Marvell Technology, Inc. vs ON Semiconductor Corporation: Strategic Comparison

This in-depth comparison examines Marvell Technology, Inc. and ON Semiconductor Corporation across revenue, market value, business model, competitive positioning, and long-term growth strategy. Whether you are researching Marvell Technology, Inc. on its own, evaluating ON Semiconductor Corporation, or weighing the two companies side by side, the breakdown below highlights where each company leads and where the gap between Marvell Technology, Inc. and ON Semiconductor Corporation is widest.

On the headline numbers, Marvell Technology, Inc. reports annual revenue of $5.6B against $7.1B for ON Semiconductor Corporation, while their respective market capitalizations stand at $72.0B and $22.5B. Marvell Technology, Inc. is headquartered in United States and ON Semiconductor Corporation operates from United States, and those different home markets shape how each company competes.

Marvell Technology, Inc.: Marvell Technology's 1.6T PAM4 digital signal processors account for nearly 40% of the total bill of materials of a high-speed optical transceiver. That concentration — one company's silicon representing four out of every ten dollars in a critical AI data center component — is not accidental. It is the outcome of two decades of portfolio restructuring and the deliberate abandonment of lower-margin businesses in mobile, consumer storage, and commodity networking. The company generated $5.56 billion in revenue in fiscal year 2024, down from $6.5 billion the year prior, while its market capitalization sat at $72 billion. The revenue-to-market-cap ratio implies that investors are not valuing what Marvell makes today — they are pricing what it will make in fiscal 2025 and 2026, when AI-related data center revenue is projected to exceed $2 billion. Data center already accounted for over 65% of fiscal 2024 revenue, a structural shift from the company's historical dependence on enterprise networking. Marvell operates as the number two player in the hyperscale custom ASIC market behind Broadcom. The company designs custom XPUs — specialized processor accelerators — for Amazon and Google, the two largest hyperscalers investing most aggressively in custom silicon as an alternative to merchant GPUs. Each custom silicon design win requires over $1.5 billion in non-recurring engineering costs and mask set fees at TSMC's 3nm process node. That barrier is not a problem for Marvell; it is the moat. The company was founded in 1995 by Sehat Sutardja, Weili Dai, and Pantas Sutardja. CEO Matt Murphy, who arrived in 2016, has executed the portfolio transformation that shed mobile and consumer businesses and concentrated resources on data center and AI infrastructure.

ON Semiconductor Corporation: And its intelligent sensing group provides image sensors and actuator drivers for advanced driver-assistance systems. These are not modest ambitions. The Power Solutions Group is ON Semiconductor's largest and most strategically important segment, selling silicon carbide (SiC) products, discrete power devices, MOSFETs, and power modules for automotive electrification, industrial power conversion, and cloud power infrastructure. The margin structure varies significantly by segment and product category. ISG's 46.7% gross margin reflects the value of image sensors and signal processors in automotive safety and industrial applications. NXP follows closely with strength in in-vehicle networking MCUs, radar, and secure connectivity. Texas Instruments and Renesas round out the top five. In silicon carbide power semiconductors, the competitive landscape is more concentrated and rapidly evolving. Infineon is a major player with its CoolSiC product line and joint development labs with Hyundai-Kia. This is a more commoditized market where price competition is intense and margins are lower. The revenue contraction was broad-based, affecting all three operating segments and all geographic regions, driven by softening demand in automotive and industrial end markets following the post-pandemic inventory correction. This 170 basis point compression from 47.1% in FY2023 was driven by lower sales volumes, manufacturing underutilization, and negative operating use, partially offset by a reduction in lower-margin manufacturing services revenue. Shares outstanding declined 3.16% year-over-year to approximately 433 million, reflecting the buyback program and disciplined capital allocation. ON Semiconductor's revenue fell across all three segments and all geographic regions. STMicroelectronics holds an estimated 32.6% share of the SiC MOSFET market and has secured exclusive supply agreements with Stellantis. ON Semiconductor's top 20 customers represent approximately 40% of revenue, and one distributor accounted for 10% of FY2024 sales. The automotive qualification process for power semiconductors takes 2-3 years, and design wins are locked in for the vehicle platform lifecycle — typically 5-7 years. A customer designing an electric vehicle traction inverter can source SiC MOSFETs from PSG, gate drivers and power management ICs from AMG, and current sensing and thermal monitoring from ISG — all from ON Semiconductor with pre-qualified interoperability. This vertical integration provides supply chain resilience, cost control, and the ability to capture margin at the manufacturing level rather than paying foundry premiums. The third layer is customer design-in and long-term supply agreements. ON Semiconductor's products are qualified into automotive platforms with 2-3 year design cycles and 5-7 year production lifecycles. The fifth layer is the management team's track record. Texas Instruments has analog breadth but limited power semiconductor and SiC presence. First is silicon carbide expansion. Second is system-level solution selling. Third is the AI data center power opportunity. Fourth is portfolio rationalization. Fifth is capital return discipline. This is an ambitious but not impossible plan, built on three visible demand drivers and two structural margin levers. The first demand driver is electric vehicle adoption. The Volkswagen Scalable Systems Platform (SSP) agreement is particularly significant because it makes ON Semiconductor the primary supplier of a complete power box solution — not just discrete devices but integrated modules with system-level optimization. If Volkswagen produces 5-7 million vehicles annually on the SSP platform by 2030, and each vehicle uses $200-300 in ON Semiconductor power electronics, this single platform could generate $1-2 billion in annual revenue. The second demand driver is AI data center power infrastructure. Generative AI models require massive computational power, and the data centers that train and run these models consume enormous amounts of electricity. The third demand driver is advanced driver-assistance systems and autonomous driving. ON Semiconductor's intelligent sensing group provides CMOS image sensors, image signal processors, and time-of-flight sensors for ADAS cameras and LiDAR systems. By optimizing its internal manufacturing footprint and using external foundries for peak demand and commoditized products, ON Semiconductor has reduced capital expenditures from 19.1% of revenue in 2023 to a target of 11% in 2027. The second margin lever is the mix shift toward silicon carbide. If revenue remains flat or declines in 2025, the 2027 targets become mathematically more difficult to achieve. The early years were challenging. The dot-com crash of 2001 hit the semiconductor industry hard, and ON Semiconductor — heavily exposed to communications and consumer markets — struggled to maintain profitability. But the defining acquisition was Fairchild Semiconductor in 2016. The integration was successful, and ON Semiconductor emerged as a major player in power semiconductors with industry-leading cost structure. The results have been significant. The 2024 downturn tested this progress. Revenue fell 14.2% as automotive and industrial demand softened.

