-generation architectures are "collapsing the conversion tree," requiring suppliers that span both high and low voltage. vGaN (the only vertical GaN being sampled, built on proprietary GaN-on-GaN in a US fab) is the highest-power-density, highest-voltage piece of that toolkit.

Assessment: The doubling of the rack-content opportunity is the most bullish AI disclosure to date and reflects real product proliferation (SiC JFET, vGaN, 5x5 SPS, Vcore) rather than TAM hand-waving. The "wall to core" breadth is a genuine, durable moat in a field where most competitors address a single conversion stage. The constraint remains scale and the XPU vendors' deployment cadence, but the trajectory is unambiguously up and to the right.

5. Wide-Bandgap Expansion: 30+ GaN Devices and the vGaN/GM Program

onsemi is preparing to sample more than 30 new GaN devices in 2026 spanning 40 to 1,200 volts (both discretes and integrated driver-plus-GaN solutions), with lateral-GaN revenue beginning in 2026 via new foundry partnerships for regional supply. For vertical GaN, the company is collaborating with GM on electric-drive systems, with first vGaN revenue expected in 2027. Management explicitly likened the wide-bandgap commercialization muscle to the GTAT acquisition that built the SiC franchise.

Assessment: The GaN ramp broadens onsemi's wide-bandgap leadership beyond silicon carbide and adds a 2026–2028 growth layer. The GM electric-drive collaboration is a credibility marker for vGaN in high-power automotive. These are durable competitive extensions; with vGaN revenue in 2027, they support the long-term mix-and-margin thesis without affecting the near-term model.

6. Non-Core Exits: A Smaller Base That Is Still Growing

onsemi will exit ~$50M of non-core revenue in Q1 and ~$300M across 2026 (weighted toward Q2–Q3 plus seasonality). Management was emphatic that the core business has grown above market even through the downturn, that Q1 is "above seasonal" excluding the exits, and that 2027 will resume above-market growth net of the exits. The exits are margin-neutral today (near corporate average) but are being shed for pricing pressure and volatility.

"If you account for the $50 million exit in the first quarter, we're actually above seasonality as a baseline… we expect '27 to resume that over market growth… now that we net out the exits." — Hassane El-Khoury, President & CEO

Assessment: The exit narrative has flipped from a drag to a quality signal. At our initiation the ~5% revenue exit looked like a base-shrinking headwind that capped the recovery; with the core now demonstrably above seasonal and the cyclical turn confirmed, the exits read as deliberate quality improvement rather than weakness. The 2027 "resume above-market growth" framing sets a clear bar.

7. Visibility Improving: Book-to-Bill, Expedites, and the Restocking Still to Come

Management cited concrete leading indicators of improving demand: book-to-bill trending up, fewer turns needed to fill the quarter than a quarter ago, and more expedites than 90 days ago. Crucially, this is happening without restocking, which management says has still not begun. OEMs and Tier-1s have drained inventory below critical levels and will eventually need to replenish, which would be incremental demand on top of the current end-demand-driven recovery.

"Book to bill is trending up. We are walking into the quarter with less turns needed than the prior quarter. Expedites, we're seeing more expedites than we did ninety days ago… We're not seeing the replenishment yet." — Hassane El-Khoury, President & CEO

Assessment: The improving order KPIs corroborate the recovery without relying on it. And the still-absent restocking is the most attractive part of the setup: onsemi is guiding to year-over-year growth on end-demand alone, with a coiled restocking spring that has not yet contributed. When replenishment begins, it lands on lean inventory and a smaller footprint with ~700bps of utilization-driven margin upside, a powerful operating-leverage scenario.

8. Capital Position and Inventory

onsemi ended 2025 with ~$2.5B of cash and short-term investments and $4B of total liquidity ($1.5B undrawn revolver). Inventory fell $58M to 192 days, including 76 days of strategic bridge inventory (down from 82) that the company will continue to deplete over two years; base inventory is healthy at 117 days. Distribution inventory was 10.8 weeks, within the 9–11-week target. Q4 capex was $69M (4.5% of revenue).

