Add as a preferred source on GoogleBank of America Says Slower AI Spending Could Extend Chip Earnings Cycle
The bank’s analysts argue that resilient demand could support chipmakers even as capacity growth slows. They cite $2.5 trillion in long-term contracted backlogs and rising cloud growth.
Bank of America analysts argue that slower growth in artificial intelligence infrastructure spending could extend chipmakers’ earnings cycle by reducing overbuilding while demand for AI services remains resilient.
The analysts say slower capacity growth could improve utilization and pricing for chipmakers. The analysts cite growing cloud activity, heavier use of AI services and contracted backlogs totaling $2.5 trillion as signs of continued demand.
They also expect hyperscalers’ cash burn to peak before narrowing. Projected free cash flow from Nvidia, Broadcom and Micron could help fund infrastructure investment, the analysts say.
The bank’s analysts also flag vendor financing and circularity risks in the spending cycle. The assessment comes amid concerns about the rising cost of AI infrastructure and whether the spending surge can continue.