ON Semiconductor is emerging from a cyclical reset with improving demand, accelerating AI data center exposure, and early traction in Treo. The stock looks constructive for investors willing to underwrite execution risk and valuation.
ON Semiconductor(ON) looks like a medium-term recovery story with real secular legs, not just a bounce off a cyclical low. The core case rests on three named facts. First, Q1 2026 revenue reached $1.513B and non-GAAP EPS came in at $0.64, with management guiding Q2 revenue to $1.535B to $1.635B and non-GAAP EPS to $0.65 to $0.77. Second, the company’s highest-value growth pockets are gaining traction: AI data center revenue grew more than 30% sequentially in Q1 and management said it expects AI data center revenue to double in 2026, while Treo revenue increased more than 2.5x sequentially. Third, ON still generates meaningful cash, with trailing free cash flow of $2.10B and a 5.83% FCF yield, even after a difficult 2025 margin reset.
The pushback is obvious. Trailing earnings are weak enough to make the stock look expensive on a trailing basis, with a 68.1x trailing P/E, and 2025 net income fell to $121.0M from $1.57B in 2024. Gross margin also compressed to 32.3% for full-year 2025. That is the danger in ON right now: investors are paying for recovery before the income statement fully shows it. In plain English, the market is buying the next lap, not the last one.
For a balanced, moderate-risk investor, the stock is attractive if the thesis is framed correctly. ON is not a low-drama compounder like a broad analog giant, and it is not a pure AI rocket either. It sits in the middle: a cyclical semiconductor name with improving execution, strong automotive and industrial positioning, and new upside from AI power, silicon carbide, GaN, and software-defined vehicle content. That mix supports a constructive stance, but only with respect for valuation and execution risk.
ON Semiconductor(ON), headquartered in Scottsdale, Arizona, is a power and sensing semiconductor company serving automotive, industrial, AI data center, computing, mobile, and automation markets. The company operates globally and employs 22,600 people. Its business is built around power conversion, power management, connectivity, image sensing, and intelligent sensing products that help customers improve efficiency, reliability, and system performance.
The company reports three operating segments: Power Solutions Group, Analog and Mixed-Signal Group, and Intelligent Sensing Group. According to business context tied to the 2025 year-end filing, Power Solutions Group represented 47% of 2025 revenue, Analog and Mixed-Signal Group 38%, and Intelligent Sensing Group 15%. That mix matters because it shows ON is still primarily a power semiconductor company, with sensing as a meaningful but smaller differentiator.
ON’s strategic identity has shifted over the last several years. Management is steering the portfolio toward electrification, energy efficiency, safety, automation, and AI power infrastructure. In Q1 2026, automotive revenue was $797M, industrial revenue was $417M, and the company said automotive and industrial together account for almost 80% of revenue in FY2025. That concentration creates both strength and vulnerability. It gives ON direct exposure to EVs, ADAS, factory automation, storage, and grid infrastructure, but it also ties results to cyclical demand in auto and industrial channels.
The current market value is about $36.0B. With trailing revenue of $6.06B and EBITDA of $2.05B, ON is no longer priced like a distressed cyclical. The market is assigning value to the company’s position in silicon carbide, GaN, AI data center power, and next-generation vehicle architectures. That premium only works if execution keeps improving.
Power Solutions Group is the workhorse. In Q1 2026, PSG revenue was $736.6M, up 2% sequentially and 14% YoY, making it the strongest segment in the quarter. This unit includes SiC, SiC JFETs, MOSFETs, power modules, vertical GaN, gate drivers, and other power devices. It is the segment most directly exposed to EV traction inverters, charging, energy storage, renewable power, and AI data center power conversion.
Analog and Mixed-Signal Group is more mixed right now. Q1 2026 AMG revenue was $540.4M, down 3% sequentially and 5% YoY. This business includes analog, ASICs, logic, isolation, power management, connectivity, and sensor interfaces. It is strategically important because products in this segment often sit closer to the system architecture and can carry sticky design-in value, but the recent revenue trend shows it has not yet become the growth engine management wants.
