|
Data Centers’ Soaring Power Needs
|
|
According to data from Cleanview, there are 1,290 data centers operating in the U.S. that can consume 61 gigawatts of electricity.
|
|
An additional 2,019 planned data centers will add another 377 gigawatts of electricity demand – a 618% jump in power needs.
|
|
The chart below shows planned data center expansions by state compared to those currently in operation.
|
|
|
|
Source: Apollo
|
|
Certain states like Texas, Virginia, and Utah are seeing the largest concentrations of planned additions. And the strain on the power grid will be enormous.
|
|
Data centers currently consume around 6% of the nation’s electricity. That figure jumps to 12% by 2030 and could rise to 20% within the next 10 years.
|
|
For all the excitement around semiconductor companies and other advanced technology surrounding AI, the new bottleneck is far more fundamental.
|
|
Simply meeting the power needs for current and future data center demand is emerging as a crunch point for enabling AI applications.
|
|
That’s why the new winners of the data center buildout are companies found operating inside the AI power stack.
|
Tune in to Trading With Larry Live 
Each week, Market Wizard Larry Benedict goes live to share his thoughts on what’s impacting the markets. Whether you’re a novice or expert trader, you won’t want to miss Larry’s insights and analysis. Even better, it’s free to watch. Visit us on YouTube to catch the latest! |
|
|
The AI Power Stack
|
|
The explosive growth of AI is reshaping not only technology but the entire U.S. energy system. The data centers that train and run AI models need steady, reliable power on a scale we’ve never seen.
|
|
Unlike traditional computing loads, AI facilities operate around the clock with minimal tolerance for interruption. In order to meet that high run time, there are critical points in the AI power stack’s supply chain.
|
|
That includes companies that produce energy for power plants to turn into electricity, where natural gas will play a major role in meeting data center demand.
|
|
But you also need to transport gas to where it’s consumed, which makes midstream energy companies essential in the power stack. Midstream companies operate pipelines along with processing and storage facilities for energy products.
|
|
Given the strains already present on public power grids, hyperscalers are also turning to independent power producers, which own power plants running on gas and nuclear.
|
|
And speaking of nuclear, the industry is seeing a resurgence given its carbon-free, always-on electricity that’s well suited for meeting data center uptime demand.
|
|
Semiconductor stocks have seen a massive profit tailwind from data center capex. Energy companies operating in the AI power stack are next in line.
|
|
That’s why, if you didn’t tune in to my AI Retirement Reset event last night, I’d encourage you to give it a look. I shared the name of a top ticker that I believe is going to benefit from the trillions about to flow into energy production. And that ticker is just the start.
|
|
|
|
|
|
Regards,
|
|
Larry Benedict
Editor, Trading With Larry Benedict
|
|
|
|
|
Get Instant Trade Alerts on Mobile!
|
|
|
|
Click the icon below from your mobile device to download The Opportunistic Trader app today for one-tap access to trade alerts, issues, and model portfolios for all of Larry’s services.
|
|
Available in the app store on Android and iPhone.
|
|
|
|
0 Response to "Why Power Could Be AI’s Next Bottleneck"
Post a Comment