Last month during its Investor Day, semiconductor company Onsemi outlined its long-term growth strategy to lead one of the largest and most consequential technology shifts of the next decade by solving the growing power constraints limiting the advancement of AI (artificial intelligence), electrification, and automation. As those markets converge around common technology requirements, the chipmaker says it is leveraging its portfolio across power and sensing to address market opportunities and solve the needs of some of the most important technology trends shaping the global economy.

“Power density is emerging as one of the defining engineering challenges of this decade,” said Hassane El-Khoury, President and CEO of Onsemi. “The companies that lead in the next era of innovation will be those that can deliver more energy while consuming less space. As that challenge becomes more urgent, power has become a foundational technology layer that will help shape the pace of innovation, economic growth, and global competitiveness.”

At the event, El-Khoury discussed charting a path to power the next decade of innovation, and was joined by Achyut Shah, Group President, Power Solutions Group, on powering the AI revolution; Sudhir Gopalswamy, Group President, Analog & Mixed-Signal and Intelligent Sensing Groups, on expanding the core of automotive and industrial; and Thad Trent, Executive Vice President and CFO, on turning technology leadership into durable financial performance.

Onsemi sees a $213 billion TAM (total addressable market) by 2030 spanning automotive, industrial, AI data centers, and emerging applications. Although these markets are often viewed separately, they increasingly rely on the same underlying capabilities of efficient power conversion, perception of the physical world, and edge intelligence and connectivity.

This convergence allows the company to deploy common technology platforms across multiple end markets. For example, a high-voltage technology developed for automotive electrification can also support industrial energy systems and AI data centers. Advances in sensing and connectivity can extend across vehicles, factories, robotics, medical systems, and other autonomous applications.

Execs expect the company’s platform-based model to increase the revenue potential of its core technologies by moving from individual components to higher-value solutions and system architectures that address a larger share of each customer’s design requirements. Through this approach, they aim to generate greater leverage from R&D investments, reduce dependence on any single end market or technology cycle, and grow faster than the markets served.

 

Solving power density and AI at the edge

Capturing the opportunities requires addressing a growing constraint facing AI infrastructure and other intelligent systems—power density, or the ability to deliver greater performance and capability using less energy and less physical space. Next-generation systems must use available energy more productively while generating less heat, occupying less space, and maintaining reliability at scale.

Increasing power density requires electrical, thermal, mechanical, and physical performance to be optimized simultaneously. Increasing power typically generates more heat, while reducing system size can make that heat more difficult to remove.

Improving efficiency may require new materials, device architectures, and integration methods. Historically, these trade-offs forced system designers to make compromises among power, size, cost, and thermal performance.

As these challenges become more complex, value increasingly shifts to companies that can optimize across the entire system.

Onsemi has expanded beyond traditional power semiconductors to build capabilities spanning technologies, products, solutions, and systems. Company execs highlighted three technology platforms that will shape the next generation of intelligent systems and drive long-term growth across its key end markets: high-voltage technologies for the device layer; Treo for the intelligence layer; and EPP (embedded power platform) for the integration layer.

The company unveiled the EPP at the Investor event, calling it a breakthrough architecture for the AI era aimed at redefining system power delivery through unprecedented power density and integration of multiple dies into a single silicon device.

While today’s AI investments are largely focused on building the infrastructure needed to train and run increasingly sophisticated models, company execs believe that the next phase of growth will occur at the edge, where intelligence moves beyond the data center into machines that interact directly with the physical world. Whether deployed in a robot, vehicle, factory, medical system, or autonomous platform, each physical AI system requires four capabilities to operate together: power, to provide the energy foundation that enables intelligent systems to operate; sense, to perceive and understand the physical environment; control, to execute decisions through precise real-world actions; and connected compute, to process, share, and continuously adapt information at the edge.

The execs say that the ability to combine these capabilities through system-level co-design and optimize performance across the full stack positions Onsemi to deliver differentiated solutions and capture a larger share of the value created by physical AI.

“The ability to efficiently deliver power, sense the environment, and enable intelligent decision-making will define the next generation of intelligent machines,” said El-Khoury. “Our differentiated technology platforms enable us to solve power density across the device, intelligence, and integration layers in a way no other semiconductor company can, positioning us to expand our content and grow as AI moves from infrastructure into the physical world.”

 

Embedded Power Platform

The EPP architecture uses the silicon wafer as the foundation of the package and introduces a highly integrated approach to power system design. Designed as a scalable platform, it brings electrical, mechanical, and thermal design together to help customers achieve higher power density, improve system performance, and accelerate development.

“For decades, the semiconductor and the package have been treated as separate technologies,” said El-Khoury. “EPP changes that by making the silicon itself part of the system architecture. EPP brings together advanced semiconductor technologies, manufacturing, and system-level optimization into a common architecture that can evolve alongside future innovations. This approach can redefine how power systems are built and create a new foundation for AI infrastructure, electrification, and automation.”

