onsemi Targets $2.5B AI Data-Center Revenue, Sets Ambitious 2030 Margin Goals

onsemi Targets $2.5B AI Data-Center Revenue, Sets Ambitious 2030 Margin Goals AI data-center growth is a central focus: onsemi expects AI data-center revenue to exceed $500 million in 2026, potentially surpass $2.5 billion by 2030, as higher-voltage infrastructure increases its semiconductor content opportunity per rack. The company introduced its Embedded Power Platform and highlighted vertical gallium nitride and silicon carbide technologies to improve power density, efficiency and thermal performance across AI, automotive and industrial applications. onsemi set ambitious 2030 financial targets: 12%–14% annual revenue growth, nearly $11 billion in revenue, 53% gross margins, 38% operating margins and more than $3.5 billion in free cash flow, while planning to return 100% of free cash flow to shareholders after investments and acquisitions. onsemi NASDAQ: ON outlined a strategy centered on power-density technologies for artificial intelligence data centers, electric vehicles, industrial systems and “Physical AI,” while setting long-term revenue, margin and free-cash-flow targets through 2030. At the company’s analyst event, President and CEO Hassane El-Khoury said AI infrastructure is increasingly constrained not only by compute and memory, but by the ability to efficiently deliver and manage power. He said future systems will require power that is distributed, efficient and able to operate under sustained peak loads. “More power is really not enough,” El-Khoury said. “We need better power.” Embedded Power Platform and high-voltage technologies The company introduced its Embedded Power Platform, or EPP, which it described as a platform rather than a single product. EPP uses a silicon wafer as an active package and embeds heterogeneous semiconductor products, including silicon carbide and gallium nitride components, into a wafer-level architecture. onsemi said the approach is intended to improve electrical, thermal and mechanical performance simultaneously while reducing system-development time. The AI Chip Stock Making a Quiet Move Toward Dominance El-Khoury said EPP is already being sampled for automotive, industrial and AI data-center applications using the company’s existing 12-inch manufacturing process. He said the platform can be used with onsemi’s high-voltage technologies and its Treo intelligence platform, which is designed to support modular power-control solutions across voltage levels and end markets. The company also highlighted vertical gallium nitride, or vGaN, technology. El-Khoury said onsemi has developed, sampled and manufactured vertical GaN devices that have been placed on customer boards for automotive, industrial and AI data-center applications. He said production is supported by the company’s Syracuse facility and would not require a new greenfield fab investment. Subaru said it is evaluating onsemi’s embedded power module technology for future vehicle platforms. A Subaru representative said the technology could offer higher power density, dual-sided cooling, improved electrical control and system-level cost reductions, though the automaker said it continues to evaluate how the technology may fit into its vehicle plans. AI data-center opportunity Achyut Shah, president of onsemi’s Power Solutions Group, said AI processors are moving from consuming hundreds of watts toward kilowatt-scale power requirements, while rack-level power needs are rising from tens of kilowatts toward megawatts. He said this transition will require a shift from traditional AC distribution architectures to high-voltage DC systems extending from the grid to computing equipment. Shah said the initial ...