Marvell Technology, Inc. and ON Semiconductor Corporation pursue distinct approaches to generating revenue, and understanding how each company operates is the foundation of any fair comparison between Marvell Technology, Inc. and ON Semiconductor Corporation.

Marvell Technology, Inc. business model: The financial structure of a custom silicon win is highly lucrative but capital intensive; Marvell charges the hyperscaler substantial non-recurring engineering (NRE) fees, often exceeding $50 million per design, to cover the massive mask set costs and engineering hours required to bring the chip to tape-out. Marvell's PAM4 DSPs are the industry standard for 400G, 800G, and the upcoming 1.6T optical modules, giving the company immense pricing power and a near-monopoly position in a market where the DSP accounts for nearly 40% of the total bill of materials of a high-speed optical transceiver. Marvell's third major revenue engine is the enterprise networking and storage controller market, where the company sells merchant silicon for Ethernet switches, PHYs, and enterprise solid-state drive controllers. Marvell's cost of goods sold is dominated by the wafer procurement costs paid to TSMC, the packaging and testing fees for advanced 2.5D and 3D chiplet integration, and the amortization of the massive IP portfolio the company has developed or acquired.

ON Semiconductor Corporation business model: AMG sells analog products, application-specific integrated circuits (ASICs), logic and isolation products, non-volatile memory, foundry services, gate drivers, and large-scale integration (LSI) devices. AMG's higher gross margin reflects the value-added nature of analog and mixed-signal design, where product differentiation and customer-specific solutions command pricing power. ISG sells actuator drivers, CMOS image sensors, image signal processors, single photon detectors, short-wavelength infrared products, and indirect time-of-flight sensors for automotive sensing, industrial automation, and consumer applications. AMG's 50.1% gross margin reflects the design-intensive nature of analog and mixed-signal products, where proprietary circuit design and customer qualification create pricing power. Non-GAAP operating margin was approximately 32.3% for FY2023 and would have been higher in FY2024 excluding restructuring charges. The SiC market is also facing pricing pressure as Chinese suppliers, backed by national industrial policy and a 25% domestic content mandate by 2025, accelerate into cockpit, ADAS, and SiC power domains. If Chinese competitors achieve scale in SiC substrates and devices, the pricing power that ON Semiconductor currently enjoys could erode. SiC devices command premium pricing and higher margins than silicon power semiconductors. If SiC pricing erodes faster than expected, the margin expansion story weakens.

The durability of a company's moat often decides long-term winners. Here is how the competitive advantages of Marvell Technology, Inc. stack up against those of ON Semiconductor Corporation.