Assessment: A clean, conservatively managed balance sheet with lean inventory positions onsemi to convert the recovery efficiently. The continued depletion of strategic SiC bridge inventory is an additional multi-quarter cash tailwind. Nothing here constrains the recovery; the setup is for operating leverage to flow through as demand builds.

Guidance & Outlook

The Q1 2026 guide of $1.44–1.54B (midpoint $1.49B) is the first year-over-year-growth quarter in over three years. By end-market, management expects automotive roughly flat (a Chinese New Year headwind), industrial down low-teens (seasonality, energy infrastructure, and factory-automation lumpiness), and "Other" up low-single-digits as AI data center grows high-teens against seasonal and exit offsets. Excluding the ~$50M of Q1 non-core exits, revenue would be above seasonal, so the underlying momentum is stronger than the headline.

The margin signal: The 30bp sequential gross-margin expansion (to a 38.5% midpoint) in a seasonally down quarter is the most important element of the guide. It is direct evidence that FabRite cost actions and the lower depreciation base are expanding margins independent of utilization, which steps up only to the low-70s in Q1. Management expects further margin expansion through 2026 as FabRite benefits compound, the ~$45–50M depreciation reduction lands (weighted to H2), and utilization recovers (with a ~two-quarter lag to the P&L).

2026–2027 framing: Management declined to guide the full year but pointed to: the ~$300M non-core exit (a deliberate, margin-neutral, quality-improving headwind), AI data center growing high-teens-plus, the lower ~15% tax rate, and 2027 as the year above-market growth resumes net of exits. The composite picture is a smaller-but-higher-quality revenue base inflecting to year-over-year growth, with margins expanding on self-help and a coiled restocking spring as upside.

Analyst Q&A Highlights

When the Business Moves Above Seasonal

The opening exchange probed the weak "Other" category and, more importantly, when onsemi can grow above seasonal given the point in the cycle. Management's answer reframed the reported in-line guide as understating underlying momentum once the deliberate exits are netted out.

Q: "You sound better on your tone about everything to do with the cycle and secular… But you're kind of still at seasonal. So when do you think you could be above seasonal given the point of the cycle that we're at right now?"
— Ross Seymore, Deutsche Bank

A: "If you exclude the exits, we are above seasonal… if you look at the reported numbers, they're in line with seasonality… But if you exclude the exits, we are above seasonal… there are also about $40 million of exits that are in there. So that's why that bucket was down sequentially."
— Thad Trent, CFO

Assessment: The "above seasonal ex-exits" framing is the analytically important point. onsemi is deliberately suppressing its own reported growth by shedding ~$300M of non-core revenue in 2026, so the underlying core business is already running above seasonal. That makes the first-YoY-growth guide more impressive, not less, and supports the view that the recovery is real beneath the optics.

AI TAM, the Path to 10% of Sales, and Differentiation

A request to size the AI total addressable market, identify when it could reach 10% of sales, and articulate the differentiation drew the clearest statement of the "wall to core" and "collapsing the conversion tree" moat.

Q: "The $250 million or slightly more than that… can you just talk a little bit about the TAM you think you can address in that? When do you think it could be 10% of sales or something larger like that? And how ON differentiates[?]"
— Ross Seymore, Deutsche Bank

A: "We're the only company or one of the very few companies that are able to do the high voltage… all the way to the SPS type devices closer to the core… our architecture is moving forward are collapsing the conversion tree. Which means you have to be able to do high voltage and low voltage. We're the only ones company that has vertical GaN."
— Hassane El-Khoury, President & CEO

Assessment: Management would not put a date on 10% of sales, but the differentiation answer is substantive: full-power-tree coverage plus a unique vGaN capability is a real moat as architectures collapse the conversion stages and move to higher-voltage rails. The reluctance to guide AI specifically is conservative; the trajectory ($250M from "almost nothing," high-teens sequential growth) implies a double-digit revenue share is a question of when, not if.

