NVIDIA CEO: HBM supply shortage to impact AI chip growth

NVIDIA's CEO has indicated that a shortage in High Bandwidth Memory (HBM) supply is expected to impact the growth of AI chip production. This means that the ability to manufacture and deliver AI chips, which are crucial for various advanced computing applications, may be limited not by demand, but by the availability of a key component.

This matters because HBM is essential for high-performance AI accelerators, including GPUs, enabling faster data processing critical for AI models. A constrained supply of HBM could slow down the expansion of AI infrastructure and data centers, potentially delaying AI advancements across industries. The bottleneck is in the memory component itself.

The mechanism involves the intricate supply chain for semiconductors. HBM is a specialized type of RAM that is stacked vertically and integrated directly with the processor, offering significantly higher bandwidth than traditional memory. Manufacturing HBM is complex and involves specialized processes, making its production capacity a critical choke point for advanced AI chip output.

This development primarily impacts NVIDIA (NVDA), as a leading AI chip designer, potentially limiting its revenue growth from AI GPUs. Other companies involved in AI hardware, such as AMD (AMD) and Intel (INTC), could also face similar HBM supply constraints. Data center operators and cloud service providers like Microsoft (MSFT), Amazon (AMZN), and Alphabet (GOOGL), who are building out AI infrastructure, may experience slower deployment of new AI capabilities due to reduced chip availability.