Why It’s Not Too Late to Swap Tech for Tech Beneficiaries, From Utilities to REITs
“Our cycle model, the Regime Indicator, just shifted from Upturn to Downturn,” wrote the stock market strategists at BofA Securities this past week. That had my attention, even though my own model, Tactical Sloth, calls for ignoring market cycles. I don’t do sector or theme rotations. I’ll do a tire rotation, but grudgingly.
Sector swappers are abuzz about power companies outrunning artificial-intelligence giants since the end of June. The iShares U.S. Utilities exchange-traded fund has returned 12%, while the iShares U.S. Technology ETF has slipped 4%. For an S&P 500 fundholder, it’s no big whoop. The stuff that’s working has outweighed the stuff that isn’t, and the fund has returned a couple of percent over that stretch. For more-tactical investors, it raises the question of whether it’s too late to take tech profits and buy something boring.
It isn’t too late, according to BofA. Tech looks “egregiously expensive,” and earnings growth there is impressive but decelerating. The S&P 500 carries “extreme concentration risk,” with Apple, Microsoft, Nvidia, Alphabet, and Amazon.com together making up more than a quarter of the index. Volatility is likely to be elevated in years ahead, says BofA, and for that, investors will want quality, stability, and income. Over the past decade, dividends have contributed just 16% of returns, but going forward, they could chip in something closer to their historical average of 40%.
That favors utilities, which pay close to 3% in dividends. They remain cheaper than average relative to the S&P 500. And BofA notes that utilities aren’t as sleepy as billed; since 1980, they’ve returned 11% a year, close to the Nasdaq’s 12%. Utilities are quiet tech beneficiaries, especially as AI data centers create soaring demand for power.
The real estate and energy sectors have even higher dividend yields than utilities. Buy real estate investment trusts, says BofA. They look cheap relative to funds from operations, and troubled office REITs have a tiny sector weighting relative to healthier players in telecom towers, retail, healthcare, and data centers. REITs have done even better than utilities since summer. Energy, on the other hand, is a “value trap,” where rising supply and slowing demand could create an oil glut next year.
You might think that consumer staples and healthcare fit well into the new call for stability and dividends. In good times and bad, we spend with snack companies to get fat, and with drug companies to treat our resulting diabetes and heart disease. But the new obesity meds threaten this symphony of capitalism. BofA is Underweight both staples and healthcare. Instead, it favors banks, which have steady earnings and lower leverage than in the past, and consumer discretionary companies, which benefit from rising wages and falling interest rates.
Changing topics, sort of: This past week I spoke with Stephen Byrd, who once worked in the power industry and now heads sustainability research at Morgan Stanley. He sees a “severe shortage” of U.S. data centers brewing, and power is a key constraint. To accommodate the number of chips being sold, the industry will need new data-center capacity equal to 10 gigawatts of electric power, but it’s building about half that. Some markets have a shortage of electricity being generated; others see long delays for new hookups. One key county for data centers in Virginia sent notices to prospective customers saying the timeline could be seven years. “Seven years in the world of AI is a lifetime,” says Byrd.
That bodes well for some obvious beneficiaries, and some less obvious ones. It suggests healthy growth ahead for solar power and natural gas; think First Solar and GE Vernova. Cummins could gain from more demand for backup power. Growing power demand makes a U.S. nuclear renaissance increasingly likely. Hyperscale cloud companies could even make deals directly with nuclear power companies.
Earlier this year, Amazon bought a 960 megawatt data-center campus from Talen Energy, which emerged from bankruptcy last year. The campus is powered directly by Talen’s nearby Susquehanna Steam Electric Station, which generates 2.5 gigawatts of power. Ongoing power sales to Amazon will give Talen steady cash flow. Talen shares have tripled to over $150 since last fall.
“We believe Constellation [Energy] and Vistra will do the exact same thing,” says Byrd. “They’ll sign extremely large power deals with very large companies that want to build supercomputers.”
And then there are the crypto miners. I’ve written some skeptical things about the long-term usefulness of crypto. It turns out that one use is that if you set up a large-scale crypto mining operation and squander vast amounts of electric power to produce make-believe internet coins, and if there’s an acute shortage of power hookups for AI data centers, then companies will pay you to stop making make-believe coins, so they can use your power hookups.
“These Bitcoin sites can often offer a three- or four-year time advantage relative to connecting a data center to the grid the old-fashioned way,” says Byrd. “For a hyperscaler, which is extremely interested in getting its chips powered up, the value of time is immense. These chips depreciate so quickly.”
An April Bitcoin event called the halving, which reduced the financial rewards for miners, has many of them eager to find new business models. Byrd points to Core Scientific, which recently emerged from bankruptcy, as an example of how that can pay off for investors. It has partnered with a Nvidia-backed start-up called CoreWeave to make its infrastructure available for data centers. Since spring, Core Scientific shares have risen from $3 and change to more than $10.
Byrd’s price-per-watt calculations point to potential upside for Cipher Mining, BitDeer Technologies, Applied Digital, and Hut 8.