Hyperscalers’ shift toward power-intensive AI inference spending boosts demand for ON Semiconductor’s power semiconductors.
AI-related spending is pivoting from data center infrastructure to inference, which will soon dominate hyperscaler budgets. Inference, the process of running AI models for real-world applications, demands significant power and thermal management, benefiting power semiconductor firms like ON Semiconductor (NASDAQ: ON).
Current spending focuses on building data centers and training models, but inference costs are expected to surpass infrastructure expenditures in the coming years. ON Semiconductor, a partner of Nvidia, stands to gain from both segments but is particularly well-positioned for the long-term growth in inference.
The shift reflects broader trends in AI adoption, where operational costs will increasingly outweigh initial capital investments. ON Semiconductor’s role in power efficiency and thermal solutions aligns with this transition.