The Magnificent Seven returned 27.5% in 2025. The S&P 500 returned 17.9%. By every headline measure, tech crushed the market for the third straight year.
Except it didn't. Not really.
Only two of the seven stocks actually beat the index. Alphabet, up 65.8%, carried the group on the back of its TPU chip business and a lucky antitrust ruling. Nvidia added 36%, still riding the GPU wave that made it the world's most valuable company. The other five? Microsoft managed 16.6%. Meta scraped 11.4%. Apple delivered 10%. Amazon, supposedly the backbone of cloud computing, returned 5.5%.
That's not a tech rally. That's two stocks dragging five laggards across the finish line.
The Concentration Problem Nobody Wants to Talk About
Here's the number that should unsettle you: the Magnificent Seven now represent 34.3% of the S&P 500. Up from 12.5% in 2016. One-third of America's benchmark index is seven companies.
Strip away tech and communication services from the S&P 500's 2025 returns, and the index returned 6%. Not 17.9%. Six percent. The "broad market rally" was a tech story wearing an index fund costume.
This isn't new. The same concentration dynamic played out in the late 1990s with the dot-com darlings. In the early 1970s with the Nifty Fifty. Both ended the same way: a violent repricing when markets remembered that concentration is a bet, not a strategy.
The Mag 7 accounted for 42% of the S&P 500's total return last year. Nvidia alone was responsible for 15%. One company, selling one product category (GPUs for AI training), generating one-seventh of the entire market's gains. That's not diversification. That's a leveraged bet with an index fund wrapper.
What Actually Drove 2025
The real engine was infrastructure spending. Hyperscalers — the Amazons and Microsofts and Googles of the world — poured an estimated $527 billion into AI infrastructure in 2025 alone. Goldman Sachs projects that figure to balloon to $1.4 trillion between 2025 and 2027.
That money flows downhill. Nvidia sells the chips. Broadcom (up 47% in 2025) builds the networking. Micron (up 229%) manufactures the memory. AMD (up 78%) picks up whatever Nvidia can't supply. The VanEck Semiconductor ETF returned 49% for the year, its third consecutive year of massive gains.
The semiconductor supply chain was 2025's real story. Not the Magnificent Seven, not AI chatbots, not whatever Sam Altman tweeted. Follow the capex, and you find the returns.
But capex is a leading indicator, not a guarantee. Every dollar spent on a data center is a bet that AI revenue will eventually justify the investment. That bet hasn't been proven yet.
Dan Ives, one of Wall Street's most vocal tech bulls, called 2026 a "prove it year" for AI. Translation: the market gave these companies the benefit of the doubt in 2025. Now it wants receipts.
The Rotation Is Already Happening
Six weeks into 2026, and the script is flipping.
Basic materials are up 9.05%. Industrials are up 5.2%. Energy is climbing. Small-cap value stocks have gained 5.94%. Small-cap growth is up 6.02%. These are the sectors that spent 2025 watching from the sidelines while Nvidia ate the market.
Tech? The worst-performing sector in 2026. Down 0.4%.
Large-cap growth, the category that houses most of the Magnificent Seven, has gained 0.13% year to date. Compare that to small-cap growth at 6.02%. That's a 46x performance gap. In six weeks.
Michael Arone, chief investment strategist at State Street, calls it a "powerful one-two punch" of better-than-expected economic growth and broadening earnings. The bull market is no longer a Mag 7 monopoly. It's spreading to the companies that actually make things, build things, and dig things out of the ground.
Small caps are trading at a forward P/E of 18x versus the S&P 500's 24x. That 22% discount is near a historic low. The last time the gap was this wide, the subsequent decade favored small caps by a significant margin.
The Historical Pattern
This rotation mirrors something we've seen before. International equities outperformed U.S. stocks in 2025 by the widest margin since 2009. South Korea's Kospi surged 76%. Germany's DAX, Japan's Topix, and the UK's FTSE 100 all posted gains above 20%.
The U.S. market's 16% gain looks modest by comparison. And that 16% was almost entirely driven by seven stocks that most investors can't influence or time.
Every major market cycle follows the same arc. A narrow group of stocks leads the charge. Valuations stretch. Concentration builds. Then the market broadens, and the laggards become the leaders.
The dot-com bubble didn't end because the internet was fake. It ended because valuations detached from cash flows. The same risk exists today with AI infrastructure spending. The technology is real. The revenue justification hasn't caught up.
Who Wins, Who Loses
The winners in 2026 are likely the companies that grew earnings in 2025 while their stock prices declined. High-quality small and mid-caps with manageable debt, pricing power to navigate tariffs, and businesses that don't depend on a single technology narrative.
The losers? Anyone who treated the S&P 500 as a diversified investment without realizing one-third of it was a concentrated tech bet. And anyone who assumes AI capex will translate into AI revenue on the timeline the market is pricing in.
Apple might be the most interesting play in the Mag 7 precisely because it was the most boring in 2025. Investors criticized its cautious AI strategy. But caution looks smart when the market starts questioning whether $527 billion in annual capex was wise. Apple has less to lose if AI sentiment sours, and room to run if it accelerates its AI roadmap.
Nvidia is the opposite proposition. If AI sentiment holds, it outperforms. If sentiment cracks, it has the furthest to fall. That's not an investment thesis. That's a coin flip at a premium price.
The Uncomfortable Truth
The 2025 "tech rally" was a story about two stocks, a semiconductor supply chain, and $527 billion in corporate spending that hasn't been justified by revenue yet.
The 2026 rotation is the market's way of hedging that bet. Money is flowing into sectors with tangible cash flows, reasonable valuations, and earnings growth that doesn't depend on whether AI becomes a $10 trillion industry or a $2 trillion one.
Information about the Mag 7's dominance is free. Everyone has the same charts, the same return data, the same analyst reports. Judgment about what happens next is what's scarce.
The conventional wisdom says tech always wins long-term. The crack is that "tech" in 2025 really meant "two stocks and a semiconductor supply chain." The insight is simpler than anyone wants to admit: concentration is risk, not conviction. And the market is starting to remember that.
Sources: CNBC, Morningstar, Statista, Visual Capitalist, Motley Fool, S&P Dow Jones Indices, State Street, Goldman Sachs, Oppenheimer