Wall Street analysts are increasingly bullish on Broadcom's custom chip business, but CNBC's Jim Cramer argues the semiconductor giant isn't really competing head-to-head with Nvidia's AI dominance. The commentary comes as both companies navigate the booming demand for AI infrastructure, with investors trying to figure out which players will capture the biggest slice of the market.
The semiconductor world's latest debate centers on whether Broadcom is positioned to challenge Nvidia's stranglehold on AI computing, and CNBC's Jim Cramer has a clear answer: they're not even in the same race.
Cramer's comments come as Wall Street analysts pile onto Broadcom with increasingly optimistic coverage, viewing the company's custom silicon business as a major growth driver. But according to Nvidia CEO Jensen Huang, Broadcom's custom chips pose no threat to his company's dominance in AI training and inference.
The distinction matters more than it might seem. While Nvidia builds general-purpose GPUs that can handle virtually any AI workload, Broadcom specializes in custom application-specific integrated circuits (ASICs) designed for specific customers and use cases. Think of it as the difference between a Swiss Army knife and a precision surgical tool.
Broadcom's custom chip business has been quietly booming, particularly with hyperscale customers like Google and Meta who need specialized processors for their unique AI workloads. The company's revenue from custom silicon jumped significantly in recent quarters, though it remains a fraction of Nvidia's $60+ billion annual revenue run rate.
The competitive landscape gets more interesting when you consider that many of Broadcom's biggest customers are also Nvidia's largest buyers. Google purchases billions in Nvidia H100s while simultaneously working with Broadcom on custom TPUs for specific applications.
"These companies serve complementary rather than competing functions," one semiconductor analyst told investors during a recent briefing. The custom chip market addresses specific optimization needs that general-purpose GPUs can't efficiently handle, while Nvidia's platforms remain essential for the broad AI development ecosystem.
Cramer's assessment reflects this nuanced reality. Rather than viewing Broadcom as an Nvidia challenger, he sees it as a beneficiary of the same AI boom that's driving Nvidia's record performance. The total addressable market for AI chips continues expanding, creating room for multiple winners with different specializations.
