The ‘Magnificent 7’ moniker, once shorthand for a group of big‑tech stocks that moved in lockstep, is starting to fracture. Investors who have relied on the collective strength of Apple, Microsoft, Alphabet, Amazon, Nvidia, Meta and Tesla are now facing a far more complicated picture. Neuberger Berman co‑chief investment officer Jeff Blazek puts it bluntly: “The basket is beginning to break at the same time more granular AI‑related equity stories are gathering momentum.” The clearest sign of this splintering lies in the numbers. In 2023 the seven stocks returned 76% as a group, dwarfing the S&P 500’s 24% gain, according to investment manager Mellon. As of May 2026 their year‑to‑date performance of 8.8% still edges the broader index’s 8.1%, but that headline masks a deep internal divergence.
The great divergence
The average pairwise correlation – the rate at which any two of the seven stocks move together – has collapsed from 75% in their 2023 heyday to just 25% today, the lowest level since 2019, Neuberger Berman notes. In the space of a few months the gap between the best and worst performer has become “striking”. Year‑to‑date returns as of mid‑May 2026 range from Alphabet’s 23% gain to Tesla’s 15% loss. Even the expected leaders have swapped places. “Consider Alphabet and Microsoft specifically: a year ago, the former was written off as lagging badly on AI, while the latter was deemed a consensus winner. That read has inverted – starkly,” Blazek said. Alphabet, Amazon and Meta all delivered first‑quarter results that exceeded analyst expectations, but the wider group’s once‑cohesive performance “may have run its course”, he added. “The story of seven stocks moving in unison to dominate large‑cap indices may be reaching its end.”
AI’s uneven footprint
All seven remain heavily exposed to the artificial‑intelligence boom, and appetite for AI shows no sign of cooling. Helen Jewell, international chief investment officer of fundamental equities at BlackRock, said: “AI capital expenditure is now $725 billion for 2026 and we expect that to reach $6 trillion by 2030 – an extraordinary amount of money is being spent on AI.” Yet the expected divergence between the companies, Jewell argues, will be determined by the breadth of their offering across what she calls the “AI stack” – the layers of computing, cloud infrastructure, models and applications. “A company like Alphabet has more elements across that whole AI stack – computing, cloud, models, applications. Others are more exposed to the parts of the stack seen to be less strong. Software, for example, is seen as a weaker part of the AI story because there’s a feeling that AI will be able to replicate a lot of what software does,” she explained.
The different strategies illustrate why the Mag 7 are now marching to different beats. Nvidia positions itself as an AI infrastructure company, building a foundational platform with its CUDA and Omniverse ecosystems and controlling critical interconnect technology. Microsoft made an early, decisive bet on generative AI through its partnership with OpenAI, embedding it across its product suite. Amazon integrates AI across virtually every operation – from AWS managed services to logistics and fraud detection – aiming for practical AI at scale. Meta is shifting towards what it calls “personal superintelligence”, focusing on empowering individuals with AI tools while leveraging its vast user data; it treats access to compute as a key competitive moat, driving significant chip deals. Apple adopts a deliberate, privacy‑focused approach, emphasising on‑device processing and selective acquisitions rather than building massive public‑facing models. Tesla, true to its “AI‑first” philosophy, applies the technology across autonomous driving, robotaxis, manufacturing and energy management. Jewell noted that the divergence already emerging will only become more pronounced, making the Mag 7 label increasingly outdated: “While the term was useful in describing their shared characteristics, it will become outdated as the divergence that’s already started to emerge continues.”
Where that leaves investors
For all the talk of a split, investors have been warned against trying to time a theme. James Norton, head of retirement and investments at Vanguard Europe, points out that people described Mag 7 valuations as stretched for years – and anyone who sold out on that narrative missed a huge run. “Clearly this year’s performance to date has been more mixed. As it’s hard to predict future share price trajectories, taking a diversified, long‑term approach is more sensible than trying to time a specific theme,” he said. Jewell is similarly wary of calling any of the stocks “expensive”. “Ultimately the multiple reflects what people think the future earnings growth is going to be. The reason Alphabet is demanding a higher price is because there’s a feeling that its earnings growth is going to be strong because they’ve got so many different parts of the AI story,” she said. She also stressed that the performance divergence is not a static story. These companies share an unusual ability to reinvest themselves, thanks to their ambition and high cash flows. “As the oldest of the Mag 7, Microsoft is a phenomenal company that has gone through many iterations. It reinvents itself continually – as does Apple – because they’ve got the cash to rethink and recreate what they do, which is what allows them to sustain for the long term.”
Norton’s advice for retail investors is to zoom out. “For investors with a diversified portfolio, these companies are just one part of the picture. It is true that historically quite a small number of companies have driven a large share of overall market returns. But in a diversified portfolio, the Mag 7 are just a part of the returns of the US market, which are in turn part of the returns from the global market, which are again balanced by the returns you are also getting from bonds.”