The Q1 Gross-Margin Puzzle: Up 30bps Through a Seasonal Decline

A modeling question flagged the apparent contradiction that utilization is rising while Q1 margins are typically down, and asked for the puts and takes. The answer surfaced the central thesis change: FabRite self-help is now expanding margins through seasonality.

Q: "Maybe can you quickly walk through the puts and takes on gross margins in Q1? I mean utilization rates are coming up, but margins are down sequentially… any other puts and takes… we should be considering[?]"
— Joshua Buchalter, TD Cowen

A: "The key element here is that the Q1 midpoint of our guidance is up 30 basis points on gross margin… Typically, Q1 is seasonally down… The fact that we're actually expanding gross margins 30 basis points shows you that our FabRite initiatives of taking cost out are offsetting the headwind."
— Thad Trent, CFO

Assessment: This exchange is the crux of the upgrade. A company whose margins expand through a seasonally down quarter has a demand-independent margin lever, which is exactly what onsemi lacked at our initiation. It de-risks the entire margin-recovery thesis: even gradual demand recovery now compounds with structural cost-out, and the utilization recovery becomes upside rather than the sole path.

AI Data Center Growth in 2026

A question on 2026 AI growth drew a qualitative bullish answer from the CEO and an unusual quantitative data point from the CFO, who put a number on Q1.

Q: "You did throw out a kind of target for the year that you exceeded. I was just curious. Thoughts about growth in that segment for '26[?]"
— Blayne Curtis, Jefferies

A: "We're starting off the year with a better growth than we did starting off last year. So I'm very bullish about that segment… [CFO Trent:] We do expect that our AI data center revenue in Q1 will grow high teens percentage-wise. To give you an indication of our trajectory here."
— Hassane El-Khoury, CEO / Thad Trent, CFO

Assessment: The CFO's high-teens Q1 sequential growth figure is a useful, rare quantification of AI momentum and confirms the segment is accelerating into 2026 from the $250M-plus 2025 base. Starting 2026 with better growth than 2025 began implies the AI revenue line compounds materially this year, which is the single most important driver of the long-term re-rating case.

The 2030 Content-Per-Rack Opportunity Doubling to $105,000

A question on the updated slide showing AI rack content rising to ~$105,000 by 2030 (versus ~$50,000 for 2027) asked what is driving the increase. The answer tied it to product proliferation mapping onto more rack content.

Q: "Now your opportunity… what that looks like per rack in 2030… you're at, like, $105,000 per rack versus, like, $50,000 for 2027. Just curious, like, what's driving that?"
— Joe Quatrochi, Wells Fargo

A: "As new generation architectures start to firm up… our overlap of products availability that we're making and investing in… versus the opportunity of content in the rack has increased… more products mapping onto more content that we can address with our portfolio."
— Hassane El-Khoury, President & CEO

Assessment: A doubling of the 2030 rack-content opportunity in a single quarter is a striking upward revision, and management grounds it in tangible product introductions (SiC JFET, vGaN, 5x5 SPS, Vcore) rather than TAM optimism. It reframes onsemi's AI opportunity from a power-discrete supplier to a broad content gainer as racks move to higher voltage. This is the disclosure that most expands the long-term bull case.

Forward Revenue Visibility: Book-to-Bill, Turns, and Expedites

A question on forward revenue KPIs (starting backlog, expedites, visibility change over 90 days) drew a clean confirmation that the leading indicators are improving, even without restocking.

