For years, the Magnificent Seven tech giants commanded investors' attention, dominating the S&P 500 and dictating which way the overall market moved. Those days are officially over: the tech-heavy Nasdaq 100 is up almost 18% in 2026 and the S&P 500 has climbed 10%, but an index of the Mag 7 stocks has gained just 1.1%. KeyToFinancialTrends reads that gap as evidence the artificial intelligence trade hasn't ended, it has simply moved – away from the companies spending the most on AI infrastructure and toward the companies supplying the physical components that infrastructure requires.
Nowhere is that shift clearer than in memory chips. The Philadelphia Stock Exchange Semiconductor Index is up 82% in 2026, on pace for its best year since 1999 and coming off its best quarter ever, driven by names like Micron Technology and Sandisk rather than Nvidia, the chip stock that has long dominated the AI narrative. Nvidia is, remarkably, only the third-worst performer in that same semiconductor benchmark this year, up a comparatively paltry 4.9%. KeyToFinancialTrends treats Nvidia's underperformance relative to memory chipmakers as the clearest sign of how specific this rotation has become: investors aren't simply chasing "AI stocks" as a category, they are distinguishing sharply between companies whose AI-driven demand is already showing up in pricing power and those still being asked to prove their AI investments will pay off.
The scale of the market's repositioning shows up starkly in correlation data. In April, the Mag 7's 40-day correlation to the Nasdaq 100 peaked above 0.95, a near-perfect lockstep; it has since dropped below 0.7, its lowest level since 2017. Big tech and the S&P 500 are now moving together about as loosely as they did back in 2015, when those same companies made up only 10% to 11% of the index, according to DataTrek Research co-founder Jessica Rabe. Brian Barbetta, co-head of Wellington Management's technology team, put it bluntly: the Mag 7 used to be one of the only reliable sources of earnings beating the market, and now investors are more focused on reasons not to like them.
Fund flows confirm the rotation is showing up in real money, not just index math. Investors pulled a record $786 million in June from the Roundhill Magnificent Seven ETF while pouring $9.3 billion into the Roundhill Memory ETF, according to data compiled by Bloomberg, and Deutsche Bank strategists noted that positioning in large-cap tech, which was "extreme" at the end of May, has since returned to a more neutral stance. KeyToFinancialTrends sets those flow numbers against the underlying concern driving them: Microsoft, Amazon, Alphabet, and Meta are all dramatically accelerating capital expenditures on AI infrastructure, weighing on cash flow at a moment when it remains genuinely unclear how well those investments are paying off, with Meta reportedly telling employees internally that its AI agent development has not accelerated as fast as expected.
The earnings data underlying the rotation is stark and getting starker. The Mag 7 are expected to post an 18.9% increase in net income next year, down from a 21.4% estimate just three months ago, while earnings estimates for chipmakers have been revised dramatically higher over the same period, to 48.5% from 34.3%, according to Bloomberg Intelligence. Not everyone expects the divergence to hold: JPMorgan strategist Nikolaos Panigirtzoglou argued the gap will narrow as hyperscalers and AI model providers eventually see monetization catch up with spending, while Morgan Stanley's Mike Wilson has separately warned that this degree of divergence "can't continue" as unsustainable. Key To Financial Trends frames that disagreement as the real trade to watch heading into earnings season: whether Big Tech's spending finally shows up as revenue before investors' patience runs out, or whether memory chipmakers keep capturing the market's attention simply because their growth is already visible in the numbers rather than promised for later.
