Nvidia (NVDA -0.84%) and Broadcom (AVGO -1.00%) are the leading players in the artificial intelligence (AI) semiconductor space, with both chip designers being the dominant forces in their respective niches.
Nvidia controls the market for data center graphics processing units (GPUs), which has been thriving thanks to the need for massive computing power required to train AI models. Broadcom, on the other hand, commands the lion's share of the custom AI processor market, which has been gaining traction because of AI inference.
However, both stocks have witnessed divergent fortunes on the market in the past year, with Broadcom's gains of 45% significantly outpacing Nvidia's jump of 9%. Does this make Broadcom the better AI stock of the two? Let's find out.
Image source: Getty Images.
The case for Broadcom
Broadcom's custom AI processors are in great demand right now as they are being deployed in huge numbers by cloud hyperscalers for AI inferencing (where a trained AI model draws conclusions based on new data). The company has three customers that are expected to deploy more than 1 million units each of its custom AI processors in their cloud clusters in 2027.
That would be a big jump over the 30,000 units of its custom AI processors deployed in each cloud cluster by its customers last year.
The reason its custom AI processors are in such great demand is because of their ability to perform the tasks they are designed for more efficiently when compared to GPUs, thereby lowering the operating cost. Reports suggest that the company's custom AI chips can be two to three times faster than GPUs, and they consume 30% less power.
This is the reason hyperscalers such as Meta Platforms, Microsoft, Alphabet's Google, Amazon, and others have been building custom AI chips for their servers. And the number of customers using Broadcom's AI chips is likely to go up to seven from the current three. The chipmaker reportedly designs chips for Alphabet, Meta, and ByteDance at present, and there are indications that it could add the likes of Oracle, xAI, and SoftBank to its list as well.
This potential expansion of the company's AI customer base could help it sustain its impressive 70% share of the custom AI processor market. This could pave the way for outstanding revenue growth in the long run considering that the custom AI chip market is expected to clock an annual growth rate of 32% through 2030.
Not surprisingly, analysts have been ramping up Broadcom's revenue estimates for the next couple of fiscal years.
AVGO Revenue Estimates for Current Fiscal Year data by YCharts.
So, Broadcom is likely to remain a top AI stock over the long run as it is building up a solid customer base and is on its way to making the most of the lucrative opportunity in custom AI chips thanks to its outstanding market share. But will that be enough to help its stock deliver more upside than Nvidia?
The case for Nvidia
While the popularity of custom chips sold by Broadcom to tackle AI workloads has been increasing, the demand for GPUs sold by Nvidia continues to remain strong as well. The company's latest quarterly report showed a 69% year-over-year increase in revenue to $44.1 billion. That was faster than the 20% increase in Broadcom's revenue in the previous quarter.
And Nvidia's data center revenue increased 73% year over year thanks to the demand for its Blackwell AI chips. Broadcom's AI revenue growth, on the other hand, was relatively slower at 46%.
So despite its larger revenue base, Nvidia is growing much faster than Broadcom. The reason Nvidia continues to be the dominant player in AI chips is because it has focused on improving the inference performance of its GPUs, which is helping its customers reduce operating costs.
And it remains miles ahead of its rivals in AI revenue, with a 90%-plus share in data center GPUs. This puts the company in a terrific position to sustain its impressive levels of growth for a long time to come.
McKinsey is forecasting an investment of $5.2 trillion worldwide in AI data centers by 2030. That would leave abundant room for growth in the company's data center business from the previous fiscal year's level of $115 billion. This lucrative opportunity suggests why Nvidia is expected to match, or even exceed, Broadcom's projected growth.
NVDA Revenue Estimates for Current Fiscal Year, data by YCharts.
Moreover, Nvidia's AI opportunity isn't just limited to data centers. The company's gaming and AI personal computer (PC) business grew 42% year over year in the previous quarter. Its automotive business recorded a 72% sales increase.
The adoption of AI is projected to increase at a nice clip in both these markets, promising outstanding growth.
The verdict
AI has given Broadcom and Nvidia a lift. But the latter is the bigger beneficiary of the huge demand for AI chips in multiple end markets, while the former is just getting started. Also, as mentioned above, Nvidia's growth is better than Broadcom's even though the former is a much bigger company.
What's more, Nvidia trades at a cheaper valuation than Broadcom even with its stronger growth.
AVGO PE Ratio data by YCharts; PE = price to earnings.
So, even though Broadcom stock may have outperformed Nvidia in the past year, the valuation and growth profile of the two companies suggests that the latter is a better AI stock to buy right now. You can't go wrong with Broadcom, either, considering that it is scratching the surface of a huge opportunity in AI chips, but if you're looking to choose from one of these two AI stocks for your portfolio, Nvidia looks like the better option.