Q: "Maybe if you can share with us a few more specifics on the forward-looking revenue KPIs like… starting backlog or… customer expedites or just any sort of revenue visibility change[?]"
— Gary Mobley, Loop Capital

A: "Book to bill is trending up. We are walking into the quarter with less turns needed than the prior quarter. Expedites, we're seeing more expedites than we did ninety days ago… We're not seeing the replenishment yet. But this is strong signals for stabilization in the market."
— Hassane El-Khoury, President & CEO

Assessment: Improving book-to-bill, fewer required turns, and rising expedites are exactly the leading indicators that precede a recovery, and they corroborate the first-YoY-growth guide. The still-absent restocking is the upside kicker: onsemi is growing on end-demand alone, with replenishment as a future tailwind on lean inventory. This is the strongest visibility commentary of our coverage.

Do the Exits Hurt Utilization or Margin?

A question on whether walking away from ~$90M/quarter of non-core revenue over two quarters helps or hurts gross margin and utilization drew a clarifying answer: the exits are margin-neutral and not a utilization headwind.

Q: "These are lower margin but… I was just curious the impact to utilization… you're walking away from $90 million a quarter over two quarters. Is that a positive or a negative for gross margin?"
— Blayne Curtis, Jefferies

A: "It's neutral today. So the margin on that business today is near the corporate average. The reason we're exiting is because we're seeing margin pressure and pricing pressure… It's not going to be a headwind to utilization… net of these exits… our utilization, we expect to go up throughout the year."
— Thad Trent, CFO

Assessment: The clarification removes a modeling concern: the exits do not dilute gross margin (they are near the corporate average) and do not pressure utilization (the freed capacity absorbs growth). They are a forward-looking quality decision, shedding volatile, price-pressured revenue before it becomes structurally dilutive. Utilization is still expected to rise through 2026 net of the exits, which underpins the margin-expansion guide.

What They're NOT Saying

1. No full-year 2026 revenue guide: Management guided only Q1 and declined to frame full-year 2026 revenue, leaving the magnitude of the recovery (beyond "year-over-year growth") undefined. The shape of H2 2026 remains an estimate.
2. AI still not formally guided: Despite a high-teens Q1 growth data point and a doubled 2030 rack opportunity, management would not give an annual AI growth target or a date for AI reaching 10% of sales, keeping the most important secular driver loosely framed.
3. Restocking timing undated: The coiled restocking spring is acknowledged but not dated; management cannot say when OEMs and Tier-1s will replenish, so the biggest potential upside catalyst remains unscheduled.
4. SiC revenue still unsized: For another quarter, management declined to size the silicon-carbide business in absolute dollars, describing it only as "coming in as expected" with continued share gains.
5. Limited detail on pricing for 2026: Management expects "normal pricing behavior" and framed any industry price increases as potential upside, but offered little specificity on blended ASP assumptions embedded in the recovery.

Market Reaction

• Pre-print setup: ON closed at $65.10 on Feb 9 (the print was released after the close, with an afternoon call), up 20.2% year-to-date and 27.0% over the trailing twelve months, at the top of its 52-week closing range ($31.95–$65.20). The stock had re-rated substantially from the ~$50 level at our Q3 recap as the recovery narrative gained traction.
• Reaction session (Feb 10): ON closed at $67.38, up 3.5% (+$2.28), after trading as high as $68.80 (+5.7%) intraday, on volume of 19.3M shares versus an 8.6M-share 30-day average (2.2x normal).
• Market context: The S&P 500 was down 0.3% on the session, so onsemi's gain was idiosyncratic and against a soft tape.

A 3.5% gain on a quarter that merely met the revenue midpoint tells us the market is paying for the forward signal, not the trailing print. After an initial after-hours wobble on the optically in-line Q1 revenue midpoint, the stock recovered and rallied as the market digested three forward positives:

The first year-over-year growth guide: The Q1 2026 guide containing the first year-over-year revenue growth in over three years is the cleanest cyclical-confirmation signal onsemi has provided, and it arrived with an "above seasonal ex-exits" qualifier that made it more credible.

Margins expanding through seasonality: A 30bp sequential gross-margin expansion in a seasonally down quarter, driven by FabRite self-help, told the market the margin recovery is no longer hostage to utilization alone, which is precisely the structural concern that had capped the multiple.