Marvell Technology, Inc. competitive advantage: The physical architecture of the modern artificial intelligence data center does not rely solely on Nvidia's GPUs; it is fundamentally enabled by a silent, multi-billion-dollar silicon ecosystem engineered by a single fabless semiconductor company that completely reinvented itself over the last eight years. As AI training clusters scaled from thousands of GPUs to hundreds of thousands, the optical interconnect became the primary bottleneck, and Marvell's DSPs became the mandatory tollbooth that every hyperscaler had to pay to achieve the necessary bandwidth density. Today, Marvell operates in a highly concentrated, extremely lucrative oligopoly within the custom silicon market, competing primarily with Broadcom to design bespoke, application-specific integrated circuits for hyperscalers who demand silicon optimized for their specific software stacks rather than off-the-shelf merchant parts. The fundamental mechanism of how Marvell makes money in its most lucrative segment — custom compute silicon — relies on the hyperscalers' strategic imperative to reduce their dependence on Nvidia's merchant GPUs and the exorbitant margins associated with them. As AI clusters scale to hundreds of thousands of accelerators, the electrical signals generated by the compute chips must be converted into light to travel across the data center fabric without latency degradation. The business model for electro-optics is characterized by high volume, rapid design cycles, and deep integration with the optical module manufacturers and the hyperscalers' networking teams. This platform strategy creates massive switching costs; once a hyperscaler designs its data center architecture around Marvell's custom compute and optical interconnect ecosystem, migrating to a competitor's silicon for the next generation would require a complete redesign of the network fabric, a risk that cloud providers are unwilling to take. If this assumption holds true, Marvell's model is a highly profitable, structurally advantaged tollbooth on the global data economy; if hyperscalers decide to bring custom silicon design entirely in-house, or if a radical breakthrough in optical interconnects bypasses the need for traditional DSPs, the fundamental economic rationale for Marvell's premium valuation would be severely compromised. Marvell operates in a highly concentrated, extremely lucrative oligopoly within the custom silicon market, competing primarily with Broadcom to design bespoke, application-specific integrated circuits that allow hyperscalers to reduce their dependence on Nvidia's merchant GPUs and achieve maximum performance-per-watt for specific AI training workloads. However, Marvell has successfully defended its position in the electro-optics market by using its Inphi heritage to dominate the PAM4 DSP market for 800G and 1.6T optical transceivers, a segment where Nvidia has no meaningful presence, ensuring that even if hyperscalers adopt Nvidia's compute and networking stack, they are still forced to purchase Marvell's optical DSPs to connect the racks together. The competitive narrative is further complicated by the fact that Marvell and Broadcom are entirely dependent on the same upstream supply chain for advanced packaging and TSMC wafer allocation, meaning that competitive advantages are often dictated by who can secure the most CoWoS capacity and the most advanced 3nm process nodes during periods of intense industry congestion. The competitive advantage in the data infrastructure market is no longer about who can manufacture the cheapest component, but about who can provide the most comprehensive, system-level platform that allows hyperscalers to optimize the entire signal chain from the compute die to the optical fiber; Marvell's victory in integrating its custom compute, networking, and electro-optic portfolios has established it as the premier architectural partner for the AI revolution, forcing Broadcom to compete on scale and Nvidia to compete on closed-loop ecosystem lock-in, ensuring that Marvell will dictate the pace of innovation in the high-bandwidth interconnect market for the foreseeable future. The financial narrative of Marvell is inextricably linked to the capital expenditure cycles of its top hyperscaler customers; when these companies increase their AI infrastructure capex by even 10%, Marvell's data center revenue can grow by 25% due to the high content per rack of its custom silicon and optical DSPs, but when they pause to digest inventory, Marvell's overall revenue collapses with equal velocity. This vertical integration poses a severe risk to Marvell's enterprise networking and DPU businesses, as hyperscalers who purchase hundreds of thousands of Nvidia GPUs are increasingly incentivized to adopt Nvidia's proprietary networking fabric to guarantee maximum cluster performance, thereby marginalizing Marvell's merchant Ethernet switch silicon and OCTEON DPUs. The company has no control over the internal strategic decisions of these hyperscalers, and the intense, zero-sum competition between Marvell and Broadcom for these custom design wins means that a single lost bid can depress the company's growth trajectory for three to four years, the typical lifecycle of a custom ASIC program. Marvell faces intense geopolitical and supply chain risks due to its absolute reliance on TSMC for the manufacturing of its most advanced 5nm and 3nm custom silicon; any disruption at TSMC's facilities in Taiwan, whether from natural disaster, geopolitical conflict, or supply chain bottlenecks in advanced packaging technologies like CoWoS, would immediately halt Marvell's ability to deliver its highest-margin products to its hyperscale customers. Finally, the massive capital expenditure required to maintain its technological lead in electro-optics and custom silicon represents a continuous financial burden; the transition to 1.6T optics and the development of co-packaged optics require billions of dollars in R&D, straining the company's free cash flow and limiting its financial flexibility to pursue additional significant acquisitions or weather an extended downturn in the hyperscaler capital expenditure cycle. This is not merely a product portfolio advantage; it is a fundamental architectural moat derived from the physical realities of scaling AI data centers, where the performance of the compute chips is entirely bottlenecked by the bandwidth and latency of the optical interconnects that link them together. By owning the PAM4 DSP market for 800G and 1.6T optical transceivers through its Inphi acquisition, Marvell controls the exact point where electrical signals from the custom XPUs must be converted into light, giving the company unprecedented visibility into the hyperscalers' network traffic patterns and the ability to co-optimize the custom compute silicon with the optical fabric. Marvell's position in the custom silicon market is reinforced by its deep, strategic integration with Arm's Neoverse compute subsystems and its exclusive access to TSMC's most advanced 3nm and 2nm process nodes, allowing the company to offer hyperscalers a complete, chiplet-based design platform that integrates high-bandwidth memory controllers, PCIe Gen 6 PHYs, and ultra-ethernet SerDes into a single, massive system-on-chip. This platform approach creates immense switching costs; once a hyperscaler like Amazon Web Services designs its Trainium accelerator around Marvell's custom compute and optical interconnect ecosystem, migrating to a competitor's silicon for the next generation would require a complete redesign of the network fabric and the software stack, a risk that cloud providers are fundamentally unwilling to take. This combination of proprietary electro-optic physics, deep TSMC manufacturing priority, and the massive capital barriers of custom ASIC design creates a competitive advantage that is virtually impossible for a new entrant to replicate, and forces existing competitors to spend billions of dollars just to reach the baseline of Marvell's current generation platform capabilities. Marvell is also pursuing a strategic expansion of its software-defined networking partnerships, working closely with companies like Arista Networks and Cisco to ensure that its Teralynx Ethernet switch silicon is deeply integrated and optimized within the cloud data center fabrics of the future, creating a smooth hardware-software ecosystem that locks in hyperscaler preference. The company is also exploring the integration of advanced thermal management technologies into its custom silicon platforms, allowing hyperscalers to push the power envelope of their AI clusters beyond 1000W per rack without degrading performance, a crucial selling point for cloud providers who are constrained by the thermal limits of their data center facilities. Marvell's roadmap calls for the continuous iteration of its custom compute platform, moving from the current 5nm XPUs to 3nm designs that integrate next-generation Arm Neoverse cores, HBM4 memory controllers, and 224G ultra-ethernet SerDes, allowing hyperscalers to double the compute density per rack without increasing the power envelope. The company anticipates that the transition to co-packaged optics will fundamentally alter the economics of the data center, allowing hyperscalers to achieve 10x higher bandwidth density at 50% lower power consumption, a value proposition that is critical as data centers hit the physical limits of their electrical grid connections. The company also foresees a growing role for its OCTEON data processing units in the edge AI market, where Marvell is developing specialized, high-performance DPUs optimized for the harsh environmental conditions of telecommunications hubs and enterprise edge data centers, attempting to capture a share of the inference market that exists outside the massive hyperscale facilities.

ON Semiconductor Corporation competitive advantage: Its power management solutions address the AC-DC conversion challenges of hyperscale data centers building out generative AI infrastructure. Chinese companies are moving rapidly into SiC power devices, automotive MCUs, and image sensors, and while they currently lag in technology and reliability, their cost advantages and government support could reshape competitive dynamics over the next 5 years. The Volkswagen supply agreement for the Scalable Systems Platform (SSP) is a major win, but if Volkswagen delays or scales back the SSP platform, ON Semiconductor's revenue pipeline would be affected. ON Semiconductor's single most defensible moat is its vertically integrated manufacturing footprint combined with a portfolio breadth that spans power, analog, and sensing technologies — a combination that enables the company to serve as a "one-stop shop" for automotive and industrial customers who need optimized system-level solutions rather than discrete components. This system-level optimization reduces the customer's bill-of-materials complexity, improves time-to-market, and creates switching costs that lock in revenue across product generations. The second layer of the moat is the Fab Right manufacturing strategy. The company has secured multi-year long-term supply agreements (LTSAs) with Volkswagen, BMW, Hyundai-Kia, Zeekr, and Stellantis for SiC power devices, creating revenue visibility and customer lock-in. The company's SiC devices reduce module size by 40% and weight by 52% compared to traditional silicon IGBT solutions, creating measurable performance advantages that OEMs cannot ignore. The combination of these five layers — portfolio breadth, vertical manufacturing, customer lock-in, SiC technology leadership, and management expertise — creates a moat that competitors cannot replicate quickly. Infineon has scale but lacks ON Semiconductor's sensing portfolio. ON Semiconductor's position as the only major semiconductor company with significant presence across power, analog, and sensing — combined with vertical manufacturing and automotive design-in relationships — creates structural advantages that should persist through the cycle. This system-level approach increases revenue per vehicle, deepens customer relationships, and creates higher switching costs.