According to Onsemi, AI infrastructure, electrified transportation, and industrial automation are all competing for the same critical resource—power. Customers need to move and manage more electricity within increasingly compact systems while controlling heat, efficiency, cost, and development time. Yet many of today’s power systems are still developed using traditional design approaches that treat power electronics, mechanical design, and thermal design as separate engineering challenges, with each layer optimized independently and sequentially. Decisions made at one stage can create compromises in another, leading to additional engineering iterations, costly late-stage changes, and longer development cycles.

EPP reimagines the package from passive housing into an active contributor to system performance. By using the silicon wafer as the package, it enables the integration and interconnection of silicon, SiC (silicon carbide), and GaN (gallium nitride) technologies within a highly integrated wafer-level architecture.

Multiple devices, including FETs (field-effect transistors), drivers, and controllers, can be embedded in a single package and co-optimized for electrical, thermal, and mechanical performance. This enables complete power-system co-design, allowing electrical, thermal, and mechanical characteristics to be evaluated and optimized together from the start. The result is higher power density, improved system performance, reduced development complexity, and faster time-to-market.

EPP also leverages Onsemi’s standard 12-in silicon wafer manufacturing capabilities, bringing key integration processes into the precision and control of the semiconductor fab. This applies mature semiconductor design tools, wafer-level manufacturing, and advanced simulation capabilities to power-system integration, helping improve performance while accelerating innovation.

The technology replaces a sequential model with a common platform that can be co-designed, co-simulated, and co-optimized. The approach is designed to help customers achieve three to five times higher power density, depending on the application, and accelerate development cycles to as little as four months. It can improve thermal performance and heat dissipation, reduce electrical losses through lower parasitic inductance, and enable greater device control and higher switching frequencies.

Crucially, it can identify design trade-offs earlier and reduce costly late-stage changes, and it can scale a common architecture across power levels, device types, applications, and semiconductor technologies.

Onsemi says EPP can help it get more of the $47 billion electrification market. For EV applications, traction inverters are often constrained by efficiency losses, thermal limitations, development complexity, and system size. EPP addresses these challenges with up to four times higher power density and 15% lower power losses compared to conventional approaches, enabling smaller, lighter, and more efficient inverter designs. Its scalable architecture supports a single inverter platform spanning low-end to high-end vehicle applications, allowing automakers to reuse a common design across multiple vehicle models and power classes.

This approach can reduce R&D and manufacturing costs, accelerate qualification and development cycles, improve vehicle range or lower system costs, and help bring new vehicle programs to market faster.

EPP is expected to begin sampling in 2026 with strategic customers and ecosystem participants across automotive and AI applications.

 

Subaru to evaluate EPP

At the event, Onsemi announced a strategic technology engagement with Subaru Corp. to evaluate EPP, giving the automaker early access to the technology and engineering expertise as the company explores next-generation power architectures for future electrified vehicles.

Building on a longstanding relationship between Subaru and Onsemi, the initial phase of the new collaboration focuses on early engineering evaluation and technical learning as the companies explore advances in power-semiconductor integration that can translate into scalable system-level solutions to meet the performance, efficiency, and thermal demands of next-generation mobility.

The two companies also collaborate on advanced safety technology, with Onsemi providing advanced CMOS (complementary metal-oxide semiconductor) image sensors for Subaru’s EyeSight driver-assist stereo cameras, and will supply Hyperlux image sensors, such as the AR0823AT, for Subaru’s next-generation AI-powered EyeSight slated for upcoming EVs.

Power electronics are playing an increasingly important role in vehicle performance, efficiency, and design flexibility as automakers expand their electrified vehicle portfolios. The engagement enables Subaru to explore how advances in power semiconductor integration can support broader system-level optimization, including new approaches to power system architecture, efficiency, and vehicle design flexibility—while providing Onsemi with practical customer insights to support the continued development of EPP.

”Our early engagement with Onsemi gives us an opportunity to evaluate EPP’s integrated approach to power system design while drawing on Onsemi’s modeling, simulation, and engineering expertise,” said Tamotsu Inui, Managing Executive Officer and Chief General Manager of the Engineering Division, Chief General Manager of Technical Research Center, at Subaru. “We are encouraged by the platform’s potential and look forward to continuing our technical engagement as we develop future electrified-vehicle architectures.”

“The transition to electrification is driving increasingly sophisticated vehicle architectures and placing growing pressure on automakers to improve efficiency, simplify development, and accelerate innovation,” added Dinesh Ramanathan, Senior Vice President of Corporate Strategy at Onsemi. “That is creating an opportunity to rethink power systems from the ground up. Subaru’s engagement brings valuable customer insight to EPP’s continued evolution, using the platform in advanced customer systems and exploring how its integrated power architecture would support the next generation of electrified vehicles.”

For future electrified vehicles, EPP can support smaller, lighter, and more efficient traction inverter systems while reducing development complexity and accelerating time to market. Because the platform can scale across multiple power levels using a common architecture, automakers can reduce engineering effort, minimize requalification costs, and accelerate development cycles. The result is expected to be a faster path from concept to production, with lower development costs and greater flexibility across future vehicle programs.