Capital return and AI: The record $1.4B free cash flow, the new $6B buyback, and the doubled 2030 AI rack-content opportunity reinforced both the valuation floor and the growth ceiling. The +3.5% reaction at the top of the 52-week range, against a down market, is a meaningful endorsement of the inflection.

Street Perspective

Debate: Is the Cyclical Recovery Confirmed?

Bull view: Industrial grew year-over-year for the first time in eight quarters, auto digestion is done, Q1 is guided to year-over-year growth (the first in three-plus years, above seasonal ex-exits), book-to-bill is rising, expedites are up, and PMIs are expanding. With restocking not yet begun, there is additional upside on top of an end-demand recovery that is already underway. This is the start of a multi-year up-cycle.

Bear view: The Q1 reported guide is still in line with normal seasonality, full-year 2026 is unguided, and management again would not call restocking. Year-over-year growth off a deeply depressed base is a low bar, and the ~$300M of 2026 exits caps reported growth. A gradual recovery against a stock already up 27% over twelve months leaves limited room for error.

Our take: Bull. The combination of the first year-over-year growth guide, the first year-over-year industrial growth, expanding PMIs, and improving order KPIs is sufficient confirmation that the down-cycle has ended. The exits-adjusted "above seasonal" base and the undated restocking spring are upside. The recovery is confirmed; the open question is only its slope.

Debate: Has the Stock Already Priced the Recovery?

Bull view: The stock is up off its lows, but earnings estimates are about to climb materially as the recovery, margin self-help, AI scaling, and buyback compound. At ~23x our FY26 EPS and ~18x FY27, a recovering cyclical with a structural margin lever and a doubling AI content opportunity is not expensive. The +3.5% reaction at the highs, against a down tape, shows the market is still adding.

Bear view: ON has re-rated ~30% since the autumn and trades near a multi-year high. The easy money (buying the trough) has been made. At a low-20s forward multiple on a still-recovering earnings stream, the stock needs the recovery to play out flawlessly. Any macro setback or margin disappointment leaves meaningful downside from these levels.

Our take: Lean bull, with eyes open on valuation. The trough-to-here move is largely behind us, but the next leg is driven by rising estimates rather than multiple expansion: FY26 EPS toward ~$2.90 and FY27 toward ~$3.50–4.00 on margin recovery and AI scaling. That earnings trajectory, plus the buyback, supports Outperform even after the re-rate. The risk is a stalled recovery, which the bear case correctly identifies.

Debate: Is the Margin Self-Help Durable or a One-Quarter Optic?

Bull view: FabRite cut fab capacity 12% in 2025, the actions lower 2026 depreciation by ~$45–50M, under-absorption fell to ~700bps, and Q1 margins expand 30bps through seasonality. As utilization recovers (each point worth 25–30bps) on top of the structural cost-out, gross margin marches toward 40% in 2026 and the 53% target over time. The self-help is structural and compounding.

Bear view: A 30bp expansion is small, SiC remains below the corporate average, and the bulk of the path to 53% still depends on a demand recovery that lifts utilization from the high-60s/low-70s. Mix and one-time cost actions can stall, and the 53% target has been a multi-year "someday" number. Mid-40s is the realistic medium-term ceiling.

Our take: Bull on the mechanism, realistic on the destination. The key is that margins now expand without utilization, which they could not at our initiation. That structural lever, plus the depreciation step-down and eventual utilization recovery, makes 40%+ in 2026 plausible and 45%+ a credible 2027 target, with 53% the long-term aspiration. The self-help is durable; the timeline to the full target remains demand-influenced.

Model Implications & Valuation Framework

Valuation framework: At the post-reaction price of ~$67.38 and our FY26 non-GAAP EPS estimate of ~$2.75–3.05, ON trades at roughly 22–24x forward earnings, and ~17–18x our preliminary FY27 estimate of ~$3.50–4.10. For a semiconductor at the start of an up-cycle, with a now-structural margin lever, a doubling AI content opportunity, record free cash flow, and a $6B buyback (roughly a fifth of the market cap), a low-20s forward multiple on rising estimates is reasonable. The thesis has re-rated the stock, but the next leg is driven by estimate revisions higher rather than multiple expansion.

