As the AI boom drives markets, investors may be looking for alternative avenues to capitalize on this emerging trend. Tim Kramer, CEO of CNIC Funds, appears on Wealth! to explore why the energy sector could be an attractive entry point into AI trading.
Cramer emphasizes that electricity is “the most consumed commodity in America.” He compares AI trading to the gold rush, suggesting that gold was the hot commodity at first, but attention quickly shifted to the gold rush and the companies and individuals who supplied the trade. He expects a similar pattern for AI and electricity.
“When you think about the context of how much electricity this is going to require, we think we're in a very interesting position to give people exposure to some of this growing segment of electricity consumption,” Cramer told Yahoo Finance.
For more expert insights and the latest market trends, click here to watch this full episode of Wealth.
This post Angel Smith
Video Transcript
In the United States, artificial intelligence data centers are expected to add 323 terawatt-hours of electricity demand by 2030.
According to Wells Fargo, that's enough electricity to power Italy in a year.
And the energy sector could be a back door into AI deals.
As part of our ETF report brought to you by Invest oqqq, we welcomed Tim Kramer, CEO of CN IC Fund.
So let’s get started!
This is particularly interesting, Tim.
So how can you position your portfolio for AI through a different area, the utilities that are actually required to power these data centers?
No, that's a great way to look at it.
This makes electricity the most consumed commodity in the United States on a retail basis.
But so far, it hasn't been included in any mutual fund, any index, any ETF.
So we partnered with ICE Intercontinental Exchange, which owns the New York Stock Exchange, to create the Power Index.
It combines electricity futures from six different U.S. regions and weights them together, and is backed by exchange-cleared futures.
And that product will allow us to actually contribute to the electrification of America.
So with that in mind, when you think about why there isn't a lot of AI trading happening in portfolios, or to what extent investors are doing this aspect of AI trading in their portfolios, you see that finally people are starting to take notice of this.
And then when AI trading emerged, it was like a gold rush, and in the early days, people wanted gold, but then they realized they were probably better off selling picks and shovels and jeans, which is probably a better risk-adjusted proposition.
And now people are trying to choose direct participants in the AI sector.
Now they're looking to look at some of the more common ones, like power and companies that might be providing copper transmission to the data center.
This is a relatively new thing – we only built this a year ago – but people are definitely starting to take notice.
With all this in mind, we looked at a few different aspects of this AI trade going forward.
So, as Mark Zuckerberg explained on the earnings call, one of the meta-platforms has applications that sit on top of language learning models, which sit on chips in the data center.
But we've never talked about what the chips in the data center sit on — the real estate, the utilities and power that goes into all of this.
So as this technology becomes more AI-enabled, what are the key ways we can see this transition over the next few years as it changes how ETFs operate?
Yeah.
of course.
And that was a great introduction with how those things overlap and the Zuckerberg quote.
That means a Google search takes 0.3 watt-hours.
Sending an email takes 1 watt-hour.
To do an AI search, it currently takes 4 watt-hours.
That's more than 10 times the normal search rate.
And by 2030, it will take 15 watt-hours to perform the same search.
The reason is that the databases are much larger and the chips they use are much more advanced, requiring more cooling.
Against this background, how much power does this require?
Well, I think we're in a really interesting position in terms of exposing people to a part of the sector that is growing electricity consumption.
This is a whole new wave of demand and it's going to force these companies, utilities, commodity industries to make sure that the demand profile of their business allows them to potentially make more money, but also how well positioned are they to make sure they have enough infrastructure, enough capacity, and know for sure what it's going to cost to bring more capacity online in the near future.
Oh, so you're kicking the hornet's nest with that.
So, if you want to put a data center, it's going to consume a huge amount of electricity, and if you want to put a data center, it's going to basically increase the electricity bill for everybody, including private investors.
And right now, nationally, I think over 20% of Americans are behind on their utility bills.
And then as demand increases, they start increasing power.
So, supply and demand comes into play when you start increasing prices and that gets passed on to retailers.
So the direct answer to your question is that people are starting to look for alternatives.
One is, if you're building a data center, can you locate it in a place that provides its own power?
And then there's the issue of developing the infrastructure to make it happen.
The second thing we're seeing is some power companies saying, “If you put a data center in, you need to go in with a contract in your right hand that says we're going to secure 90% of your power for the next 10 years.”
You will have a contract in your left hand stating that you can pay for any system upgrades or transmission enhancements that may be required at your data center.
The impact is there, the question is how this will progress.
It's not clear yet, but it will be more expensive for everyone.
Tim Kramer, CEO of CN IC Fund.
Thank you so much for taking the time to explain this here.
Thank you for giving me the opportunity.
appreciate.
