Magnificent moats?

Earlier this year, the world’s biggest sovereign wealth fund – Norway’s NBIM – posted its highest quarterly return since its creation in 1996, a result it attributed mainly to its vast holdings in Mag7¹ companies. They were not the only ones; with Mag7 stocks hugely outperforming – and carrying – broader indexes in 2023 off the back of rampant artificial intelligence (AI) optimism, current valuations seem to imply almost boundless faith in the unassailability of their economic moats.

Indeed, viewed through the prism of economies of scale and network effects, there is even a case for arguing the Mag7 are cheap. A recent study found that “the top-performing 2.4% of firms account for all of the USD 75.7 trillion in net global stock market wealth creation from 1990 to December 2020.”² By some estimates, such disproportionate market capitalizations mean “the economic profit of the Magnificent Seven is around 40% greater than the aggregate economic profit for the Russell 3000” as professor and strategist Michael Mauboussin told the Financial Times back in February.³

The performance of these companies reflects well known – and colossal – competitive advantages, which many observers deem practically insuperable, at least for the foreseeable future. As Mike Nell, Senior Equity Analyst and Portfolio Manager at UBS Asset Management, puts it, “in our model, we’ve baked in a 35% return on equity at terminal value (or very near) for all these companies, which is sort of an act of faith that they won’t lose their moat.”

Such forecasts are not made lightly. Each member of the Mag7 cohort has its own distinct variety of moat, enabling them all to create enormous value – and profits.

Notwithstanding Microsoft’s big ambitions for its rival search engine, Bing,⁴ Google’s 90% share in global search is so dominant that many consumers barely realize alternatives exist anymore. Embeddedness in business infrastructure make its removal seem virtually unthinkable for most companies, while Apple’s brand strength exerts a hold over its customers that can border on cultish.⁵

In many cases, Mag7 dominance is so entrenched that experienced analysts find it impossible to imagine where meaningful competitors will emerge. “You can't see anything that would supplant, say, the iPhone,” says Nell. “And I certainly don't think the market is going to assume it's been supplanted until it can see the whites of the eyes of the thing that is displacing it. It's really hard to imagine what that would be; Apple’s ecosystem is so sticky that it's unlikely to be upset in the foreseeable future.”

‘Foreseeable’ carries a lot of weight in that verdict, of course. Like all great empires before them, the Mag7’s aura of immutability is only that: an aura. As all investors know, nothing lasts forever. On a long enough time horizon, all champions are supplanted. Angus Muirhead, Head of Thematic Equities at UBS Asset Management, cautions that “nearly all of the biggest companies are publicly listed, and the bigger they get, the more successful they get, the more significant they are in these indexes. But look back over 10-year intervals, and the largest companies have changed. Only two or three tend to survive the next ten years. But we can never imagine that at the time.”

Gill posits that “The big question here is whether AI is a sustaining or disruptive innovation as per Clayton Christiansen’s framework. If it proves to be the former, as many believe, it should reinforce the moats.

AI: levelling the playing field, or tilting it?

Since many believe it to be the most disruptive technology in a generation, might AI be the catalyst to rattle the Mag7 Titans?

It is not hard to find voices that think so, even inside the walled gardens of the Mag7 itself. A leaked 2023 internal memo from Google fretted that open-source models were eating the company’s lunch, and that by comparison with their own AI offerings “open-source models are faster, more customizable, more private, and pound-for-pound more capable.”⁶ The conclusion of the memo was frank: “we have no moat.” Other voices have raised similar concerns, with plenty of commentators cautioning against AI ‘hype’ around the Mag7.

At the same time, it is just as easy to find those who think AI’s high-tech, capital-intensive, talent-scarce nature is more likely to compound existing Mag7 advantages than erode them. “There's a lot of slack in the R&D budgets of these big companies," says Nell. “They can spend USD 100 million, and it's meaningless to them. But if you're a smaller company, trying to spend USD 10 million extra might be a reach.” All the Mag7 companies enjoy R&D budgets in the billions of dollars, substantial portions of which have been devoted to AI in recent years, including strategic acquisitions such as Microsoft’s of OpenAI, Google’s of DeepMind, and Apple’s of Xnor.ai.

