You ask AI a question. A whole town worth of infrastructure has to switch on.
Open ChatGPT. Type a question. Three seconds later, an answer. Feels like magic.
It is not magic. It is a building the size of a hockey arena, full of thousands of computer chips, gulping more electricity than a small town. Miles of copper pipes keep it cool. Old nuclear plants are being pulled out of retirement just to feed it.
This is the part of the AI story almost nobody is talking about. And it is where we think the real money will be made. The four biggest tech companies will spend $725 billion in 2026. That is up 77% from last year. It is more than the entire economy of Sweden. Every one of those dollars has to land somewhere physical. A chip factory. A power line. A pipeline. A copper mine.
Think back to the California gold rush. Most of the people who got rich were not the prospectors. They were the ones selling the picks, the shovels, and the blue jeans. Levi Strauss never panned a single nugget. He made a fortune anyway. AI is the new gold. The chips, the power, the copper, the pipelines, those are the shovels. That is what we own.
The world is about to need a whole new Japan worth of electricity.
Here is the number that should stop you. Today, all the world's data centres combined use about as much electricity as a mid-sized country. By 2030, they will use as much electricity as all of Japan. In six years.
Try this on for size. Imagine every single house in Canada plugged in four extra fridges and never unplugged them. That is roughly the new electricity demand coming down the pipe. Now ask yourself: who built that grid? Who supplies it? Where does the metal come from?
Honest answer: nobody has fully solved it yet. New big power lines take ten years to build. New power plants take longer. AI does not have ten years. That gap is the whole opportunity.
You can build an app in a weekend. You cannot build a power plant in a weekend.
The boring metal that just quietly hit an all-time high.
Copper hit an all-time high price of $14,527 per tonne in January 2026. It is still trading near those levels today. Over a five-fold gain since 2008 and up roughly 30% in the past year. The kind of move you usually only see in a once-in-a-generation supply crunch.
Why copper? Take a quick look around your house. Every wire in your walls. Every power line out the window. Every transformer on every pole. The motor in your fridge. The cable behind your TV. All copper. There is no electricity, anywhere, that does not run through it. And there is no AI without it.
Here is a stat that makes it click. A regular gas car has about 25 kilos of copper in it. An electric car? Closer to 100 kilos. Roughly the weight of a small person. Now add data centres, plus an entire grid that has to be rebuilt, plus the global push to electrify everything.
The world wants way more copper than it can dig out of the ground. And new copper mines? They take ten to fifteen years to permit, finance, and build. There is no quick fix. That is the setup.
Microsoft is restarting Three Mile Island. Yes, that one.
In 2024, Microsoft signed a 20-year deal to restart Three Mile Island. The famous nuclear plant in Pennsylvania. The one from the 1979 partial meltdown that basically killed nuclear power in America for forty years. Microsoft is paying to bring it back to life. All 835 megawatts of it. Just to feed its AI.
And that is not a one-off. It is the new playbook.
Picture this. Imagine Costco signed a 20-year contract to buy every single egg from one specific farm. Not some of the eggs. All of the eggs. For two decades. That is essentially what the tech giants are doing with America's nuclear power plants.
They stopped waiting for the power grid to catch up. They are paying for it themselves.
For twenty years, nuclear was stuck. The plants were too expensive to build, and nobody wanted to commit to buying their power for decades. The economics did not work. The tech giants just changed that overnight. A 20-year contract from a company worth three trillion dollars turns a struggling plant into a goldmine. That is why the companies that own these plants have been some of the biggest winners of the AI boom so far.
Four pillars. One simple idea.
This is what ALPHA AI is built to own. Not the AI apps. Not a bet on which chatbot wins. The four physical things every AI app needs to actually work.
Chip foundries
Equipment makers
Nuclear operators
Uranium
LNG exports
Power-plant fuel
Lithium producers
Rare earths
Examples shown to illustrate the categories. Not recommendations to buy.
Why four pillars instead of just one? Because they each move on their own rhythm. Chips move when tech earnings come out. Power companies move slowly on long-term contracts. Pipelines move with global demand for gas. Mining moves on multi-year cycles. When one is having a rough year, the others usually are not. That is the whole idea.
Think of it like owning every team in the playoffs instead of trying to pick the winner.