Why Morgan Stanley Is Telling Clients to Trim Tech and Buy Energy Before Summer

There is a specific type of Wall Street call that appears as a quiet, methodical repositioning rather than a dramatic reversal; it doesn’t make the front page on any one day but influences portfolio allocations for months afterward. In recent weeks, Morgan Stanley has been advising clients to increase their energy exposure and decrease their technology overweights in anticipation of what the bank believes will be a structurally favorable summer for power infrastructure and oil demand. It’s not a tech panic. Beyond that, it’s more measured. However, because it touches on a convergence of factors that most market participants haven’t fully connected yet, the reasoning is worthwhile to comprehend.

Let’s start with technology. For the better part of three years, US equity returns were dominated by a small number of mega-cap technology companies known as the Magnificent Seven. These companies have experienced a quiet but significant deflation. Not a breakdown. a steady decline. Over the course of about three quarters, Nvidia’s forward price-to-earnings ratio shrank from the low 30s to about 20, despite the company’s earnings continuing to grow at an astounding rate. As a result, the underlying business is essentially unchanged, but the company is trading at a fraction of its previous premium. The S&P 500’s forward P/E multiple has already dropped 18% from its six-month peak, according to a recent weekly note from Michael Wilson, chief US equity strategist at Morgan Stanley. This level of repricing typically indicates a correction that is well advanced in both time and price rather than one that is just getting started.

Peter Oppenheimer of Goldman Sachs observed the same thing from a different perspective. He claimed that since the early 1970s, the technology industry has recently experienced one of its worst periods of relative underperformance in comparison to the larger global market. Eighteen months ago, this positioning would have seemed unrealistic. Today, the IT sector trades at a forward P/E below consumer discretionary, consumer staples, and industrials. Wilson and Oppenheimer both consider the Magnificent Seven, which are currently trading at about 24 times forward earnings, to be genuinely appealing from a strict value perspective in relation to their growth profiles. After a year of multiple compression, the math has significantly improved.

Therefore, the tech deflation serves as both a prelude to the energy rotation and an acknowledgment that portfolios that rode the AI wave for three years have probably drifted overweight technology without clients consciously choosing to make that wager. The practice aspect of this was explained simply by David Royal, chief investment officer at Thrivent: discipline demands you to rebalance, regardless of how much you like the names, when you’re aiming for a 65% equity allocation and the rally has pushed you 5% above that. Even in cases where the fundamental view remains unchanged, a sector may experience selling pressure due to a quiet overweight correction.

The energy call, however, is more than just a relative value ploy against technology. Since late 2025, Morgan Stanley has been constructing the structural case for energy demand. If you’re watching power sector stocks without considering data centers, it’s easy to overlook the AI connection. Global electricity demand is expected to increase by more than one trillion kilowatt-hours annually through 2030, with AI-driven data centers accounting for nearly one-fifth of that growth, according to a thorough analysis published in February 2026 by Morgan Stanley’s own research institute. Attendees at the bank’s own conferences on powering AI, which took place in November and December of 2025, came to the same fundamental conclusion: the grid wasn’t built for this degree of demand concentration, and it isn’t keeping up.

The problem is made tangible by the numbers. According to Morgan Stanley Research, the power demand for US data centers could reach 74 gigawatts by 2028, despite an estimated 49 GW of power access shortfall. In Arizona, Texas, and Virginia, data centers of a size that would have been unheard of just five years ago are being planned. These sites require 1 to 4 gigawatts per campus, and they are unable to obtain grid connections due to interconnection queues that last for years and equipment costs that have increased by 30% since 2019. The biggest hyperscalers are not waiting for a solution. The goal is to supply their own energy via battery storage, nuclear power purchase agreements, microgrids, and natural gas generators. At businesses that were previously totally grid-dependent, the phrase “bring your own power” has evolved from novelty to operational strategy.

The energy trade is thus created. Outside of China, natural gas is expected to supply about one-fifth of the world’s new electricity needs. Once politically dubious as an investment thesis, nuclear is now drawing significant funding from tech firms looking for long-term power contracts with a reliability profile that intermittent renewables cannot match. Infrastructure developers, grid stability software providers, and suppliers of power equipment are all well-positioned to profit from what Morgan Stanley’s Stephen Byrd called a “upcoming infrastructure CapEx cycle” that will generate concentrated pockets of value creation throughout the whole power supply chain. According to the bank, growing power spreads, or the difference between the cost of producing electricity and the price at which it is sold, could increase by 15%, generating $350 billion in value for the sector.

It’s difficult to ignore the fact that this call necessitates holding two positions at the same time: a bullish assessment of AI’s long-term economic effects and a bearish-to-neutral assessment of AI stocks in particular for the near future. The AI revolution is far from over, according to Morgan Stanley. Wilson is not giving up on technology; rather, he is rebalancing it by recommending high-quality hyperscalers in addition to cyclicals. According to the bank, the physical infrastructure that powers AI will be more valuable in the upcoming phase of its implementation than the software companies that write the prompts.
This article is not financial advice; it is merely meant to be informative. Prior to making any investment decisions, always seek the advice of a qualified financial advisor.