Intelligent Sensing Group remains smaller but strategically useful. Q1 2026 ISG revenue was $236.3M, down 5% sequentially and up 1% YoY. The product set includes image sensors, image signal processors, SWIR sensors, and photon-counting technologies. This segment gives ON a differentiated angle in automotive sensing, industrial autonomy, robotics, and safety systems. It also helps the company sell more complete solutions rather than just discrete power devices.
The annual revenue history shows why ON’s segment mix matters. Total revenue peaked at $8.33B in 2022, stayed high at $8.25B in 2023, fell to $7.08B in 2024, and dropped again to $6.00B in 2025. The company is trying to replace lower-quality or non-core revenue with higher-value content in EV power, AI power, and intelligent systems. Management said Q1 included about $50M of planned non-core exits and expects another $30M to $40M of non-core exits in Q2. That is painful in the short run, but it is usually the right move if the replacement revenue carries better margins and stronger design-win durability.
Get AI research on any stock
Instant reports, daily intelligence, and an AI analyst in your pocket.
The flagship product family to watch is Treo. Management highlighted Treo repeatedly in the Q1 2026 earnings call, and for good reason. Revenue from Treo increased more than 2.5x sequentially in Q1, and the platform is moving from product proliferation into revenue ramp and design wins. That is the kind of sentence investors should circle. It means the product is leaving the lab and entering customer programs.
Treo is showing up in several applications: 10BASE-T1S zonal automotive architectures, ADAS park-assist systems using ultrasonic sensing, power management for AI client platforms, and inductive position sensing for humanoids and advanced automation. ON also said it began production shipments of Treo-based T1S Ethernet solutions for a leading North American customer’s next-generation zonal architecture, with more than 30 Treo devices integrated into the platform. That is not a science project. That is production content.
That quote is the real hook. A product line with 60% to 70% gross margin can materially change the earnings profile if it scales. ON’s company-wide gross margin was 38.5% in Q1 2026 and 32.3% for full-year 2025. So Treo is not just a growth product. It is a mix-upgrade product. In semiconductors, that is how a company stops being treated like a commodity supplier and starts earning a better multiple.
EliteSiC is the other flagship family that matters. Management said ON is already in production with customers on next-generation 900-volt EV platforms and that its silicon carbide share of new EV models deployed at the 2026 Beijing Auto Show was about 55%. Expanded collaborations with Geely and NIO reinforce that ON is not merely participating in EV power. It is embedded in platform transitions where semiconductor content per vehicle rises.
GaN is earlier in the monetization curve but still important. ON said its GaN solutions design funnel exceeds $1.5B, with products spanning 40 to 1,200 volts. Ten products are already sampling and another 20 are sampling in 2026. Revenue contribution is expected to begin ramping in 2027. That makes GaN more of a forward option than a current earnings driver, but it is a credible option, not a slide-deck fantasy.
ON’s competitive advantage is strongest where efficiency, power density, reliability, and system-level integration matter more than the lowest unit price. That is why the company is leaning into silicon carbide, GaN, AI power conversion, zonal automotive networking, and multimodal sensing. These are not the easiest semiconductor categories, but they are the ones where customers care about performance and qualification, not just price per chip.
The company’s moat looks moderate but real. Business context points to four pillars: vertically integrated SiC manufacturing, design-in switching costs, customer relationships in EV platforms, and a global manufacturing footprint. Recent product tools such as Elite Pairing Studio are designed to help customers choose matched SiC MOSFET and gate-driver combinations early in the design cycle. That kind of workflow creates stickier customer relationships because once a power architecture is qualified, customers do not swap suppliers casually.
Management also framed ON as the only broad-based U.S. power semiconductor supplier serving AI data center needs across the full power tree, from high-voltage conversion to intelligent power stages, protection, control, and system-level integration from the grid to the processor. That does not mean ON owns the market. It does mean the company has a differentiated pitch at a time when hyperscalers care about efficiency losses from the wall to the rack and from the rack to the processor.
