Citi strategist Scott Chronert and his team said in a new report that the “Magnificent Seven” is no longer a useful framework for tracking large-cap growth stocks in the U.S. market. The bank is replacing that lens with what it calls a “growth cluster,” a basket of companies spanning six sectors and representing roughly half of the S&P 500’s total market value.
The shift comes as the old mega-cap trade has lost its uniform leadership. Roundhill Magnificent Seven ETF is up just 1% this year, far behind the S&P 500’s 9% gain over the same period. After leading the broader market since the AI boom began in late 2022, the group has faded as a collective trade. Microsoft has been the weakest member, down 17% in 2026, with the latest selloff tied to market concerns over heavy AI capital expenditure, or capex.
Citi said valuation concerns, spending pressure, and uncertainty over the software sector’s outlook as AI tools become more widely used have all contributed to a harsher investor response toward some members of the group.
A broader basket across six industries
Citi said it first introduced the growth cluster concept several years ago and has recently refined it. The basket is made up of growth companies that have contributed the most to S&P 500 earnings over recent quarters. Together, they span six sectors and account for about half of the index’s market capitalization.
Its performance this year has shown deeper swings than the benchmark, but also stronger rebounds:
- First quarter: the S&P 500 fell 4.6%, while the growth cluster dropped 9.4%
- Second quarter: the S&P 500 rose 14.9%, while the growth cluster jumped 24.7%
- Year to date: the S&P 500 is up 10.1%, while the growth cluster has gained 11.8%
Citi said the group underperformed more sharply in the first quarter, then rebounded harder in the second, leaving it slightly ahead of the broader market on a year-to-date basis. The bank added that it also outperformed Citi’s own cyclical and defensive clusters.
Earnings leadership is no longer concentrated in a few names
The first reason Citi gave is earnings dispersion. The bank said a weighted index built from the 25 stocks making the largest contribution to S&P 500 returns this year would be up 7% year to date. Using the same method on the Magnificent Seven produces a gain of only 2%.
“Even a Mag 10 expansion would miss important earnings contributors,” the strategists wrote, pointing to strong earnings from Intel, Applied Materials, and Lam Research.
Citi also said this broader growth cohort has continued to beat analyst expectations by a wide margin. Its contribution has risen to about 48% of the S&P 500’s expected earnings over the next 12 months.
Valuations look cheaper, but the risk case remains
The second argument is valuation. Citi said the Magnificent Seven has turned choppy as investors take profits and rotate into less expensive names. On a PEG ratio basis, the growth cluster now sits at a 15-year low.
Citi wrote: “In the current setup, forward growth expectations reflect sustained momentum in semiconductors and hardware from the AI capex tailwind, as well as a temporary surge in growth for commodity semiconductors due to current bottlenecks. The result is a market where stocks do not appear to fully reflect the long-term growth opportunity.”
Still, the bank said the same argument can be read in reverse. If the AI capex tailwind peaks earlier than expected, or if existing bottlenecks ease sooner, the current burst in earnings may not last, and cheap-looking valuations may prove misleading.
AI rotation in 2026 has been uneven
Citi said AI-related trading in 2026 has not moved evenly through the market. Semiconductor and memory shares have been the hardest-hit areas in the latest rotation. iShares Semiconductor ETF has fallen 18% over the past month, while Roundhill Memory ETF is down 32%.
The bank also acknowledged the limits of any framework used to measure how much of the S&P 500 reflects the AI trade. “We do not think there is any single method that can perfectly describe how much of the S&P 500 reflects the AI trade. We believe the cluster method is intuitive, reasonable, and close,” Citi wrote. “The conclusion is that, using the growth cluster approach, about 55% of the S&P 500 is directly exposed to AI tailwinds or headwinds, and nearly half of index earnings can be attributed to this group.”
From AI enablers to AI adopters
Chronert had already argued in December 2025 that the market’s 2026 AI trade would shift from “AI infrastructure providers” to “AI adopters.” Citi said that view is consistent with the growth cluster framework, where earnings are spread across six sectors rather than concentrated in a small group of tech giants.
Recent market action has lined up with that rotation. Apple has rebounded as it avoids the spending race around AI data centers, while Microsoft and Meta remain under pressure over questions about when large capex plans will translate into returns. In its July 5 outlook for the second half of 2026, Citi set a year-end S&P 500 target of 8,100, citing the continuation of the AI capex supercycle.