But deep pockets aren’t the only advantage enjoyed by the Mag7, many of whom have data-driven business models that stand to benefit immensely from AI. Despite the ‘no moat’ memo, for instance, Google’s dominance in search and cloud computing alongside its vast data accumulation give it a rock-solid foundation for AI tools to polish its core search and advertising functions.

Amazon too has vast quantities of consumer data, which its AI can use to sharpen its preeminent logistics and marketing capabilities still further, while Facebook’s ad targeting and content personalization are also set to benefit.

China analyst Jia Tan, Head of Research, China Equity Long/Short, UBS O’Connor, points out that Nvidia’s stratospheric recent performance is premised heavily on AI’s rapidly manifesting potential, and that Microsoft’s investment in AI Copilot is strengthening its moat too. Meanwhile, the data lake Tesla has accumulated from millions of self-driven miles is likely to prove a huge asset in the ongoing race for fully autonomous vehicle networks, deepening its advantage over legacy US and European competitors:

“What's happening in that market is a transition from a product that was largely mechanical, to something that's basically a large smartphone on wheels,” says Nell. “Most car manufacturers weren't born in Silicon Valley, which means that their software expertise is minimal. Detroit is not Silicon Valley. Germany is not Silicon Valley. These companies are trying to take on Tesla, which is just a completely different animal.”

Acknowledging the power of such advantages, Jia Tan is dubious of any imminent breach of the Mag7 moats: “investors have talked about a potential threat to Google’s search function from Bing and OpenAI, but so far it hasn’t materialized. It’s ultimately not possible to forecast where the true threat will come from; all we can do is try to track all those disruptive innovations closely.

Acknowledging the power of such advantages, Jia Tan is dubious of any imminent breach of the Mag7 moats: “investors have talked about a potential threat to Google’s search function from Bing and OpenAI, but so far it hasn’t materialized. It’s ultimately not possible to forecast where the true threat will come from; all we can do is try to track all those disruptive innovations closely.”

Far from threatening the Mag7, AI might turn out to add yet more width to their moats. Albert Tsuei, Lead Portfolio Manager for UBS Asset Management’s digital transformation equity strategy, points out that investors have historically misjudged the value of big tech stocks at moments of paradigmatic technological change: “investors have often underestimated the sheer power of these big tech business models when they are at scale. Historically, many of these companies turned out to be significantly cheaper than they should have been as they started to flex their economic power.” A point backed up by the stock market wealth creation study mentioned earlier.⁷

Regulation on the cards

Whatever AI’s effects on Mag7 business models, it is already fomenting a major discussion about regulatory intervention. Summits on this theme are coming thick and fast, and both the European Parliament and UN adopted major AI resolutions in March of this year.⁸

Historically, regulatory activity has sometimes functioned as a drawbridge to competitors looking to storm incumbents’ moats, and the conversation around AI could serve to embolden the many ongoing regulatory interventions against big tech. Google and the US Justice Department are already embroiled in the most significant antitrust case in years, with many lawmakers on both sides of the Atlantic eager to see the company – and others in the Mag7 – broken up.⁹ Apple and Amazon too are being sued by the US government, and accused by the Justice Department and Federal Trade Commission, respectively, of monopolistic practises.¹⁰ Meanwhile, Facebook has faced scrutiny from a number of inquisitors over the same issue – as well as data privacy, free speech, and content moderation.

But again, regulation can just as often magnify pre existing commercial advantages as diminish them, since it is generally easier for bigger companies to absorb the compliance costs and friction it imposes. Last year’s Senate hearing, in which OpenAI CEO Sam Altman appeared to plead for greater regulation for the AI sector, provoked competitors to fret publicly about unintended consequences, with Stability AI’s Emad Mostaque telling reporters that regulation invariably favors incumbents and can stifle innovation, and Clem Delangue, CEO of Hugging Face, tweeting that: “Requiring a license to train models would be like requiring a license to write code. IMO, it would further concentrate power in the hands of a few & drastically slow down progress, fairness & transparency.”¹¹

Since most of the Mag7 place a significant premium on public trust, they may well benefit from the reassurance regulation can provide consumers as they learn to engage with AI, especially in contentious use cases such as fully self-driven vehicles or medical diagnostics. In addition, the Mag7 have powerful public affairs presences that are accustomed to lobbying the world’s major legislatures and governments, and may be better positioned than most to shape any nascent national or global regulatory framework to their advantage.