Future prospects matter as much as current results. The growth strategies below explain how Marvell Technology, Inc. and ON Semiconductor Corporation each plan to expand from here.

Marvell Technology, Inc. growth strategy: The narrative of Marvell is no longer that of a diversified semiconductor company fighting for scraps in the consumer and enterprise markets; it is the story of a highly focused, technologically elite design house that has successfully positioned itself as the indispensable co-architect of the AI revolution, proving that in the race to build the infrastructure of the future, the companies that control the custom silicon and the optical interconnects will capture the vast majority of the economic value. The economics of Marvell's business are defined by massive upfront research and development expenditures, extreme reliance on advanced semiconductor manufacturing partners like TSMC, and a revenue structure that is increasingly dominated by high-margin, multi-year custom silicon design wins and recurring electro-optic component shipments. The financial architecture of the company is designed to maximize cash flow during the upcycles of the data center buildout, using the massive free cash flow generated by high-margin custom silicon and electro-optics to fund aggressive share repurchase programs and invest in the next generation of silicon photonics and co-packaged optics technologies. The narrative of Marvell is no longer that of a diversified semiconductor company fighting for scraps in the consumer and enterprise markets; it is the story of a highly focused, technologically elite fabless manufacturer that has successfully positioned itself as the co-architect of the AI revolution, proving that in the race to build the infrastructure of the future, the companies that control the custom silicon and the optical interconnects will capture the vast majority of the economic value. Broadcom currently holds the dominant position in this segment, using its massive scale and deep historical relationships to capture the majority of the custom AI accelerator market, but Marvell has successfully closed the technological gap by aggressively investing in its Arm-based compute subsystems and advanced chiplet integration capabilities, allowing it to win critical second-source and next-generation design bids that hyperscalers require to maintain supply chain leverage. Nvidia's strategy is to offer a complete, closed-loop compute and networking stack, bundling its GPUs with its proprietary networking silicon to guarantee maximum cluster performance, a move that directly threatens Marvell's merchant Ethernet switch silicon and OCTEON DPU businesses. In the enterprise storage controller market, Marvell faces intense competition from Intel, Microchip, and a host of Asian fabless designers, but the company has strategically de-emphasized this segment, choosing to focus its engineering resources on the high-margin data center and electro-optics markets rather than engaging in a suicidal price war in the commoditized merchant silicon space. Marvell's growth strategy for the next three years is laser-focused on the aggressive commercialization and market penetration of its 1.6T electro-optic DSP platform and its next-generation 3nm custom compute silicon, aiming to capture 100% of the new optical interconnect demand in the hyperscale AI market by offering bandwidth densities that competitors simply cannot match. The company's primary strategic initiative is the rapid scaling of manufacturing yield for its 1.6T PAM4 DSPs, which requires the complex integration of advanced analog front-ends and high-speed SerDes into the high-volume production lines at TSMC; achieving a 90% manufacturing yield on these DSPs is the single most important operational metric for the company, as it directly dictates the gross margin and the ability to fulfill the massive backlog of orders from the optical module manufacturers. To accelerate this growth, Marvell is investing heavily in the expansion of its silicon photonics research and development, forging strategic partnerships with specialized laser manufacturers to ensure an uninterrupted supply of the continuous-wave lasers required for co-packaged optics, a critical bottleneck that could constrain growth if not managed properly. The second pillar of the growth strategy is the penetration of the custom silicon market with its comprehensive Arm-based compute subsystem platform, specifically targeting the next-generation AI inference accelerators at Microsoft and Meta, allowing Marvell to win design bids that require deep integration of machine learning tensor cores with high-bandwidth memory and ultra-ethernet networking. The company's growth strategy also includes a deliberate and managed exit from the low-margin consumer and legacy carrier markets, reallocating those engineering resources to the production of higher-margin data center and electro-optic products, a portfolio optimization move that will artificially suppress unit growth but dramatically improve the overall profitability and return on invested capital. Marvell is investing in advanced packaging technologies, working directly with TSMC to secure allocation for CoWoS and InFO packaging, ensuring that its massive custom XPUs can be integrated with HBM3E memory stacks without supply chain constraints. Marvell's management expects the data center segment to grow to represent over 75% of total revenue by fiscal 2027, as the company continues to exit the low-margin consumer and legacy carrier markets, effectively transforming Marvell from a diversified semiconductor manufacturer into a pure-play data infrastructure platform for the AI cloud. However, the future outlook is not without significant risks; if Nvidia successfully bundles its networking and DPU silicon with its GPUs to create a closed-loop ecosystem that marginalizes merchant Ethernet, or if a breakthrough in wireless optical interconnects bypasses the need for traditional DSPs, Marvell's massive investment in electro-optics and custom silicon could be rendered obsolete, making the successful execution of the 1.6T and 3nm roadmaps an absolute existential imperative for the company's long-term survival. The founding philosophy of the company was radically different from the established semiconductor giants of the era; while Intel and AMD were focused on the microprocessor, and Cisco was dominating the routing market, Marvell focused entirely on the physical layer — the analog and mixed-signal silicon that actually moved the data across the copper wires. The team worked 100-hour weeks, operating on a culture of extreme frugality and technical perfectionism, focusing entirely on creating a gigabit Ethernet PHY (physical layer) chip that could be manufactured at a cost low enough to be deployed in every enterprise switch and network interface card on the planet.

