The One Unshakeable Truth Behind A.I.

By TradeSmith Research Team

Listen to this post
Investors face constant, nearly irresistible temptations for promising investment stories that haven’t yet borne fruit.

Pie-in-the-sky promises that wind up less impactful than anticipated, or even just take a long time to reach mass-market appeal, are what bubbles in premature tech are made of. (See the past bubbles in blockchain, 3D printing, and the then-unproven internet of the late ’90s for evidence.)

I get the temptation. Because many of those narratives mint millionaires. Still, we must be wary of stories.

Now, I’m about to say something potentially controversial.

The elephant-in-the-room investment narrative, right now, is A.I.

Many folks are convinced that A.I. is here to stay, and with good reason. Already, it’s seen a lot of useful applications — like digital twin testing, which we covered not long ago. ChatGPT, with its 180 million users in just over a year, speaks for itself.

Still, it’s uncertain just how much A.I. is set to disrupt. Can it really replace 2.4 million U.S. jobs by 2030, as research firm Forrester predicts? Will it really boost GDP by nearly 15% by the same time, as PricewaterhouseCoopers estimates?

It might… but we should never invest on hopes and projections alone.

So what can we do?

Focus on the certainties.

If A.I. is here to stay — and it probably is — what can we count on happening?

Well, for one, we can count on A.I. consuming a tremendous amount of energy. Data centers, some of which house A.I. models, already consume an estimated 1% of the world’s energy consumption.

And that’s just getting started. A.I. is forecast to grow its energy consumption alongside its adoption, with one report expecting it to consume more than some small countries by 2027.

So if A.I. is already consuming so much energy… And is set to consume more… It tracks that we should have exposure to energy technologies that are positioned to meet A.I.’s growing demand.

Today, we’ll look at the current energy situation with A.I. — and some possible solutions.

And I’ll share one critical energy technology that’s woefully underinvested but is clearly the trade to make.

What Powers A.I.?

To understand the A.I. energy picture, it’s helpful to zoom out and think about what A.I. really is.

The best analog for A.I. today is cloud computing, which is simply racks upon racks of powerful computers whose power we can access remotely through the internet.

A.I. is powered much the same way: in data centers full of computers running A.I. algorithms. That’s why semiconductor companies, like Nvidia and AMD, caught such a tremendous bid last year. Any company that makes market-leading computer chips is in a great position to take advantage of the A.I. trend.

Computer chips are one side of the equation… And there’s plenty of trades to make there — especially in companies that flash one of the most important buy signals in the market.

But energy is the other, equally important side that’s yet to see nearly as much interest. A.I. is set to demand more energy than any technology that came before, bar none. Even right now, it’s a more energy-intensive technology than what it’s working on replacing.

OpenAI found last year that a Google Search query consumes an estimated 0.0003 kWh (kilowatt hours) of electricity. A query with ChatGPT consumes anywhere from 0.001 to 0.01 kWh of electricity — or anywhere from 3 to 33 times more.

Granted, Google is still pulling many times more queries a day than ChatGPT, but the search giant is only growing at about 10% per year. A.I. is growing much faster, with ChatGPT alone growing to 100 million users in less than a year.

The point is, this is a big shift… and that’s just ChatGPT. There are dozens of other generative A.I. tools out there sucking up energy.

And even this isn’t counting the A.I. capacity set to come online. Consider that Nvidia plans to ship 1.5 million dedicated A.I. servers over the next four years.

And the fact that Alex de Vries, a researcher at the School of Business and Economics at the Vrije Universiteit Amsterdam, recently estimated these new servers alone would demand more than 85 terawatt hours of energy annually. That’s more than some small countries.

His research also estimated that on the whole, global datacenter energy demand could jump 50% by 2027.

These are projections, just like the projections about job losses and GDP boosts. But an unshakeable truth about A.I., regardless of how it gets used in the future, is that it consumes a ton of energy.

And with the push from world governments to shift away from fossil fuels and toward renewables, we can be sure that these A.I. datacenters won’t be powered exclusively by coal-fired plants and oil pipelines.

We need to focus on the technologies that can provide clean energy at a global scale.

And if you’ve been following along with TradeSmith Daily, you’re already well ahead of the crowd.

The Nuclear, Green, A.I. Future

Add A.I. to the growing pile of reasons why you should invest in nuclear energy stocks right now:
  1. They’re still woefully misunderstood, with a bad PR problem despite being cleaner and safer than coal, oil, natural gas, and even wind turbines…
  2. They’re showing the first strong uptrend in decades…
  3. Even though it’s up another 30% since I last wrote you about nuclear, the sector is still cheap compared to the S&P 500 (22.4 vs. 24.3)…
But don’t just take my word for this. The preeminent mind in A.I., OpenAI CEO Sam Altman, is saying the same thing.

Here he is speaking at the World Economic Forum in Davos, Switzerland, last week:

“There’s no way to get there without a breakthrough,” he said. “It motivates us to go invest more in fusion.”

And Reuters noted:

In 2021, Altman personally provided $375 million to private U.S. nuclear fusion company Helion Energy, which since has signed a deal to provide energy to Microsoft in future years. Microsoft is OpenAI’s biggest financial backer and provides it computing resources for A.I.

Altman said he wished the world would embrace nuclear fission as an energy source as well.

The guy who brought ChatGPT to the world is saying we need more nuclear energy capacity, and is directly investing in nuclear energy himself.

It’s no wonder that uranium, the commodity underlying nuclear energy stocks, has been on an absolute tear.

I’ve been tracking this chart for several months now — of uranium against oil, natural gas, and coal.

You can see that uranium continues to leave the fossil fuels in its dust… and while they’ve all barely treaded water in 2024, uranium is up more than 15%.
Long story short, you need to have nuclear energy stocks in your A.I. investment plan, right alongside high-quality semiconductor companies.

As far as I know, there is no better speculation on the future of not just A.I., but the global energy story period.

The easiest way to do this is with the Global X Uranium ETF (URA), which holds a basket of energy companies. This ETF will give you broad exposure to the trend.

But if you want to find the best stocks, that’s where TradeSmith’s best-in-industry tools come in.

One of the easiest ways to find high-quality stocks in any trend, sector, or market is with Ratings by TradeSmith.

With a few taps of your keyboard, you can immediately see if a stock is worth your investment dollars… or one to avoid entirely.

For example, here’s the rating for Cameco (CCJ), the largest publicly traded uranium company.
Cameco earns a Strong Bullish Rating at 90, making it a prime target for any investor looking to invest in the nuclear energy trend. And you can do this with any nuclear stock in the URA ETF, helping you find the best stocks to buy.

I also like to plug some of my own holdings into Ratings to see if I should keep holding, cash in my gains, or cut my losses.

It all starts in your TradeSmith Finance dashboard, which you can access right here.

To your health and wealth,

Michael Salvatore
Editor, TradeSmith Daily