The AI Gold Rush No One’s Talking About

In 1849, thousands of people rushed to California to find gold. Most came up empty.

But the ones who sold the tools, like Levi Strauss with his rugged jeans, built real wealth. It wasn’t the gold diggers who got rich. It was the suppliers of the "picks and shovels".

Today, we’re watching history repeat itself. Artificial intelligence is the new gold. The excitement is real, and so is the opportunity. But the money isn’t just in the apps or the algorithms.

Smart money is moving into the foundation: the computing power, fiber networks, and data centers that make AI possible.

These are the digital picks and shovels of the 21st century.

The Backbone of AI Is Physical, Not Virtual

When you ask a chatbot a question or let your car use self-driving mode, it feels like magic.

But this experience isn’t powered by code alone.

It requires rows of powerful servers, stacks of GPUs, and vast networks of high-speed fiber. AI demands an enormous amount of data movement and processing. And none of it works without the proper infrastructure.

Despite its “invisible” nature, it’s about physics and engineering.

The “cloud” is just someone else’s data center, not up in the sky, as some politicians like Kamala Harris have described.

Every AI function needs fast hardware to keep up.

As these systems become more complex, the pressure grows on the physical layers.

AI companies are hitting the limits of today’s network speeds. Nvidia CEO Jensen Huang has said networking—not chips—is now the bottleneck.

Fast chips are no longer enough. If the networks lag, everything slows down.

AI Doesn’t Work Without a Strong Foundation

Think about what powers AI.

There’s no AI without compute clusters, fiber lines, switches, and reliable cooling.

These components are just as important as the algorithms. Every tool, from smart assistants to image recognition software, is only as good as the hardware behind it.

That’s why tech giants are spending aggressively. Microsoft alone plans to invest $80 billion in infrastructure in fiscal year 2025. This investment covers everything from new data centers to cutting-edge GPUs. They are not doing this for show. It’s necessary to keep AI performance growing.

Amazon, Meta, Google, and Microsoft are now pouring a combined $50 billion per quarter into digital infrastructure. This is the floor, not the ceiling. These companies know that they risk falling behind if they don’t scale up. Infrastructure has become a strategic priority.

The Missed Opportunity Most Are Ignoring

Most investors are focused on AI software. That’s understandable. It’s where the headlines are. But it’s also where the crowd is. The deeper value often sits beneath the surface.

An autonomous car can generate gigabytes of data every second. It needs to transmit and receive data in milliseconds to make decisions.

This requires powerful edge compute and next-generation wireless networks.

Or think about AI that manages global financial transactions. These systems rely on real-time data feeds and synced compute centers across continents.

Without a robust physical backbone, these use cases fall apart. A recent report stressed that future data centers will need more compute and stronger connectivity. In other words, it’s not enough to build powerful servers. They have to be connected at lightning speed.

The digital highways have to scale, or AI innovation stalls.

Despite their importance, many infrastructure companies trade below their true value. They were overlooked when revenues were flat and AI wasn’t mainstream. But that’s changing quickly. Now, the same companies that were ignored are becoming essential.

Companies Quietly Leading the Shift

Some firms are already positioned to benefit from this transition. Arista Networks makes the high-speed switches that move data through AI compute clusters. Marvell Technology builds the interconnect chips that allow GPUs to work in unison.

These are not speculative startups. They are established firms with real revenues and reliable cash flow. Many even pay dividends. Yet they are about to benefit from one of the most powerful technology trends of our time.

In the Premium Section below, I look deeper at the seven companies and investable areas my team and I are currently looking at. If you’re a paid-up Subscriber, make sure to print out (it’s a long one) that entire section because you’ll want these companies on your watch list.

The Global Picture Is Bigger Than Most Realize

The AI competition isn’t just happening in Silicon Valley. It’s playing out on the world stage. Nations are waking up to the fact that infrastructure is strategic. Countries that fail to invest in compute and networks will fall behind.

The United States currently leads in AI partly because its corporations have invested heavily in cloud and compute capacity.

But China is accelerating fast.

They’re building AI supercomputers and laying vast fiber networks. One of China’s cloud services now claims over 21 exaflops of compute power. With 21 exaflops, China’s AI system may rival the combined compute of dozens of advanced human brains working at full tilt, every second.

This kind of compute power was once reserved for nuclear simulations and military-grade space modeling. Now, it’s used to train AI.

In other words, the U.S. Department of Energy's Frontier supercomputer (the first official exascale machine in the U.S.) runs at 1.1 exaflops. It was built for national security-grade modeling like nuclear detonations, weather-driven battlefield dynamics, and advanced missile physics.

China’s 21-exaflop setup is twenty times faster.

And if that wasn’t mind-boggling enough, the total demand for data center capacity is expected to triple by 2030.

This will require upgrades in compute, power delivery, bandwidth, and cooling.

Governments will start treating data centers like oil refineries, and compute power is becoming a matter of national security.

What Investors Should Pay Attention To

If AI keeps growing—and all signs say it will—then the infrastructure behind it will become more valuable by the day. The best opportunities might not be in the companies building apps, but in the firms enabling them to run at scale.

We’re already seeing signs of this. Telecom companies are adding compute capabilities at the edge of their networks. Data center operators are partnering with chipmakers to design AI-optimized facilities. Equipment manufacturers are developing hardware that increases performance and reduces energy use.

AI needs to run 24/7, across the globe. That kind of scale can only happen if the underlying infrastructure is ready. Investors who understand this will have a real edge.

Many of these firms already have stable businesses. What’s changing is the growth potential. As AI workloads multiply, the value of owning the physical layer of the internet will rise.

And that layer is still undervalued.

AI’s impact won’t come from one killer app. It will come from systems that work together on time, at scale, without lag.

That’s why infrastructure is the real story.

We’re talking about the towers that carry signals, the fiber that moves data, the chips that process commands, and the servers that do the heavy lifting.

These components are critical to making AI practical and profitable.

For investors, this is a clear takeaway. The companies that build and operate the core of the AI economy will have staying power. While others chase short-term buzz, you can focus on long-term value.

In every era of innovation, those who owned the roads and rails often profited more than those who rode on them.

AI will be no different.

The builders of this digital backbone are in a strong position. The time to recognize that—and act on it—is now.

Double D

P.S. For paid-up Premium Subscribers, I dive into seven of the best companies to watch for. You can find it below:

🔓 Premium Content Begins Here 🔒