ON Semiconductor Corporation growth strategy: The strategic bet is that ON Semiconductor can grow revenue at 10-12% annually through 2027 — three times the semiconductor industry average — while expanding gross margins from 45.4% to a target of 53% and operating margins from approximately 25% to 40%. The 2024 downturn, driven by soft automotive and industrial demand, tested this strategy. The real question for investors is whether the cyclical recovery in automotive and industrial demand, combined with the secular ramp of SiC adoption in EVs and AI power infrastructure, will deliver the revenue growth and margin expansion that management has staked its credibility on by 2027. If this segment's growth were to stall — whether due to slower EV adoption, competition from Infineon or STMicroelectronics in SiC, or a technological shift away from silicon carbide — ON Semiconductor would lose not only its largest revenue stream but also its highest-potential growth engine, and the company's path to its 2027 margin targets would be blocked. The foundry services component, which includes manufacturing services at the EFK location, carries lower margins and was a deliberate reduction target as part of the Fab Right strategy. This segment is the smallest but serves high-growth markets including advanced driver-assistance systems (ADAS), autonomous driving, and industrial machine vision. The company's capital allocation reflects its strategic priorities and financial discipline. Rohm and Denso are growing in SiC MOSFETs for Japanese and Asian EV markets. The company's competitive strategy in SiC is to use its automotive customer relationships and system-level integration capabilities to win platform-level design wins rather than competing solely on device specifications. The Volkswagen SSP power box agreement and BMW drivetrain LTSA are examples of this strategy — ON Semiconductor is not just selling SiC dies but complete power solutions that include modules, gate drivers, and thermal management. The company's strategy is to focus on automotive-grade and industrial-grade products with higher reliability requirements, where its vertical manufacturing and AEC-Q101 qualification capabilities create differentiation. ON Semiconductor's strategy is to focus on application-specific analog products for automotive and industrial markets rather than competing with TI in broad-based analog. ON Semiconductor's response is to deepen relationships with Western and Korean automotive OEMs, expand European manufacturing through the Czech Republic investment, and maintain technology leadership through R&D investment in next-generation SiC and sensing technologies. The weighted-average interest rate is exceptionally low, reflecting the company's investment-grade credit profile and the low-rate environment in which the convertible notes were issued. Return on assets was 11.2%, return on equity was 17.8%, and return on invested capital was approximately 14.5% — all healthy metrics that reflect the company's asset efficiency despite the cyclical downturn. Achieving these targets would require revenue to grow at roughly twice the rate of the overall semiconductor industry while expanding margins by 760 basis points in gross margin and 1,500 basis points in operating margin. Whether the cyclical recovery in automotive and industrial demand, combined with the secular ramp of SiC adoption, can deliver this growth trajectory. The second major challenge is competition in silicon carbide power semiconductors, the company's highest-growth and highest-margin product category. Rohm and Denso are growing their SiC MOSFET presence for EV inverters. Similarly, the BMW and Hyundai-Kia relationships are critical to the SiC growth story. The Fab Right strategy, while successful in reducing capital intensity, also creates dependency on external foundry partners for peak demand periods. If the downturn extends, the company may need to preserve cash rather than return it to shareholders, which could disappoint investors who have priced in the capital return program. Wolfspeed has SiC focus but lacks the diversified revenue base to weather downturns. ON Semiconductor's growth strategy for 2025-2027 is organized around the "Fab Right" manufacturing model and a product portfolio pivot toward silicon carbide power semiconductors, intelligent power management for AI data centers, and advanced sensing for automotive ADAS. The strategy has five pillars. The new facility would be one of the largest private investments in Czech history and is subject to regulatory approval and government subsidies. This investment complements the company's existing SiC capacity in South Korea and the United States and creates a geographically diversified manufacturing footprint that reduces supply chain risk. The company has also secured a long-term wafer agreement with GTAT for SiC substrates, addressing the substrate supply constraint that has limited SiC industry growth. The Volkswagen SSP agreement exemplifies this strategy — ON Semiconductor is the primary supplier of a complete power box solution for next-generation traction inverters, not just a vendor of discrete SiC devices. The company is expanding its power management portfolio to address the AC-DC conversion, DC-DC regulation, and power delivery challenges of hyperscale AI data centers. This capital return program is designed to enhance shareholder value while maintaining sufficient liquidity for growth investments. The overall growth strategy is disciplined. ON Semiconductor is not pursuing growth for its own sake — it is pursuing growth in segments where its vertical manufacturing, automotive design-in relationships, and system-level integration capabilities create defensible margins. ON Semiconductor's specific bet for the next three years is that the electrification of vehicles and the power efficiency demands of AI data centers will drive 10-12% revenue CAGR while the company's Fab Right strategy and SiC mix shift expand gross margins from 45.4% to 53% and operating margins from approximately 25% to 40%. As vehicles add more cameras and higher-resolution sensors for Level 2+ and Level 3 autonomy, the addressable market for automotive image sensors grows. The first margin lever is the Fab Right manufacturing strategy. As SiC grows from a smaller percentage of revenue today to a larger percentage by 2027, the overall gross margin expands. The company would need to accelerate growth in 2026-2027 to compensate, which depends on EV adoption rates, AI data center buildout timing, and the company's ability to win additional design wins. STMicroelectronics, Infineon, and Wolfspeed are all investing heavily in SiC capacity, and Chinese competitors are emerging with government support. This strategy has already produced the Volkswagen SSP win and could be replicated with other OEMs. The company's stock price languished, and management focused on cost reduction and operational efficiency rather than growth. In 2006, the company acquired LSI Logic's consumer and computing products division, adding custom ASIC capabilities. In 2008, it acquired Catalyst Semiconductor, expanding its portfolio of analog and memory products. In 2010, it acquired California Micro Devices, adding protection and filtering products for mobile devices. In 2011, it acquired SANYO Semiconductor, gaining significant manufacturing capacity in Japan and a foothold in the automotive and industrial markets. At ON Semiconductor, El-Khoury applied the same playbook: focus on high-margin, differentiated products; deepen automotive customer relationships; and invest in secular growth markets. Non-GAAP operating margin expanded from the mid-teens to over 32%.