Price-target framework (12-month):

• Base case ~$78–82: ~20–21x our preliminary ~$3.80 FY27 EPS (or ~27x ~$2.90 FY26). Recovery proceeds, margins reach ~41–43%, AI scales toward high-single-digit revenue share. Implies ~16–22% upside.
• Bull case ~$92–100: ~24–26x a ~$3.90 FY27 EPS if restocking begins in H1 2026, utilization recovers sharply (margins toward mid-40s), and AI plus GaN accelerate. Implies ~37–48% upside.
• Bear case ~$54–58: ~21–22x a ~$2.60 FY26 EPS if the recovery stalls, restocking is deferred, and margins plateau in the high-30s. Implies ~14–20% downside.

The base case implies ~16–22% upside from $67.38, with a favorable up/down ratio (roughly 1.5:1 base-to-bear, wider in the bull case). The swing factors, restocking timing and the slope of margin expansion, are now skewed to the upside given the structural self-help lever. That asymmetry, combined with the confirmed cyclical turn, is the quantitative basis for the upgrade to Outperform.

Thesis Scorecard Post-Earnings

Grading the standing thesis (established at initiation, carried through Q3) against this quarter's print, guide, and call.

Overall: Thesis materially strengthened. The two bear points that justified Hold (no confirmed recovery; margin pinned on demand) both moved decisively in our favor this quarter: the recovery is confirmed (first YoY growth guide in three years) and margins now expand on self-help. Bull #1 (AI/wide-bandgap) strengthened with the doubled rack-content opportunity, and Bull #3 (capital return) strengthened with the record FCF and $6B buyback.

Action: Upgrade to Outperform. The conditions we set at Q3 for an upgrade have been met, with restocking as additional, undelivered upside.

Bottom Line: The Turn We Were Waiting For — Upgrading to Outperform

Rating decision: We are upgrading onsemi to Outperform from Hold. At our Q3 recap we maintained Hold but said the path to an upgrade was "clearly visible (restocking + utilization inflection + AI scaling + margin breaking out)" and that we were "one or two quarters of evidence away." Q4 delivered the evidence, and in the most convincing form: the first year-over-year revenue-growth guide in over three years, the first year-over-year industrial growth in eight quarters, and gross margin expanding 30bps through a seasonally down quarter on FabRite self-help. The two bear points that anchored our Hold, an unconfirmed recovery and a demand-gated margin, both broke bullish.

The margin development is the structural game-changer. At our initiation, onsemi's margin recovery was entirely hostage to a demand cycle management would not forecast. It now has a working, demand-independent self-help lever (FabRite, a 12% capacity reduction, a $45–50M depreciation step-down) that expands margins even as utilization sits in the high-60s/low-70s. Add the utilization recovery as upside, the record $1.4B free cash flow, the $6B buyback, and an AI content opportunity that just doubled to $105,000 per rack by 2030, and the risk/reward has shifted decisively. The stock has re-rated with the thesis, but the next leg is driven by rising earnings estimates through 2026–2027, with a coiled restocking spring that has not yet contributed.

What would make us more constructive (toward the bull case): The restocking cycle beginning in H1 2026; utilization recovering toward the mid-to-high-70s with gross margin breaking decisively above 40%; AI data center scaling toward a high-single-digit revenue share with an annual growth target disclosed; or a full-year 2026 guide that frames double-digit core growth.

What would move us back to Hold: The recovery stalling (Q2/Q3 2026 failing to grow year-over-year ex-exits); gross margin plateauing in the high-30s despite FabRite; or the valuation running well ahead of the earnings trajectory (a mid-20s forward multiple without estimate support).

Signposts for Q1 2026 earnings (early May 2026):