That statement matters because ON does not need to become the face of AI to profit from AI. It needs to become a trusted power enabler behind the scenes. In semiconductors, the quiet supplier that solves a painful bottleneck often does very well.
The acquisition angle also supports the innovation case. Management said 2025 acquisitions such as Aurasemi and a JFET silicon carbide acquisition strengthened the company’s power technology platform. Those deals fit the strategy rather than distracting from it. ON is not trying to become everything. It is trying to become more valuable in the places where power complexity is rising.
Operations are improving, but this remains a key execution area. In Q1 2026, manufacturing utilization increased sequentially to 77% as ON ramped production to meet stronger demand signals. The CFO said each point of utilization improvement adds 25 to 30 basis points of gross margin. That is a very useful operating lever because it turns a demand recovery into visible margin expansion.
Management’s FabRight actions are another major lever. The company said it expects sequential gross margin expansion throughout 2026 and sees another 200 basis points of longer-term improvement from FabRight initiatives, more than 200 basis points from favorable mix, and another 200 basis points longer term from prior fab divestiture actions that should start showing up more clearly in 2027. When management starts stacking basis points like that, the message is simple: the factory is becoming less of a drag and more of a weapon.
Inventory still deserves scrutiny. Inventory increased by $60M in Q1 to 201 days from 192 days in Q4. Management said that reflected higher internal loadings and customer commitments, while strategic inventory declined to 75 days from 76 days. Excluding strategic builds, base inventory stood at 126 days, and distribution inventory was flat at 10.8 weeks. This is not a clean inventory picture yet, but it is more controlled than the headline 201 days first suggests.
The 2025 10-K and business context also highlight inventory risk directly. ON recorded $268.2M of excess and obsolete inventory charges in 2025, including $230.3M tied to ISG inventory and strategy changes under the 2025 Manufacturing Realignment Program. That is a serious reminder that portfolio shifts and demand resets can get expensive fast in semiconductors. The good news is that management is confronting the issue rather than pretending it will solve itself.
Capital spending is currently restrained. Q1 2026 capex was just $22M, or 1.4% of revenue, and the company guided Q2 capex to $25M to $35M. That supports free cash flow in the near term and reflects a more disciplined posture after heavier investment periods. It also means ON is trying to squeeze more from its existing footprint before reaching for the checkbook.
ON operates in several semiconductor markets that are moving at different speeds. Broad industry data from Gartner shows global semiconductor revenue is expected to reach $733.0B in 2025 and $814.8B in 2026, with AI as the main growth engine. That backdrop helps ON, but not evenly. The company is more exposed to power, analog, and sensing than to leading-edge compute silicon, so it benefits from AI indirectly through power infrastructure rather than directly through accelerators.
That indirect AI exposure is still meaningful. MarketsandMarkets estimates the data center semiconductor market will grow from $86.8B in 2024 to $265.8B by 2029, a 25.1% CAGR. ON’s role in the power tree, battery backup, supercapacitors, high-voltage conversion, and rack-level power means it can capture value as AI racks become more power-hungry and more complex. Management said a 120-kilowatt rack carries roughly $9,500 of content today, while an 800-volt or high-voltage rack could carry about $115,000 of content. That is a large content step-up.
Automotive remains the company’s largest and most strategic market. TrendForce projects the global automotive semiconductor market to rise from about $67.7B in 2024 to $96.9B by 2029, driven by electrification and vehicle intelligence. ON is positioned for both trends. In Q1, automotive revenue was $797M, roughly flat sequentially and up nearly 5% YoY, marking the first YoY growth after seven quarters of decline. Management also said China automotive revenue grew YoY despite a 6% decline in the China passenger vehicle market. That points to content gains and share gains, not just a market rebound.
