Citigroup calls “the U.S. stock seven giants are dead,” and shifts its bets to growth companies spanning six major industries

Citigroup strategist Scott Chronert’s team said bluntly that the old framework for gauging large growth stocks—the “Magnificent Seven”—has failed; in its place, the “growth cluster” spans nearly half the market value of the S&P 500, and year-to-date returns are almost double the broader market. (Background: The Magnificent Seven have started to underperform the broader market! Report: AI capex is eating Mag 7 free cash flow) (Additional context: Viewpoint: Semiconductors are taking over and leading the rally as the “US stock seven” move in the next leg, and the S&P 500 is preparing to break through 8,000 points)

Table of Contents

Toggle

• The Magnificent Seven fade, with Microsoft leading the decline
• Growth cluster: Three numbers show the gap
• Profit diffusion, cheaper valuations
• Uneven AI rotation; semiconductor rout is a warning
• The baton pass from enabler to adopter

Citigroup’s latest report threw out hard words: as a framework for assessing the dynamics of large growth stocks, Mag 7 (the “Magnificent Seven”) is dead. Replacing it is a basket of companies Citigroup calls a “growth cluster,” spanning six major industries, with a combined market cap of nearly half of the S&P 500, and year-to-date returns that are almost twice the broader market.

The Magnificent Seven fade, with Microsoft leading the decline

The Roundhill Magnificent Seven ETF is up only 1% this year, far behind the 9% gain in the S&P 500 over the same period.

Ever since the AI boom kicked off in late 2022, the Magnificent Seven have led the broader market; now as a group, their shine is fading. The worst performer is Microsoft: down 17% since 2026, with the latest round of selling triggered by market fears over massive AI capital expenditures (capex).

Valuation concerns, worries about capital spending, and—after AI tools became widespread—uncertainty about the software sector’s outlook are why investors are punishing some members particularly harshly.

Growth cluster: Three numbers to see the difference

Citigroup first introduced the concept of a growth cluster years ago, and has recently refined it: it is a basket of growth-oriented companies that have contributed the most to S&P 500 earnings over the past few quarters, spanning six industries. In total, they account for about half of the S&P 500’s total market cap. Since the start of the year, this group of stocks has easily left the broader market behind:

• Q1: S&P 500 fell 4.6%, growth cluster plunged 9.4%
• Q2: S&P 500 rose 14.9%, growth cluster surged 24.7%
• Year-to-date: S&P 500 up 10.1%, growth cluster up 11.8%

The growth cluster fell deeper than the broader market in Q1, but the rebound in Q2 was even stronger; ultimately it edged ahead slightly year-to-date while also beating both Citigroup’s own cyclical and defensive clusters.

Profit diffusion, cheaper valuations

The first reason is profit diffusion: take a weighted index formed from the top 25 stocks that contributed the most to S&P 500 returns year-to-date—up 7% since the start of the year. Applying the same algorithm to the Magnificent Seven yields only 2%. The strategist added: “Even a breakthrough of the Mag 10 would miss important earnings contributors,” and highlighted strong earnings from Intel, Applied Materials, and Lam Research.

These growth stocks have continued to substantially outperform analysts’ expectations, lifting their contribution share to about 48% of the S&P 500’s expected earnings for the next 12 months.

The second reason is valuation: the Magnificent Seven have been choppy as investors take profits and rotate into cheaper names, while the growth cluster—measured by the price-to-earnings-to-growth (PEG) ratio—sits at a 15-year low.

Citigroup wrote: “Under current conditions, forward-looking growth expectations reflect the sustained momentum in semiconductors/hardware driven by the AI capex tailwind, as well as a phase of growth surge for commodity-type semiconductors due to current bottlenecks. So what we get is a situation where the stock seems not to fully reflect long-term growth opportunities.”

Uneven AI rotation; the semiconductor drop is a warning sign

But there is also a counterpoint to this argument. AI trading in 2026 has been extremely uneven across capital rotations, with semiconductors and memory stocks bearing the brunt: iShares Semiconductor ETF is down 18% over the past month, while the Roundhill Memory ETF is down 32%.

“Valuation not fully reflecting” can also be read the other way: if the AI capex tailwind peaks sooner or bottlenecks ease sooner, today’s earnings surge may not persist, and the seemingly cheap valuations could be an illusion.

Citigroup itself also admits the measurement limitations: “We do not believe any one method can perfectly describe how much of the S&P 500 is reflecting AI trading. We believe that using a cluster approach to evaluate the S&P 500 is intuitively reasonable and very close. The conclusion is: measured by the growth cluster, AI influence shows that about 55% of the S&P 500 is directly affected by the AI tailwind/headwind, and nearly half of the index’s profits can be attributed to this group.”

The baton pass from enabler to adopter

Chronert actually predicted back in December 2025 that in 2026 the logic of AI trading would shift from “AI infrastructure providers” to “AI technology adopters,” matching the observation that the growth cluster spans six industries and earnings are no longer concentrated in a few mega-cap leaders.

Market conditions also confirm this rotation: Apple’s stock rebounded because it avoided joining the AI data center arms race on spending, while Microsoft and Meta remain under pressure due to continued doubts about when their massive capital expenditures will translate into returns. In its 2026 second-half outlook released on July 5, Citigroup also set its S&P 500 target price at 8,100 points, with the rationale being the continuation of the AI capex super-cycle.