A closer look at the financial trajectory of Marvell Technology, Inc. and ON Semiconductor Corporation rounds out the comparison.

Marvell Technology, Inc.: Revenue fell from $6.5 billion in fiscal 2023 to $5.56 billion in fiscal 2024 — a counterintuitive decline for a company positioned in the fastest-growing segment of the semiconductor market. The explanation is inventory correction in enterprise networking, carrier infrastructure, and legacy storage, businesses that compressed while the data center segment was accelerating. The fiscal 2024 revenue number masks a divergence: one set of businesses declining sharply while another set was growing rapidly. Net income reached $618 million in fiscal 2024 against $5.56 billion in revenue. Market capitalization of $72 billion reflects the projected fiscal 2025 and 2026 data center revenue trajectory, not the trailing twelve months. Fiscal 2025 AI-related revenue was projected to exceed $2 billion, driven by custom ASIC production for Amazon and Google and PAM4 DSP sales into the hyperscale optical interconnect market. R&D expenditure exceeded $1.6 billion in fiscal 2024, representing nearly 29% of revenue. Custom silicon design wins at Amazon and Google require non-recurring engineering fees often exceeding $50 million per design — costs Marvell charges to the hyperscaler while retaining the intellectual property and the manufacturing relationship with TSMC. The financial structure of each custom win is front-loaded on NRE fees, then transitions to volume revenue as production scales. Revenue was $5.52 billion in fiscal 2022, $6.5 billion in fiscal 2023, and $5.56 billion in fiscal 2024. The trajectory is not linear. But the data center composition shift — from a minority of revenue to over 65% — is the financial fact that explains a $72 billion valuation on $5.56 billion in trailing revenue.

