Broadcom and Nvidia Just Delivered Blowout Earnings. Which Is the Better Growth Stock to Buy in March?

Broadcom (NASDAQ: AVGO) and Nvidia (NASDAQ: NVDA) make up a combined 9.7% of the S&P 500 and 17.9% of the Nasdaq-100. So when they report earnings, the market listens.

Here's what you need to know about where Broadcom and Nvidia are after their latest quarterly earnings releases and where they could be headed to help you decide which growth stock is the better buy for you.

Image source: Getty Images.

Broadcom's strategic partnerships with hyperscalers provide a clear runway for sustained growth

In just a few years, Broadcom has transformed from a somewhat stodgy hardware company to an infrastructure software and artificial intelligence (AI) powerhouse.

Broadcom makes custom AI accelerators for training and inference, such as Tensor Processing Units (TPUs) for Alphabet and Anthropic, as well as Meta Platforms' Meta Training and Inference Accelerator (MTIA).

Broadcom is experiencing massive growth in data center solutions. Graphics processing units (GPUs) remain the workhorses of data centers. But Broadcom customizes its XPU for workloads that vary based on the large language model (LLM).

CEO Hock Tan noted on the March 4 earnings call that "the one-size-fits-all with general-purpose GPU gets you only that far." This is why Broadcom believes its custom XPUs will eventually overtake the traditional GPU design in the data centers being built going forward.

Importantly, Broadcom designs different XPUs for different functions. Instead of having GPUs do all the work, Broadcom makes XPUs for training (teaching AI) and separate XPUs for AI inference (applying what the AI has been taught). Inference will be critical in the widespread adoption of agentic AI.

Broadcom also makes a variety of network and connectivity hardware tailored to maximize bandwidth for AI clusters, such as its Tomahawk switches and Jericho routers. Broadcom estimates that AI networking components will account for 33% to 40% of total AI revenue in any given quarter.

At the midpoint, that would mean AI networking contributed about $3.1 billion, or 16.1%, to Broadcom's first-quarter fiscal 2026 revenue; AI chips contributed about $5.3 billion (27.5%); non-AI semiconductor revenue about $4.1 billion (21.2%); and infrastructure software $6.8 billion, or 35.2% of total revenue.

In addition to being a highly diversified business, Broadcom consistently returns substantial capital to shareholders -- including $3.1 billion in dividends and $7.8 billion in stock buybacks in its latest quarter. Broadcom has boosted its dividend for 15 consecutive years. Many of the dividend raises have been massive, resulting in a mind-numbing 13-fold increase in the dividend in the last decade.

Nvidia remains the data center workhorse

Like Broadcom, Nvidia recognizes the massive opportunity in AI inference and the need to evolve beyond the basic catch-all GPU model. Its product upgrades follow its AI roadmap, with Blackwell Ultra and the newest platform -- Rubin -- delivering strong efficiency improvements and lower costs compared to the Hopper platform that debuted just four years ago.

Nvidia's Rubin platform consists of six different chips, only one of which is a GPU. By codesigning the GPU alongside a CPU and networking hardware, the Rubin platform can reportedly reduce inference costs by 90%. Its NVL72 solution includes the entire chip bundle, offering a plug-and-play supercomputer for megascale data centers.

So, like Broadcom, Nvidia is capitalizing on demand for high-efficiency inference by integrating networking hardware and software directly into its rack-scale offering. However, Nvidia is significantly less diversified than Broadcom.

Both companies depend on a handful of hyperscalers to drive the bulk of their data center revenue. But the key difference is that the data center made up 91.5% of Nvidia's revenue in its latest quarter, while Broadcom's non-AI semiconductor and infrastructure business is still larger than its AI chips and networking business.

Two top growth stocks to buy now

Broadcom and Nvidia have what it takes to be foundational AI growth stocks for long-term investors. Both semiconductor stocks have produced monster gains in recent years, but remain reasonably valued because they are growing rapidly while maintaining high profit margins.

Hyperscalers are turning to Broadcom for its custom-built chip solutions, including design, process technology, packaging, and networking, tailored to their specific workloads, while Nvidia's rack-scale Rubin platform is ideal for quickly building and deploying AI systems.

Ultimately, the pie is large enough for both solutions to succeed. So some investors may want to simply buy a 50/50 split of Broadcom and Nvidia rather than picking one stock over the other.

However, investors looking for diversification beyond AI and a stable, growing dividend may prefer Broadcom, while Nvidia could be the better choice for those seeking a more concentrated bet on growing demand for data center compute.

Daniel Foelber has positions in Nvidia. The Motley Fool has positions in and recommends Alphabet, Meta Platforms, and Nvidia. The Motley Fool recommends Broadcom. The Motley Fool has a disclosure policy.