Industrial is recovering more slowly but still matters. SEMI’s 2025 outlook says industrial and automotive sectors are expected to improve from Q2 2025 onward after inventory normalization. ON’s industrial revenue was $417M in Q1, down 6% sequentially but ahead of management’s expectations. The company also said it expects more than 40% YoY revenue growth in its energy storage and microgrid business in 2026, with market share approaching 60%. That is a niche, but a valuable one.
Like what you're reading?
Get full access to AI-powered research reports, market analysis, and portfolio tools.
ON’s customer base is concentrated in demanding, qualification-heavy markets. Automotive OEMs and Tier 1 suppliers, industrial automation customers, renewable and storage system makers, AI power supply vendors, and hyperscaler-linked infrastructure partners all sit inside its target set. The company’s product design and manufacturing model is built around long qualification cycles, application engineering, and supply continuity rather than quick-turn consumer gadget demand.
Named customer and partner references from Q1 2026 help make this concrete. ON cited expanded collaborations with Geely and NIO in 900-volt EV platforms, a production shipment for a leading North American OEM’s zonal architecture, more than 30 active programs with FlexPower, and a design win with a leading global robotics platform for high-resolution image sensors and indirect time-of-flight technology. It also highlighted a Sungrow Electric announcement tied to hybrid power modules for utility-scale solar inverters and liquid-cooled energy storage platforms.
This customer profile has two implications. First, design wins can be sticky and valuable once qualified. Second, revenue can be lumpy because customers in automotive and industrial do not move in a straight line. That is why ON’s management keeps emphasizing content growth over SAAR and system value over unit volume. In Q1, the company said it is more leveraged to content than to global auto production and that some of its gains also reflect market share gains.
Ownership data also shows ON is institutionally owned and closely watched. Institutional ownership stands at 109.47%, insider ownership at 0.34%, short interest is 11.02% of float, and the short ratio is 2.67. That combination usually means the stock can move sharply when the cycle turns because both institutions and shorts are active in the name.
ON competes in a crowded field, but not all competition is equal. In power semiconductors, the company specifically names Infineon, STMicroelectronics, Wolfspeed, ROHM, and Nexperia as primary competitors. In adjacent analog, automotive, and industrial categories, relevant rivals include Texas Instruments, NXP, Renesas, Analog Devices, Power Integrations, and Microchip. This is not a sleepy neighborhood.
Against those peers, ON’s strengths are focus and positioning. The company is unusually concentrated in automotive and industrial, where power efficiency, high-voltage capability, and sensing matter. It also has a stronger identity in SiC and power modules than many broad analog peers, and a broader product set than some narrower power specialists. That middle ground can be attractive if the company executes well.
The weakness is diversification. Compared with Texas Instruments or Analog Devices, ON has less insulation from auto and industrial swings. Compared with Infineon or STMicroelectronics, it has fewer scale advantages across a broad global portfolio. Compared with a pure specialist, it can look less singular. In other words, ON’s competitive position is good, but it does not come with much room for sloppy execution.
The most important competitive battlegrounds over the next few years are silicon carbide, GaN, AI power, and software-defined vehicle architectures. ON has credible assets in all four. The company said its GaN funnel exceeds $1.5B, its SiC share of new EV models at the 2026 Beijing Auto Show was about 55%, and its AI data center revenue more than doubled YoY in Q1. Those are the right scoreboards.
Macro matters a lot for ON because semiconductors remain cyclical, and ON is tied to auto and industrial demand. Gartner noted that 2025 semiconductor growth is being led by AI while automotive and industrial demand remains weaker near term. That tension explains ON’s current setup: secular growth opportunities are real, but the company still has to work through the usual inventory, utilization, and customer digestion issues in its core markets.
Geopolitics also matters because ON has global manufacturing and assembly operations across the U.S., Czech Republic, Japan, South Korea, Malaysia, and other regions, while also selling into China and working with Chinese EV players such as Geely and NIO. That footprint supports supply continuity, but it also exposes the company to trade friction, export controls, and regional demand shifts. In semiconductors, a global footprint is both moat and headache. The passport helps, but customs still exists.