ON Semiconductor Corporation: In fiscal year 2024, a semiconductor company that most consumers have never heard of generated $1.21 billion in free cash flow — a 3x increase from the $438.4 million it produced in 2023 — despite revenue declining 14.2% to $7.08 billion. That company, ON Semiconductor Corporation, achieved this cash flow surge not by cutting research and development but by executing a "Fab Right" manufacturing strategy that reduced capital expenditures from $1.54 billion in 2022 to $694 million in 2024 while maintaining a 45.4% gross margin that would have been unimaginable during the company's prior downturns, when margins compressed to approximately 30%. El-Khoury, who immigrated to the United States from Lebanon at age 17 and rose from application engineer at Cypress Semiconductor to CEO of that company before its $9.3 billion acquisition by Infineon in 2020, has restructured ON Semiconductor around two secular megatrends: the electrification of vehicles and the power efficiency demands of AI data centers. They require the company to maintain its 35-40% estimated market share in SiC power devices, successfully ramp its $2 billion Czech Republic SiC manufacturing expansion, and fend off competition from Infineon, STMicroelectronics, and Wolfspeed in a market where design wins are locked in years before vehicles reach production. Revenue fell across all three segments: Power Solutions Group declined 13.7% to $3.35 billion, Analog and Mixed-Signal Group fell 14.7% to $2.61 billion, and Intelligent Sensing Group dropped 14.5% to $1.13 billion. Net income fell 28% to $1.57 billion. Yet the company maintained pricing discipline, reduced inventory from $2.11 billion to normalized levels, and returned 54% of its free cash flow to shareholders through $650 million in share repurchases. With $2.69 billion in cash, $3.35 billion in long-term debt, and total shareholders' equity of $8.81 billion, ON Semiconductor's balance sheet supports both its growth investments and its capital return program. ON Semiconductor Corporation is a $7.08 billion revenue semiconductor company headquartered in Scottsdale, Arizona, that designs intelligent power and sensing solutions for automotive, industrial, cloud power, and IoT markets. The company operates three segments: Power Solutions Group (47% of revenue, $3.35B in FY2024), Analog and Mixed-Signal Group (37%, $2.61B), and Intelligent Sensing Group (16%, $1.13B). Despite a 14.2% revenue decline in FY2024 due to cyclical weakness in automotive and industrial markets, the company generated $1.21 billion in free cash flow — a 3x year-over-year increase — while maintaining a 45.4% gross margin. ON Semiconductor supplies EliteSiC devices to Volkswagen, BMW, Hyundai-Kia, Zeekr, and Stellantis, and is investing up to $2 billion in a Czech Republic SiC manufacturing facility. In fiscal year 2024, the company's $7.08 billion in revenue broke down as follows: Power Solutions Group (PSG) contributed $3.35 billion (47.3% of total revenue), Analog and Mixed-Signal Group (AMG) contributed $2.61 billion (36.8%), and Intelligent Sensing Group (ISG) contributed $1.13 billion (15.9%). This segment generated $3.35 billion in FY2024 revenue, down 13.7% from $3.88 billion in FY2023, with gross profit of $1.38 billion at a 41.3% gross margin. The revenue decline was driven by weakness across all three PSG divisions: Multi-Market Power fell $250.8 million, Industrial Power fell $162.2 million, and Automotive Power fell $119.1 million, all primarily due to decreased demand in automotive and industrial end markets. The segment's profitability is highly sensitive to manufacturing use — when demand weakens, fixed costs at the company's internal fabrication facilities create margin pressure, which is why the 170 basis point gross margin compression across the company in 2024 was concentrated in underutilized manufacturing assets. The Analog and Mixed-Signal Group contributed $2.61 billion in FY2024 revenue, down 14.7% from $3.06 billion in FY2023, with gross profit of $1.31 billion at a 50.1% gross margin — the highest of the three segments. The revenue decline was driven by weakness in the Power Management Division (down $269.1 million), Sensor Interface Division (down $101.5 million), and Integrated Circuit Division (down $77.4 million), again due to soft automotive and industrial demand. The Intelligent Sensing Group contributed $1.13 billion in FY2024 revenue, down 14.5% from $1.32 billion in FY2023, with gross profit of $525.4 million at a 46.7% gross margin. The revenue decline was driven by the Industrial and Consumer Solutions Division (down $107.8 million) and the Automotive Sensing Division (down $82.7 million). In FY2024, distributors accounted for $3.76 billion (53.1%) of revenue and direct customers accounted for $3.32 billion (46.9%). PSG's 41.3% gross margin in FY2024 reflects the capital intensity of power semiconductor manufacturing and the competitive pricing in discrete and MOSFET products, offset by premium pricing on SiC devices. The company's overall GAAP gross margin was 45.4% for FY2024, down 170 basis points from 47.1% in FY2023, primarily due to lower sales volumes and manufacturing underutilization. Non-GAAP gross margin was 45.5%, reflecting minimal impact from acquisition-related amortization. Operating expenses were $1.45 billion for FY2024, up from $1.34 billion in FY2023, driven by $133.9 million in restructuring, asset impairments, and other charges related to the 2024 business realignment that affected approximately 1,600 employees. Excluding these special items, non-GAAP operating expenses were $1.25 billion, representing 17.6% of revenue — well below the 2027 target of 13% as a percentage of revenue. In FY2024, ON Semiconductor generated $1.91 billion in operating cash flow and $1.21 billion in free cash flow, up from $438.4 million in FY2023. Capital expenditures were $694 million, down from $1.54 billion in 2022, as the company completed its major capacity buildout and shifted to a more capital-efficient Fab Right model. The company returned 54% of its free cash flow to shareholders through $650 million in share repurchases at a weighted-average price of $71.21 per share, and it has a $3 billion share repurchase authorization in place. The company has no meaningful debt maturities in the next 12 months, and its $3.35 billion in long-term debt carries a weighted-average interest rate well below 3% thanks to the 0.50% convertible notes due 2029 and 0% notes due 2027. The balance sheet is strong: $2.69 billion in cash, $2.99 billion in cash plus short-term investments, and a current ratio of 5.06. The 2027 financial model targets $3.5-4.0 billion in free cash flow, 25-30% free cash flow margin, and returning 50% of free cash flow to shareholders. This would represent a near-tripling of FY2024 free cash flow and requires revenue to grow at 10-12% annually while gross margins expand 760 basis points to 53% and operating margins expand 1,500 basis points to 40%. The key driver of this expansion is expected to be the mix shift toward higher-margin SiC products, which currently carry EBITDA margins exceeding 40% according to industry analysis, and the operating leverage from revenue growth on a right-sized cost base. ON Semiconductor Corporation generated $7.08 billion in revenue for fiscal year 2024, a 14.2% decline from the prior year, yet produced $1.21 billion in free cash flow — a 3x year-over-year increase — while maintaining a 45.4% gross margin through a cyclical downturn that would have crushed margins in previous cycles. CEO Hassane El-Khoury, who led Cypress Semiconductor through its $9.3 billion sale to Infineon before joining ON Semiconductor in December 2020, has restructured the company around silicon carbide (SiC) power semiconductors that are designed into the next-generation electric drivetrains of Volkswagen, BMW, Hyundai-Kia, Zeekr, and Stellantis. The company's 2027 financial model targets 10-12% revenue CAGR, 53% gross margins, and 40% operating margins — ambitious goals that depend on the cyclical recovery in automotive and industrial demand, the secular ramp of SiC adoption, and the successful execution of the $2 billion Czech Republic manufacturing expansion. With $2.69 billion in cash, $3.35 billion in low-cost long-term debt, and a $3 billion share repurchase authorization, ON Semiconductor's balance sheet supports both growth investment and aggressive capital return. In the automotive semiconductor market — valued at $68.68 billion in 2024 and projected to grow to $133 billion by 2030 at an 11.4% CAGR — ON Semiconductor competes with Infineon Technologies, NXP Semiconductors, STMicroelectronics, Texas Instruments, and Renesas Electronics. Infineon is the clear leader with more than $8 billion in automotive sales in 2024, commanding approximately 13% market share and dominating Si/SiC power modules, drivers, and microcontrollers. Texas