Pay Close Attention to This Crucial Revenue Source for Artificial Intelligence (AI) Giant Nvidia

After years of high-flying success, Nvidia (NASDAQ: NVDA) has hit some turbulence in 2025. The company’s stock has declined around 20% year to date, losing nearly $1 trillion in market cap. It’s also underperforming the major stock market indexes this year.

So where is Nvidia stock headed next, and what key metrics should investors keep a close eye on? Let’s dig in and find out.

Where to invest $1,000 right now? Our analyst team just revealed what they believe are the 10 best stocks to buy right now. Continue »

Image source: Getty Images.

Over the last three years, Nvidia’s revenue has gone through the roof. In its fiscal 2022, the company generated $26.9 billion in revenue. But its trailing-12-month revenue reached $130.5 billion — a fivefold increase from just three years ago.

The engine behind this growth is Nvidia’s data center unit, which generated $115.2 billion, or 88%, of the company’s overall revenue in fiscal 2025.

The data center unit designs and sells graphics processing units (GPUs), which are the workhorses used in all the critical stages of artificial intelligence (AI) development: training, inference, and deployment.

As AI models like ChatGPT have taken the world by storm, Nvidia has become the world’s leading provider of data center GPUs. In fact, according to Motley Fool research, Nvidia now generates about three times as much data center revenue as its three closest competitors — Intel, Advanced Micro Devices, and IBM — combined.

While Nvidia is miles ahead of the competition now, can its lead endure?

Two major worries are hanging over Nvidia right now. One is specific to the AI industry; the other relates to trade policy.

First is whether the red-hot growth of AI models and data centers will cool off. There is major disagreement on this point. Some analysts point to the emergence of DeepSeek AI as a sign that AI models will become more efficient, thus requiring less computing power and, therefore, fewer GPUs to operate. Still, other analysts believe that more efficient AI models will actually increase the need for GPUs as the barriers to building new models decrease.

Whichever side you might be on, one thing is clear: The argument here is whether the growth rate in the GPU market slows, not whether the market shrinks. There is a wide consensus that the overall GPU market will continue growing for many years to come.