The “Four AI Waves” Driving NVIDIA's Explosive Growth (the 2nd One Will Change Everything)

AI is ushering in a monumental technological revolution—and NVIDIA is powering it.

The company dominates the market for AI chips, with nearly 80% market share. And that fact is showing in its stock price:

The company's stock rose 17% after it announced blockbuster earnings. The company's stock has risen 63% in 2024 alone. And NVIDIA is now worth nearly $2 trillion.

(In fact, if you put $10,000 into the stock a decade ago, you'd have $1.7 million today.)

Why is the Street salivating over the company? What does it mean for the future of AI?

I got the answers on Episode 85 of The Artificial Intelligence Show from Marketing AI Institute founder/CEO Paul Roetzer.

NVIDIA is just getting started

NVIDIA produces the essential hardware needed to power modern AI applications. Specifically, the company produces special chips called graphics processing units (GPUs). GPUs were originally designed for graphics-intensive tasks like video games. But it turned out they were really good at powering AI models, too.

Historically, that growth has been driven by selling GPUs for training AI models, says Roetzer. Everyone (OpenAI, Google, Microsoft, etc.) is training large models that require a massive amount of compute from NVIDIA's GPUs.

"But the real growth in the future is what's called inference," says Roetzer. Once the training is done, we all use the model. (ChatGPT, Gemini, etc.)

The moment when it goes and computes information to output something is called inference. You need GPUs for inference, too.

"So once enterprises actually start using these generative AI tools into their daily workflows, there's a whole new level of growth that happens and a whole new level of need for these GPUs, for these processing units," says Roetzer.

NVIDIA CEO Jensen Huang conceptualizes this growth in four waves: 

1. Wave One: The initial training and infrastructure build-out (This is where we are mostly at now).
2. Wave Two: Widespread enterprise adoption driven by AI agents.
3. Wave Three: AI usage in heavy industries (manufacturing, energy, etc.). 
4. Wave Four: Sovereign AI, where every government relies on AI.

Now, we're possibly at the very, very beginning of Wave Two, says Roetzer. At best, we're in the very early phases of enterprise adoption.

And the most important thing to understand in the near-term is something Huang mentioned in Wave Two:

The rise of AI agents.

AI agents are about to change everything

AI agents are systems that can learn and act on their own. And they can be embedded into existing software platforms that companies already use.

In a call after the company's stellar earnings, Huang said:

“The world's enterprise software platforms represent approximately a trillion dollars. . . . These application-oriented, tools-oriented platforms and data-oriented platforms are all going to be revolutionized with these AI agents that sit on top of it. And the way to think about that is very simple. Whereas these platforms used to be tools that experts would learn to use, in the future, these tools companies will also offer AI agents that you can hire to help you use these tools or to help you reduce the barrier of using these tools.”

This is key to understanding where AI is headed, says Roetzer:

"You have to understand that they all believe AI agents are going to be key. Those agents are going to need a lot of computing power, which is going to need a lot more chips and a lot more data centers. And that's basically the next five to 10 years of business.”

Agencies and VARs need to evolve ASAP

This is going to have some profound implications pretty quickly for any company that resells software or provides services to support it. (Think: marketing agencies and VARs.)

That’s because AI agents have the potential to automate the tasks we all spend lots of time doing within software platforms. For example, an AI agent could observe a worker, learn a simple task that takes dozens of clicks, then go do it on its own. (Rinse and repeat for any other software task in a business.)

In other words: AI agents will be able to use your software for you.

Which means no more need to consult trusted partners on how to get value out of the software. Or hiring them to do things using the software for you.

"All of a sudden, I can go into HubSpot or Salesforce or whoever has VAR networks and agency ecosystems. And I can hire an agency—or I can turn on an AI agent,” says Roetzer.

Agencies and VARs must be thinking about the impact of having AI agents that can perform tasks in software at almost no cost.

"Realistically, the cost versus what it would cost to traditionally hire humans to do this work is going to be next to zero,” he says.

That includes thinking about how to reinvent their services and value creation in radical ways. For instance, agencies may want to offer services related to training these agents or configuring them for clients, rather than trying to compete with machines to do easily automated tasks.

And every knowledge worker needs to evolve, too

And it's not just agencies and VARs who need to act now.

"Anyone who works in these platforms, even on the brand side, you're using tools that an AI agent is going to be trained to use," says Roetzer.

That makes it critical to conduct “AI impact assessments,” says Roetzer. These are assessments that look at your job responsibilities, company processes, and software to determine what is vulnerable to AI automation.

The time to start doing this is now.

The mainstream adoption of AI agents may be years away. But once it happens, the change will be far larger and faster than most businesses anticipate.

“When we have true AI agents, it reinvents knowledge work,” says Roetzer.

2024 is likely to see some AI agent breakthroughs, followed by increasingly widespread adoption through 2026, says Roetzer. AI agents may even be commonplace as soon as 2027.

The good news?

"There's time to plan for this," says Roetzer. "But this may be way more disruptive than generative AI.”