Ray Dalio Warns of AI Market Concentration and Negative U.S. Stock Returns Over 5–10 Years

Author: Ray Dalio

DeepChain TechFlow

DeepInsight Summary: Ray Dalio, founder of Bridgewater, posted an investment note on X, crunching the numbers on a market currently dominated by a few AI giants. His assessment is blunt: high risk is a fact; low returns are a projection—U.S. stock market real returns over the next 5 to 10 years could range from -5% to -10%. He’s not telling you not to invest in AI; he’s advising you not to bet everything on it. This is the “holy grail” of investing he’s distilled over 50 years, now shared openly with everyone.

This note explains how to play the investment game under the current circumstances.

You can think of it like bridge, poker, backgammon, or chess. It’s your turn to play, and beside you is a computer helping you assess the situation and offering suggestions. That’s how investing feels to me. Whether or not you have this computer at your side, you should always ask yourself: Given the current layout of the cards, what’s my best move here? (In other words, what are the current market conditions and what forces are influencing them?)

I’ve played this game for a long time. At this stage, my goal is to pass on my approach—and take the next step by building a platform that allows anyone to explore investing however they like—learning, reflecting on how they would have acted in the past, and improving upon it. I believe there are right and wrong ways to play the hand you’re dealt. So when you face a specific situation, ask yourself: “How should I bet in this scenario?”—and be able to provide a solid answer.

Below, I’d like to share my perspective on what the market looks like right now and what I believe we should do—which is also what I’m currently doing.

What are the most important conditions today, and how should you bet on them?

In my view—and likely in the view of everyone else—we are currently in a market where a very small number of companies, concentrated in a sector powered by remarkable new technologies (mostly AI), are driving the entire market’s direction. These companies account for a significant portion of market capitalization and exert a substantial influence on the market and the broader economy. As always during such times, this emerging technology sector is filled with excitement, uncertainty, and volatility, which ripple through global stock markets. Therefore, the fluctuations and uncertainties in this sector are of critical importance.

Besides that, there are several other equally important major variables—what I call the "Five Forces": first, what is happening with debt and currency; second, what is happening with political and social issues (which significantly impact taxes and other politically driven market factors); third, how geopolitics affects the market (such as wars); fourth, what is happening in nature; and fifth, what is happening with new technologies. I feed these conditions into my investment system, which calculates how to position myself under these conditions, while I also independently consider what bets to make.

When considering how to place your bet, the most important question to ask—and answer clearly—is: Do you want to a) overweight new technologies, this sector, or the few companies you believe are best, beyond what’s already implied by broad market indices like the S&P 500; b) maintain weights similar to the index; or c) diversify away from this concentration?

Almost everyone wants to buy the best assets and is desperately trying to do so, and this new technology appears to be changing nearly everything. But history tells us that, at this stage of the cycle, the vast majority of people fail when they bet heavily on the few leading companies producing such technology. There’s logic behind this—this pattern has played out every time before. AI is indeed a unique new technology, but history has also seen many other unique, comparable innovations. You should study them. If you choose to ignore them, you must provide a compelling reason why this time is different.

The story of every great new technology in the past has unfolded in the same way, driven by the same logic. High risk and immense uncertainty are inherent traits of these new technology companies. Looking back at how they performed under similar circumstances, you’ll find that even revolutionary companies that ultimately succeeded in the long term—like Microsoft and Apple—were nearly destroyed at moments like these. And at the very moment when new technology companies first emerge (not in hindsight), it’s extremely difficult to tell who will succeed and who will fail—IBM is a prime example. When you examine all these cases together, you’ll understand that the future of new technology companies is highly uncertain—that’s simply their nature.

For example, they either bet too much or too little. The reason is: betting too little guarantees a loss, but they can’t precisely predict the future, so they can’t know if they’ve bet too much. Both overbetting and underbetting come at a cost.

They also cannot accurately predict all changes that affect them, including exogenous ones—monetary tightening, wars, drastic tax shifts. As a result, they all experience wild swings, first exciting investors, then terrifying the timid and washing them out, thereby amplifying market volatility. Digging deeper: these new technologies and companies once disrupted their predecessors, yet most will eventually be disrupted themselves by even newer technologies and companies—in ways we cannot currently imagine. We must consider whether the same fate could befall today’s companies. The impact of quantum computing is one “known known.” But what about those yet to be imagined?

What about the risks posed by competitors? For example, China is currently producing and distributing AI technology, and its policymakers hold entirely different views on the economy and AI. We are now in the midst of a new technology war, where leaders around the world believe they must win. From China’s perspective, AI should be made freely or cheaply available to everyone because of its massive productivity gains and its potential to raise overall living standards. In their view, profit is less important than the broad societal benefits that come from widespread adoption of these new technologies. I expect they will enter international markets with the same aggressive strategy they used for automobiles, solar panels, and batteries.

This situation resembles many historical moments that have offered us lessons. I can’t help but think of the late Dutch Empire and the early British Empire, when Britain surpassed the Netherlands in shipbuilding and other key industries. The geopolitical tensions surrounding Taiwan should also make us consider one possibility: Could China use “preventing chips from leaving Taiwan” as a tool in geopolitical strategy? AI stocks carry other risks as well—such as the risk of rising wealth taxes and other levies, which could force investors with heavy exposure to these stocks to sell; or the growing anti-AI sentiment, which might impose restrictions on company expansion.

