ON Semiconductor (ON) Poised for Growth as AI Spending Shifts to Inference

ON Semiconductor (ON) is positioned as a key beneficiary as artificial intelligence (AI) spending transitions from data center infrastructure to "inference" computing in the coming years. Analysts believe the market underestimates ON's long-term potential as inference, running AI models for real-world applications, is projected to dominate AI spending. Inference computing, distinct from infrastructure build-out for model training, is highly power-intensive and requires robust thermal management, driving demand for power semiconductors. ON, a partner with NVIDIA (NVDA) for power technology in next-generation data centers, reported its data center business grew 30% year-over-year in Q1 2026. CEO Hassane El-Khoury projects ON's data center revenue to double to approximately $500 million by 2026, from an estimated $250 million in 2025. Though a smaller portion of the projected $6.47 billion total revenue for 2026, this rapid growth highlights ON's leverage to the expanding inference market, including edge applications in EVs and industrial automation.