Is Broadcom Ready to Compete with Nvidia's AI Market Dominance?

Nvidia (NVDA) has been a leader in the field of artificial intelligence (AI) computing hardware throughout the ongoing AI race. However, a competitor is making strides to change that narrative: Broadcom (AVGO). While Nvidia is known for its top-tier graphics processing units (GPUs), Broadcom is introducing an alternative approach with its custom AI accelerators, referred to as XPUs.

But the question remains: Can these XPUs truly rival Nvidia’s supremacy in the AI investment landscape? Or is this simply hype from Broadcom’s management?

Understanding Broadcom's XPUs and Their Potential

GPUs have gained their reputation due to their ability to perform numerous calculations simultaneously. This parallel processing makes them ideal for complex tasks like AI training and inference. Moreover, multiple GPUs can be linked to amplify their processing capabilities, allowing organizations to scale to data centers with over 100,000 GPUs. Such training power is significant, as GPUs are built to be adaptable for a range of applications.

On the other hand, XPUs are uniquely designed to handle specific types of workloads efficiently. While this specialization may limit their utility in other computing tasks, it provides a distinct advantage for AI processes tailored for their configuration. By focusing on essential capabilities, XPUs can potentially outperform GPUs when optimized for particular tasks, which is the driving force behind Broadcom’s optimistic growth projections.

In 2024, Broadcom anticipates generating $12.2 billion from its AI-related products. The company believes that the market for XPUs and supplementary AI hardware could reach between $60 billion and $90 billion by 2027, primarily fueled by three key customers.

This trend is on the rise, as four additional clients are currently developing their XPU designs. Such growth could significantly enhance Broadcom’s already ambitious market forecast, making it a stock that warrants attention from investors.

But does this mean trouble is brewing for Nvidia?

Nvidia's Position and Future Outlook

During a recent conference call addressing Broadcom's Q1 FY 2025 (ending February 2), an analyst raised the pertinent question of whether the burgeoning inference-heavy market would favor GPUs or XPUs. Broadcom’s CEO, Hock Tan, noted that the anticipated market growth includes a substantial role for GPUs and doesn’t indicate a complete segregation between the two types of hardware.

This insight is crucial, as GPUs will continue to play a vital role in AI training and inference. Not every application can be efficiently designed for XPU optimization, which means Nvidia’s GPUs will remain integral to AI budgets from major cloud service providers.

Additionally, a significant portion of GPU utilization occurs in data centers leased to clients in the cloud computing sector. Many of these clients are familiar with optimizing GPU workload, making the transition to XPUs a challenging endeavor.

Overall, it is likely that Nvidia will continue to experience growth in the upcoming years, although it can be expected that Broadcom will gradually encroach upon Nvidia's market share as XPUs gain popularity.

Both Nvidia and Broadcom present appealing investment opportunities at their current valuations. With each company forecasting significant growth moving forward, a forward price-to-earnings ratio of under 30 appears to be an attractive proposition, especially when considering historical trading patterns.

Investor sentiment towards Nvidia seems to have adapted to some degree of pessimism, whereas optimism for Broadcom's potential has yet to materialize fully. As a result, both stocks may represent excellent investment options, and investors could consider acquiring shares while they are still attractively priced.

Note: The author has a vested interest in Nvidia. The Motley Fool has investment positions in Nvidia and recommends both Nvidia and Broadcom.

Nvidia, Broadcom, AI