Mean reversion

The width of Mag7 moats is, of course, a separate question from their value. Today, market concentration at the top of the index looks as precarious as it has in a generation. While most investors expect these companies to maintain their dominant position, it may be that the bigger opportunities now lie elsewhere in the market.

“It’s very difficult for these stocks to outperform,” Nell explains, “because the thing they're trying to beat is themselves. If you take Apple and Microsoft together, they're almost 50% of the benchmark in the IT sector. So, if you want to own stocks that are going to perform in line with the IT sector benchmark going forward, it would seem logical you'd have to look elsewhere. It's not a question of whether their moats are sustainable, but whether their valuations are excessive relative to the other opportunities within tech.”

So far this year, there has already been some divergence in fortunes among the Mag7, with Tesla falling 28% and Apple posting a relatively modest 10%.¹² Naturally, many analysts are talking about a mean reversion. “I believe that the mega-cap stocks are fully valued, and there's a mean reversion trade coming,” says Nell. “That could come in one of two forms: either the mid-cap, and smaller-cap stocks catch up with the mega-cap stocks, or the mega-cap stock valuations decline to match those of the small-cap and mid-cap stocks where they are today. Personally, I think it's more likely the former.”

The price of success: vigilance

The Mag7’s relative immunity to competition stems from powerful elements of each company’s DNA and will not be simple for tomorrow’s challengers to overcome. But the same was true of previous corporate giants, and history has shown that such advantages cannot be sustained forever.

Recent results have been impressive, but the AI hype cycle is young and it is still hard to determine where AI’s impacts will be felt most profoundly. With regulators watching carefully and more disruptive technologies emerging every year, the Mag7 will need to do all they can to justify today’s sky-high valuations as competitors continue to test their defences.

¹ The Magnificent Seven (Mag7) is the recently minted collective noun for Amazon, Apple, Alphabet, Microsoft, Meta Platforms, Nvidia, and Tesla.
² Long-Term Shareholder Returns: Evidence from 64,000 Global Stocks, Financial Analysts Journal, Vol 79, 2023.
³ Michael Mauboussin on increasing returns to scale, Financial Times, Feb 23, 2024.
⁴ In an interview with the Wall Street Journal in August 2023, Satya Nadella said that “Bing with AI is gonna completely change what people can expect from search”. Earlier in the year he told CBS that “we are looking forward to the new search wars.” (CBS, Feb 7, 2023)
⁵ Using an MRI scanner on an ‘Apple fanatic’, a 2011 BBC documentary found that images of Apple gadgets triggered responses in the same parts of the brain as images of a deity do for religious people; https://www.bbc.co.uk/news/business-13416598
⁶ https://www.semianalysis.com/p/google-we-have-no-moat-and-neither; https://www.theverge.com/2023/7/10/23790132/google-memo-moat-ai-leak-demis-hassabis
⁷ Long-Term Shareholder Returns: Evidence from 64,000 Global Stocks, Financial Analysts Journal, Vol 79, 2023.
⁸ Artificial Intelligence Act: MEPs adopt landmark law , European Parliament, Mar 13, 2024; General Assembly adopts landmark resolution on artificial intelligence, United Nations, Mar 21, 2024
⁹ Google, Apple breakups on the agenda as global regulators target tech, Reuters, Mar 25, 2024
¹⁰ Justice Department Sues Apple for Monopolizing Smartphone Markets, Office of Public Affairs, U.S. Justice Department, Mar 21, 2024
¹¹ @ClementDelangue, Twitter, May 17, 2023.
¹² Advisors weigh in on the future of the Magnificent Seven, Investment News, April 29, 2024