3 Reasons Broadcom Could Be a Better AI Play Than Nvidia

Nvidia (NVDA 3.59%) is often considered the simplest way to invest in the expanding artificial intelligence (AI) market. It controls over 90% of the market for data center GPUs, which the world’s leading AI companies use to train their AI algorithms. It also locks in those customers with its proprietary software and services, so AI applications optimized for Nvidia’s chips usually need to be rewritten to work on competing GPUs.

From fiscal 2026 (which ended this January) to fiscal 2029, analysts expect Nvidia’s revenue and EPS to grow at CAGRs of 37% and 38%, respectively, as the AI market expands. Those are incredible growth rates for a stock that trades at just 22 times this year’s earnings. However, another emerging AI superpower – the chip and infrastructure software maker Broadcom (AVGO 2.59%) – could actually outperform Nvidia this year for three simple reasons.

Image source: Getty Images.

1. It’s better diversified than Nvidia

Nvidia generated 91% of its revenue from its data center chips in its latest quarter. Broadcom’s business was more diversified: 61% of its revenue in its most recent quarter came from semiconductor solutions, while the remaining 39% came from infrastructure software. Therefore, any concerns about slower AI and data center spending could hurt Nvidia much more than Broadcom, which still sells non-AI chips across a wide range of industries.

Expand

NASDAQ: NVDA

Nvidia

Today’s Change

(-3.59%) $-6.41

Current Price

$172.27

Key Data Points

Market Cap

$4.3T

Day’s Range

$171.84 - $176.52

52wk Range

$86.62 - $212.19

Volume

4.6M

Avg Vol

176M

Gross Margin

71.07%

Dividend Yield

0.02%

2. Its AI accelerators could loosen Nvidia’s grip on the market

Broadcom doesn’t produce GPUs like Nvidia. Instead, it produces custom, application-specific integrated circuits (ASICs) to accelerate AI tasks. Unlike GPUs, which are primarily used for training AI algorithms, custom ASICs can be used for both training and inference.

Hyperscalers (large cloud service providers and AI research firms) are buying many Broadcom custom AI accelerators to handle their inference tasks, dilute data center expenses through economies of scale, and reduce their dependence on Nvidia.

Nvidia is pushing back with its own licensed Groq inference chips. However, Broadcom still expects its AI chip revenue to surge from $20 billion in fiscal 2025 (which ended last November) to $60-$90 billion by the end of fiscal 2027 (39%-58% of its projected revenue).

3. It could grow faster than Nvidia

From fiscal 2025 to fiscal 2028, analysts expect Broadcom’s revenue and EPS to grow at CAGRs of 46% and 56%, respectively, as its AI business expands. Its sales of non-AI chips and infrastructure software should also accelerate again in a warmer macro environment.

At 37 times this year’s earnings, Broadcom’s stock looks a bit pricier than Nvidia’s. But it’s still reasonably valued relative to its growth potential – and it could attract more attention than Nvidia this year as custom AI accelerators steal the spotlight from general-purpose GPUs.