I have had the privilege of knowing Nvidia's CEO, Jensen Huang, for decades. I met him before he started Nvidia and was one of the few analysts he briefed when he launched the company.

Even then, he was always the smartest person in the room. I marveled at his vision for Nvidia and have watched him become one of the most important tech CEOs in my lifetime. One thing I learned over the years is that when he speaks, especially about the future, one should listen closely.

About 15 years ago, he brought together a small group of analysts to share an ambitious vision for the company. At that time, he believed the future would require a new kind of computing infrastructure. It would go beyond GPUs, which were the core of his business at the time. He envisioned systems ranging from supercomputers to specialized PCs, such as high-powered gaming machines, as well as new forms of personal computing.

From Early Vision to “Owning the Stack”

This meeting stood out to me because it was the first time I heard the term “owning the stack.” At the time, it was a new idea in tech—one that meant controlling the entire computing stack, from hardware and infrastructure to software and applications, rather than relying on separate players at each layer.

Today, it has become a central principle across industries and within the investment community.

Back then, Jensen did not mention artificial intelligence in this meeting, but in hindsight, he clearly saw our AI future and its importance to them and the world.

If you want to understand why AI represents the most significant technology transition of our lifetime, start with a simple premise: AI is fundamentally about thinking.

Real thinking — the kind that understands a question, grasps its context and goes out to gather information before formulating a response. This mode requires serious computational processing. That's not a minor implementation detail. It's the architectural foundation of an entirely new industry.

Tech leaders call this AI infrastructure, and it is important to define what that means. This is not a typical product cycle or a simple software upgrade. What is being built today belongs in the same category as the electrical grid and the internet. It is permanent, foundational infrastructure that societies around the world will rely on. Viewed this way, the scale of the transformation is difficult to overstate.

What makes this moment particularly compelling for industry observers is the simultaneous transformation occurring across the entire technology stack. Each layer reflects a distinct but interconnected dimension of this shift.

A Full-Stack Boom: Where Value Is Created

Start with energy. The power demands of AI are already reshaping utility planning and data center investment in ways we haven’t seen since the dot-com build-out, only bigger. Move up to silicon and you see companies like Nvidia posting results that would have seemed implausible three years ago, with no meaningful slowdown in sight. The chip industry isn’t just growing , it's being reorganized around AI workloads as its primary design target.

Above that sits the infrastructure layer of the hyperscalers and cloud platforms. Azure, AWS, Oracle Cloud Infrastructure and CoreWeave are all expanding capacity at a pace that strains the imagination. These aren't speculative bets. They are responses to demand that is already there and accelerating.

Then come the model developers: OpenAI, Anthropic, xAI and others. Each of these are racing to push the frontier of what AI can reason through and accomplish. And finally, at the application layer, genuinely new categories of software emerge. Cursor is redefining how developers write code. Open Evidence is changing how clinicians access medical knowledge. These aren’t AI-enhanced versions of existing apps. They are new species entirely.

What stands out is how many layers of the stack are advancing at the same time, and how quickly. Past platform shifts were more concentrated. The PC era mainly expanded the application layer. The internet reshaped networking and distribution. Smartphones introduced a new interface. AI is different. It is driving change across multiple layers at once—how intelligence is generated, how compute is delivered, and how software is built.

This is what sets this investment thesis apart from past cycles. Investing across the full stack is not just a diversification strategy; it reflects how value is created in a build-out of this scale. The gains do not concentrate in one segment. They spread across the ecosystem. Energy providers benefit. Chip designers benefit. Cloud platforms, model developers and application builders all benefit. In this case, the “rising tide lifts all boats” idea is unusually accurate.

For those who’ve watched the industry long enough to have seen the electrical grid get built out, the interstate highway system expand and the internet go from a government research project to the backbone of the global economy, there’s a familiar feeling to this moment. But this has an urgency and a compression of timelines that feels genuinely new.

The way computing is delivered is changing. The way intelligence is produced and distributed is changing. The infrastructure now being built—across data centers from Virginia to Singapore, fabs in Arizona and Taiwan and research labs from San Francisco to London—will endure for generations.

This is the foundation of what comes next. By any historical measure, it represents one of the most consequential inflection points the technology industry has ever faced.

Disclosure: Nvidia subscribes to the research reports from the company I founded, Creative Strategies, along with many other high-tech companies around the world.