Instruments is the dominant player with approximately $17.5 billion in analog revenue and a manufacturing scale that ON Semiconductor cannot match. ON Semiconductor Corporation reported revenue of $7.08 billion for fiscal year 2024, a 14.2% decline from $8.25 billion in FY2023 and a 15.0% decline from the FY2022 peak of $8.33 billion. GAAP net income attributable to ON Semiconductor was $1.57 billion ($3.64 per diluted share), down 28.0% from $2.18 billion ($4.89 per share) in FY2023. Non-GAAP net income was $1.70 billion ($3.98 per share), down 24.5% from $2.26 billion ($5.16 per share) in FY2023. The divergence between GAAP and non-GAAP is modest — $136.1 million in share-based compensation, $52.0 million in acquisition-related intangible amortization, and $133.9 million in restructuring and asset impairment charges — reflecting a company with limited accounting complexity. Gross profit was $3.22 billion, yielding a GAAP gross margin of 45.4% and a non-GAAP gross margin of 45.5%. Operating income was $1.77 billion, yielding an operating margin of 25.0% on a GAAP basis. The company's balance sheet as of December 31, 2024, showed total assets of $14.09 billion, total liabilities of $5.28 billion, and total stockholders' equity of $8.81 billion. Cash and cash equivalents were $2.69 billion, short-term investments were $300 million, and total liquidity was approximately $2.99 billion. Long-term debt was $3.35 billion (net) or $3.38 billion (gross), consisting of a $375 million revolving credit facility due 2028, $1.5 billion in 0.50% convertible notes due 2029, $804.9 million in 0% notes due 2027, and $700 million in 3.875% notes due 2028. Interest expense was $62.3 million for FY2024, down from $74.8 million in FY2023. Operating cash flow was $1.91 billion for FY2024, down modestly from $1.98 billion in FY2023 despite the revenue decline, demonstrating strong working capital management. Capital expenditures were $694 million, down 54.9% from $1.54 billion in FY2022 and down 10.6% from $776 million in FY2023, reflecting the completion of major capacity buildouts and the shift to the Fab Right capital-efficient model. Free cash flow was $1.21 billion, a 176.5% increase from $438.4 million in FY2023 and a 3x increase that management highlighted as a key achievement. The company returned 54% of this free cash flow to shareholders through $650 million in share repurchases, buying approximately 9.1 million shares at a weighted-average price of $71.21. As of December 31, 2024, the company had $1.79 billion remaining under its $3 billion share repurchase authorization. The first quarter 2025 results, reported in May 2025, showed revenue of $1.45 billion, down 22.4% year-over-year, with GAAP gross margin of 20.3% and non-GAAP gross margin of 40.0%. The GAAP operating margin was negative 39.7% due to $431.5 million in non-cash asset impairment charges related to the 2025 Manufacturing Realignment Program, but non-GAAP operating margin was 18.3%. Free cash flow was $454.7 million, up 71.7% year-over-year, representing 31.4% of revenue. The company returned 66% of Q1 free cash flow to shareholders through share repurchases. The 2027 financial model targets revenue of approximately $10-11 billion (implied by 10-12% CAGR from $7.08B), gross margin of 53%, operating margin of 40%, capital expenditures of 11% of revenue, and free cash flow of $3.5-4.0 billion. The company's FY2024 operating income of $1.77 billion was down 30.4% from $2.54 billion in FY2023, and while cost management partially offset the revenue decline, the operating leverage of a semiconductor manufacturing business means that revenue recovery is essential for margin recovery. Infineon Technologies leads the automotive semiconductor market with more than $8 billion in automotive sales in 2024 and dominates Si/SiC power modules and drivers. The company's $2 billion planned investment in a Czech Republic SiC facility — described by management as potentially "one of the largest private investments in the history of the Czech Republic" — is subject to regulatory approval and government subsidies. The company has $3.35 billion in long-term debt and $2.69 billion in cash, creating a net debt position. While the debt is low-cost (0% and 0.50% convertible notes), the company has been aggressive with share repurchases — $650 million in FY2024 and $300 million in Q1 2025 alone — at a time when revenue is declining. The goodwill balance of $1.59 billion, including $748.9 million in accumulated impairment losses in the AMG segment, also creates balance sheet risk if future acquisitions underperform. Unlike fabless competitors who rely entirely on external foundries, ON Semiconductor operates 19 manufacturing sites in 9 countries, including internal fabrication for power semiconductors in South Korea ($1.42 billion in net PPE), the United States ($1.41 billion), and the Czech Republic ($612 million). The Fab Right strategy optimizes this footprint by internalizing high-margin, differentiated products while using external foundries for commoditized or peak-demand products, creating a capital-efficient model that generated $1.21 billion in free cash flow in FY2024 on $694 million in capex — compared to $438 million in free cash flow on $1.54 billion in capex in 2022. CEO Hassane El-Khoury led Cypress Semiconductor from a struggling commodity memory company to a focused automotive and IoT semiconductor leader that commanded a $9.3 billion acquisition price from Infineon. ON Semiconductor is investing up to $2 billion in a brownfield SiC semiconductor facility in the Czech Republic, which would establish a Central European supply chain to service European automotive OEMs' rapidly increasing demand for intelligent power semiconductors. The January 2025 acquisition of Qorvo's SiC JFET technology for $118.8 million specifically targets high energy efficiency and power density in the AC-DC stage of AI server power supply units. The October 2025 acquisition of Aura Semiconductor's Vcore power technologies for up to $144 million further expands the power management portfolio for data center applications. The 2024 business realignment affected approximately 1,600 employees and incurred $133.9 million in restructuring charges, but management believes these actions will improve long-term profitability. The company has committed to returning 50% of free cash flow to shareholders through share repurchases, and it has a $3 billion share repurchase authorization in place. In FY2024, the company returned 54% of free cash flow through $650 million in buybacks. In Q1 2025, it returned 66% of free cash flow through buybacks. The 2027 targets — 10-12% revenue CAGR, 53% gross margin, 40% operating margin, $3.5-4.0 billion in free cash flow — are ambitious but built on specific initiatives with measurable milestones. The January 2025 acquisition of Qorvo's SiC JFET technology for $118.8 million specifically targets the AC-DC stage in power supply units for AI data centers, complementing the company's existing EliteSiC portfolio. This capital efficiency directly flows to free cash flow, which management targets at 25-30% of revenue ($3.5-4.0 billion in 2027) compared to 17.1% in FY2024. Industry analysis suggests ON Semiconductor's SiC products carry EBITDA margins exceeding 40%, well above the company average. The company is investing up to $2 billion in a Czech Republic SiC facility to capture this opportunity, with production expected to ramp in the 2026-2027 timeframe. The first quarter 2025 revenue of $1.45 billion, down 22.4% year-over-year, suggests that the downturn is not yet over. ON Semiconductor paid $2.4 billion in cash — approximately $20 per share — to acquire Fairchild, a pioneer in power semiconductors that had been founded in 1957 and had invented the planar transistor and the integrated circuit. The acquisition created a top-10 non-memory semiconductor supplier with almost $5 billion in pro forma revenue and a comprehensive power management portfolio. Management projected $160 million in annual cost savings by the end of 2017, $200 million by 2018, and $225 million by 2019. At Cypress, he had transformed the company from a struggling commodity memory supplier into a focused automotive and IoT semiconductor leader, culminating in its $9.3 billion acquisition by Infineon in April 2020. Revenue grew from $5.26 billion in 2020 to a peak of $8.33 billion in 2022, a 58% increase in two years. Free cash flow surged from negative territory to $1.21 billion in 2024. And the company's market capitalization grew from approximately $8 billion at the start of El-Khoury's tenure to $22.5 billion by April 2025. But the company's structural improvements — Fab Right manufacturing, portfolio rationalization, and capital discipline — allowed it to generate $1.21 billion in free cash flow despite the revenue decline, a performance that would have been impossible in the pre-El-Khoury era.