Tech companies have been getting all the fame recently, but they’re not the only star performers this year.
Taken over by fear that they have a moral duty to push forward the artificial intelligence revolution, tech companies have thrown themselves into propping up and building data centers. The devoutness is most certainly admirable.
But in this desperate rush, they haven’t stopped to figure out how they’re going to keep powering these data centers. That oversight has put shares of power providers neck-in-neck with the goliath tech stocks.
The S&P 500’s information technology and communication-services sectors have jumped 15.65% and 14.97%, respectively, ranking at the top. Following closely behind this year are the industrials sector at 14.28% and utilities sector at 13.55%.
It’s entirely understandable, because some data centers require as much energy as a big city, making it a much harder area to tap into. And since META 1.07% Chief Executive Mark Zuckerberg has declared that “developing superintelligence is now in sight,” it’s likely that the global race heats up.
Zuckerberg has spent millions to secure talent that can help him achieve that goal, and way more on infrastructure. His company’s shares are up 28% this year. Microsoft MSFT -0.89% , Amazon AMZN -1.46% , and Alphabet GOOG -0.38% are also rapidly trying to expand in the space. Shares are up 25%, 1.1%, and 2.5%, respectively, this year.
It’s certainly the right move. Data centers that handle our AI-driven search inquiries and other loads are projected to require $5.2 trillion in capital expenditures by 2030, according to a McKinsey report published in April.
But equal, if not more, attention needs to be paid to the U.S.’s constrained power-generation capacity. Goldman Sachs Research expects a 2.5% increase in the compound annual growth rate for power demand between 2023 and 2030, with data centers making up 1 percentage point of that.
Another way to illustrate just how much that would be—The International Energy Agency estimates annual energy usage for data centers worldwide could jump above 945 terawatt hours by 2030, which would be equal to the current electricity consumption of the entire country of Japan. In 2024, data centers utilized 415 TWh.
The funniest part is that while these numbers may seem huge, it’s entirely possible that in the quest for superintelligence they all end up being quite modest estimates.
Tech companies have figured out that the easiest way to achieve their goals is to vertically integrate the business cycle, including owning physical assets. It’s ironic because for years these companies enjoyed an edge from not being bogged down by property, plant, and equipment due to large depreciation concerns. Nonetheless, their expansion is spurring a real-estate boom of sorts in Atlanta, Columbus, Dallas, and Phoenix, the FT reported.
Whoever can figure out the energy side of the equation stands to be the biggest winner. After all, it’s possible to get a data center up and running in a relative short time span, say 3-6 years, but a new power plant can take more than 10 years. Why? Because power plants face a much higher level of regulatory scrutiny, and their setup involves greater complexity. That’s why some data centers are being built alongside energy sources to ensure coverage. In October, Google started working with Kairos Power to develop advanced nuclear reactors close to data centers from 2030.
And to be sure, many of the Magnificent Seven have signed long-term power purchase agreements with nuclear developers. Microsoft, for example, plans to get energy from Constellation Energy’s CEG 1.82% restarting of the Three Mile Island nuclear plant in Pennsylvania. Amazon signed a 10-year agreement to get its power from Talen Energy’s TLN 0.02% Susquehanna nuclear plant, also in Pennsylvania.
It’s unclear if this will be enough. When training a new model, AI data centers note a huge uptick in power demand above the projections mentioned earlier, due to the processing power required to complete the calculations.
The Utilities Select Sector SPDR fund XLU 0.99% , which closely tracks the sector, has also hit all-time highs this month and is up 13% this year. Within its top 10 holdings, Vistra shares VST -0.95% have added 35% this year, DUK 1.27% Duke Energy has gained 16%, and Constellation Energy has increased 50%, with all three also hitting records in August.
Among industrial stocks, GE Vernova GEV 2.00% has been a big contributor and has emerged as the second-best performer in the S&P 500. It has posted an eye-popping gain of 92% this year and has hit fresh all-time highs.
GE Vernova GEV 2.00% is one of three companies that spun out of General Electric last year. It’s really set itself apart through its gas turbines, demand for which continues to be red hot. Equipment orders jumped more than 135% “which gives more confidence in the ‘stronger for longer’ narrative in the global gas market,” CNBC wrote, citing Wolfe Research analysts.
Shares of Bloom Energy BE -3.76% , another constituent, are up 55% this year, skyrocketing in July after it was announced that the company will deploy its fuel cells at select Oracle Cloud Infrastructure data centers in the US.
The renewed use of nuclear energy has powered a rally in uranium shares, too. The Global X Uranium ETF URA -2.43% is up over 57% this year. In its top 10 holdings, Cameco CCJ -1.59% has added 47% this year, while Nuscale Power SMR -4.14% has added 122%.
Conclusion
Traditional grids weren’t built to power AI. While Big Tech is taking steps to build new ones, if it doesn’t fill the gap soon, it’ll continue to lead to soaring electricity prices across the country. It will also contribute to water scarcity, as chips in data centers require large amounts of water to cool down. As long as power generation demand remains high, shares of the same could continue to do well.
As always, Signal From Noise should not be used as a source of investment recommendations. We encourage you to explore our full library, which includes deep dives on sovereign AI, military drone makers, the America First trade, and wonky economic indicators. You’ll also find discussions about consumer stocks, the TikTok demographic, and weight loss-related investments.
