Everyone knows that the stock market’s super-run in recent years has revolved around seven core stocks: The “Magnificent Seven” of Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia, and Tesla (in alphabetical order).
Most of the seven, if not all, are assumed to be the beneficiaries of the stupendous growth of artificial intelligence. The “Magnificent Seven” have a combined market value of more than $17 trillion, and accounted for over half of the S&P 500’s total return in 2024. Over the last five years, all seven have seen annualized appreciation rates of 20 percent or more, with Tesla hitting 73 percent and Nvidia exceeding 90 percent.
Is this sustainable? How does it all end? And what does it say about the AI boom that plays a pivotal role in the seven’s success?
Two years ago, I published a Forbes column about Sam Bankman-Fried, FTX, and the bitcoin mania. Back then, the price of a single bitcoin hovered around $20,000, while it hovers around $100,000 today.
Then, like now, a key question was being asked: Is bitcoin a boom, a bubble, or a “con”?
The Wall Street Journal’s Vicky Ge Huang recently compiled the warnings that some of the world’s great investors gave the public about cryptocurrencies like bitcoin, and yet the $100,000 valuation suggests we’re not listening. JPMorgan Chase CEO Jamie Dimon has called bitcoin a useless “pet rock,” while Citadel CEO Ken Griffin wonders aloud, “What problem does it solve for our economy?”
But what about AI? Is artificial intelligence something that the truly sophisticated suspect to be another bubble, and we’re ignoring that risk? Or is AI an actual boom, and the question is how and when the “landing” comes?
My conclusion is the latter. AI is a real investment in something that even less experienced investors can understand where big-league returns may arise. From medical applications to breakthroughs in content creation or computer science, AI is already revolutionizing the world of work, and there’s no reason for the innovation to stop now.
And yet, will the returns be big enough or likely enough to warrant the kind of investment it now takes to “get in the game”? That’s where I get more skeptical. There is a long list of things that have to go right for the value creation of the magnitude assumed by today’s AI pricings to materialize:
- There has to be a solution to the power consumption problem. Even today’s AI’s energy demand is well beyond America’s existing generation capacity and grid strength. Yet, Americans are struggling to invest in either. So who will? Those who can afford to build and run their own proprietary nuclear generation plants?
- There must be a solution to the data training problem. The level of data volume that fourth- and fifth-generation AI models need currently exceeds everything on the internet today. And, as we add new sources of data, the energy requirements continue to grow exponentially.
- What is the “killer app” that becomes a must-have? What people want to purchase will ultimately provide returns on the investments in generation capacity and grid strength, but the proverbial cash cow for companies remains to be seen. Consumer access to enhanced search options is a given, but the greatest AI applications won’t have to do with generating eyeballs or page views to sell advertising. The automation of data centers, coding jobs, and repetitive tasks are likelier targets, but the income has to come from the companies whose cost structures are being improved by AI.
- The clamor for government regulation will only get stronger as AI displaces more jobs, but the great unknown is the impact of intellectual property laws on the use of trademarked and copyrighted text, images, likenesses, and other forms of content. That outcome will be decided in the legal and political spaces, where regulatory approaches come with their own unintended consequences for AI research and development.
One thing is clear: AI is a very real and transformative technology, but scale will be integral to addressing such problems. The market will only want—and can only afford—four or five surviving players. As product competition leads to commoditization, competitive pricing and the ability to “stay current” will be the keys to survival.
As always, a select few investors will make enormous sums of money either by riding the wave and getting out of the water in time, or picking one of the ultimate survivors and hanging on for the ride. Countless start-ups and venture capital firms will see 100 percent losses as they stall out on scale or fall outside the cone of competitive capability.
However, all of the evidence suggests AI is neither a bubble nor a con; it is a boom with lasting potential. While there is plenty of irrational exuberance related to AI, that is the free market’s way of experimenting and learning, which have always given the U.S. economy its edge. AI’s biggest winners will have solutions that address real problems and create new forms of competitive advantage, and we should be ready for those benefits to materialize in the coming years. On the other hand, hindsight will become 20-20 for the losers who assumed it was a “sure thing.”
When only a few numbers will come up winners, I don’t see anyone winning by placing bets across the board. AI will be a long-term winner for a select—and potentially magnificent—few who successfully sell their wares with scale to customers who happily pay for the cost and productivity gains they will create.