I could list a long roster of legitimate concerns, and just as long a list of AI opportunities I’d bet on. I’m not saying these risks will play out in any particular way, nor am I saying you shouldn’t buy AI companies. I’m simply saying that the market is undeniably loaded with concentrated risks—and you need to know how to play it. Based on my research into similar cases and sound logic, I’m confident: the risks are high, and the best way to navigate this situation is:

As you likely know, my mantra is diversification—my "investment holy grail" is to hold 15 well-balanced, uncorrelated positions. In other words:

A well-diversified portfolio of good bets will outperform a concentrated bet (it offers a higher return-to-risk ratio and can be engineered to deliver better returns at the same level of risk). The more concentrated your exposure to a particular segment of the market, the more you should diversify—especially when the market is driven by a revolutionary new technology that inherently brings great uncertainty.

This is not an opinion—it’s a mathematical certainty. For example, suppose a single bet has a reward-to-risk ratio of 0.3 (such as a 6% return with an 18% standard deviation, typical of stocks). Now compare holding 5, 10, or 15 uncorrelated bets: I still achieve the same 6% return, but the risk, measured by standard deviation, drops to 8%, 6%, and 5% respectively. In other words, 15 well-chosen, uncorrelated investments can increase my reward-to-risk ratio by 4.3 times—from 0.3 to 1.29. If you wish, you can even apply leverage to achieve significantly higher returns at the same level of risk. This is a fact.

My confidence stems from backtesting, the actual returns I’ve delivered over my more than 50 years of investing, and a highly probable logic: by properly diversifying your bets and then adjusting them to your desired level of volatility, the long-term returns will far exceed those achieved by most investors’ preferred concentrated bets. More specifically, effective diversification delivers a superior risk-reward ratio than any concentrated bet; and when you adjust it to your target risk level, you can achieve higher returns at that risk level than any other approach.

Because I’ve made this approach public, it’s now my “less secret” path to success. But I rarely encounter people who think about investment strategies this way—meaning, very few consider portfolio construction: how a well-structured, adequately diversified set of bets might perform compared to a concentrated position in just a few stocks within a single transformative industry. Most people only ask whether these stocks or this industry will rise, and how to bet on them. The performance gap between those who think about portfolio construction and those who don’t is enormous. I’ll share a more complete explanation of how to do this properly at another opportunity.

Given all of the above, in my view, contemplating how to play this hand should lead one to ask: How large should my concentrated position be, and then diversify.

It is undeniable that the risk is high. Here’s my next point, which may be wrong: expected returns are low. This assessment comes from my valuation analysis and my bubble indicators—the actual stock returns over the next 5 to 10 years appear to be around -5% to -10%, though there is considerable uncertainty around these figures. In my view, these stocks are long-duration assets with high risk, because it is difficult to reliably see far into the future, and they appear both expensive and held by unstable hands.

At the most recent meeting, someone on the team asked me: What gives you the basis to think the market’s current allocation is wrong? How do you know that the lack of diversification in today’s market isn’t justified—for instance, because some investors believe AI stocks offer extremely high expected returns? Or because it’s natural for an index to exhibit such concentration when an industry accounts for such a large share of market capitalization? Or because, when there’s intense enthusiasm for a particular sector, many investors simply buy into those stocks without carefully and reliably calculating future earnings or determining how those earnings should be priced?

Price increases have many different causes, not all of them positive. Some investors watch the price and push it higher because they believe it’s attractive relative to fundamentals; others hold these stocks because they believe they’re investing in a groundbreaking new technology and view the rising price as confirmation that “this is a good stock”;还有一些投资者通过指数敞口被动地 overweight these stocks. In my view, you can wrestle with these issues and try to figure out exactly what you want to do; or you can acknowledge that you don’t need to wrestle with them at all—you simply don’t have enough information to make a confident bet. You can honestly say, “I don’t understand this well enough to place this bet,” and then truly walk away.

What gets people into trouble is the belief that “I must form an opinion, and my opinion has value.” But the reality is far more likely that you simply can’t form a sufficiently reliable, worth-betting-upon opinion. (To be clear: I’m not suggesting you avoid betting altogether—you can’t avoid it anyway, since you always have to put your money into some kind of investment or hold cash, and most people wrongly assume cash is the lowest-risk option, when in fact it’s often the worst long-term investment. What I’m suggesting is that even if you have no tactical judgment about which markets are good or bad, you should still know how to diversify your bets properly. The approach is: when you have no confident tactical insights, maintain a balanced, strategic asset allocation. But that’s a topic for another time.)

So I believe that knowing what you don’t know—and therefore deciding when not to bet—is just as important as knowing what you do know—and therefore placing a bet.

To put it simply, I follow this principle: since it’s usually hard to gather enough information to justify putting all your money into one bet, the best approach is to create a diversified portfolio of your most confident, uncorrelated bets, and then engineer that portfolio to your desired risk level. This is my “investment holy grail.”

At this moment, faced with the current market conditions, I don’t believe anyone can be clear enough about what’s coming next in this technology-driven market to place a large, concentrated bet. In my view, avoiding concentration and maintaining diversification is the best way to respond to this uncertainty. I know this contradicts the theories you read in textbooks—textbooks that generally assert markets are efficient, so you should simply “trust the market.”

In summary: We currently have a market that is excessively concentrated around a revolutionary new technology—this should remind us not to confuse excitement about a new technology with the appeal of its stocks, nor to abandon caution in favor of holding a portfolio of high-risk, highly correlated bets—especially when we could achieve similarly attractive returns with much lower risk through smart diversification.

P.S. I won’t share my positions or tactical judgments because I don’t want to be your investment advisor. But I’ll soon share some key perspectives behind these judgments, including my bubble indicator readings and the logic behind them.