Energy infrastructure and EV policy trends are more favorable. Management tied higher energy costs to accelerating EV demand and highlighted durable demand for high-efficiency EV platforms and silicon carbide content. It also linked AI data center power demand to rising scrutiny over energy use, which supports ON’s pitch around reducing losses from the grid to the processor. These are practical tailwinds, not abstract themes.
The main external risk is that non-AI end markets recover more slowly than management expects. If automotive or industrial customers stay cautious, ON’s margin recovery could lag because utilization and mix are central to the earnings rebound. The company has already said it believes the cycle has bottomed. That is a strong statement, and the market will hold management to it.
Trailing free cash flow reached $2.10B with a 5.83% FCF yield, giving ON real financial flexibility even after a difficult 2025 reset.
Unlock the full analysis
Premium members get the complete breakdown — pick rationale, financial metrics, and recent earnings detail.
Get Full Access →Q1 2026 revenue rose to $1.513B and non-GAAP EPS was $0.64, while full-year 2025 gross margin compressed to 32.3% and net income fell to $121.0M.
Unlock the full analysis
Premium members get the complete breakdown — pick rationale, financial metrics, and recent earnings detail.
Get Full Access →Management guided Q2 revenue to $1.535B-$1.635B and non-GAAP EPS to $0.65-$0.77, with AI data center revenue expected to double in 2026.
Unlock the full analysis
Premium members get the complete breakdown — pick rationale, financial metrics, and recent earnings detail.
Get Full Access →ON trades at 68.1x trailing earnings, but the market is also pricing in growth from silicon carbide, GaN, and AI power rather than last year's depressed profits.
Unlock the full analysis
Premium members get the complete breakdown — pick rationale, financial metrics, and recent earnings detail.
Get Full Access →The report's valuation framework centers on a $108 fair value, with upside toward $126 and $144 if execution and margin recovery continue.
Unlock the full analysis
Premium members get the complete breakdown — pick rationale, financial metrics, and recent earnings detail.
Get Full Access →ON Semiconductor(ON) is one of those stocks where the right question is not whether the last twelve months looked messy. They did. The right question is whether the next twelve to eighteen months can look materially better than the last twelve. The evidence says yes. Q1 2026 showed 5% YoY revenue growth, 38.5% gross margin, a 14% YoY increase in PSG revenue, more than 30% sequential AI data center growth, and a Q2 guide that points to further sequential improvement.
The company also has strategic assets that are easy to understand and hard to dismiss: silicon carbide leadership in 900-volt EV platforms, a $1.5B-plus GaN design funnel, Treo products with 60% to 70% gross margin, and a role in AI power infrastructure that reaches from the grid to the processor. Those are not side projects. They are the backbone of the medium-term bull case.
The risks are real. Revenue remains below prior peaks, 2025 profitability was sharply lower, inventory is still elevated, and the stock already reflects some recovery optimism. But ON has the cash generation, liquidity, and product positioning to justify a constructive stance. For a balanced investor with a medium-term horizon, the setup supports a Buy rating with a fair value estimate of $108. In this part of the cycle, that is a respectable place to be: not blindly bullish, not timid, just paid for being early enough.
Get AI-powered research reports, daily market intelligence, and a personal analyst in your pocket.
Get Full Access →Not ready to subscribe? ·
ON Semiconductor Corporation (ON) falls sharply after announcing its largest acquisition ever, an all-stock deal for Synaptics. Investors are weighing the company’s expanded AI and Physical AI ambitions against dilution, integration risk, and a long closing timeline.
ON is being priced like a cyclical auto chip name even as its AI data-center power business is starting to show real traction. The selloff looks misaligned with a fresh product launch, management's growth commentary, and a September Analyst Day that could force a reframe.
ON Semiconductor is a cyclical power and sensing chipmaker working through a weak earnings base while building exposure to AI data center power and wide-bandgap semis. The stock looks like a selective recovery story with real cash flow support, but valuation still assumes a meaningful turnaround.