Author: Ray Dalio
Compiled by: Deep Tide TechFlow
Guide: Bridgewater founder Ray Dalio posted an investment note on X, doing the math on the current market dominated by a few AI giants. His judgment is tough: high risk is a fact, low returns are an opinion—the real returns on US stocks over the next 5 to 10 years could be in the range of -5% to -10%. He's not telling you not to buy AI; he's advising you not to put all your chips on AI. This is the 'Holy Grail of Investing' he has summarized over more than 50 years, and now he's sharing it publicly with everyone.
Investment Principles: How to Play the Hand You're Dealt
This note is about how to play the game of investing in the current situation.
You can think of it like bridge, poker, backgammon, or chess. It's your turn to move, and there's a computer next to you helping you assess the situation and make suggestions. For me, investing feels like that. Whether you have that computer or not, I think you should ask yourself one question: given how the cards are laid out now, what move should I make (i.e., what are the current market's characteristics and what forces are influencing it).
I've played this game for a long time. At this stage, my goal is to pass on my way of playing and, going a step further, to create a platform where all sorts of people can use it to explore the subject of investing, however they like—to learn, to look back at what they would have done, and to do it well. I believe there are right and wrong ways to play the hand you're dealt. So when you encounter a specific situation, you should ask yourself: 'How should I bet in this situation?' and be able to give a reliable answer.
Below, I want to talk about what the market looks like to me now, and what I think should be done (which is also what I am doing).
How to Play This Round Now
What are the most critical conditions today, and how should you bet on them?
In my view—and probably in everyone's view—we are currently in a market where a very small number of companies, concentrated in a sector with astonishing new technology (mostly AI), are dominating the direction of the entire market. These companies account for a high proportion of market capitalization and have a huge impact on the market and the economy. It's the same every time this happens: the new technology sector is filled with great excitement, uncertainty, and volatility, and these emotions spread to global stock markets. Therefore, the ups and downs and uncertainty of this sector are of great significance.
Beyond that, there are several other equally important major variables, what I call the 'Five Great Forces': 1) what's happening with debt and money, 2) what's happening with political and social issues (which significantly affect taxes and other politically-driven market factors), 3) the impact of geopolitics on markets (like those wars), 4) what's happening in the natural world, and 5) what's happening with new technologies. I feed these conditions into my investment system, and it calculates how to bet on them, while I'm also thinking about what to bet on.
When considering how to bet, the most important question to ask and answer clearly is: Do you want to a) bet more heavily on the new technology than the market index (like the S&P 500) already implies, overweighting this sector or the few companies you think are best; b) maintain a weight roughly similar to the index; or c) diversify away from this concentration?
Almost everyone wants to own the best assets and is desperately trying to do so, and this new technology in front of us seems to be changing almost everything. But history tells us that at this stage of the cycle, putting a high proportion of chips on the few leading companies producing this technology has failed for the vast majority of people. There's logic behind this, and it's played out this way every time in the past. AI is indeed a unique new technology, but there have been many unique new technologies in history that can serve as references. You should look at them. If you choose to ignore them, you need a good reason explaining why this time is different.
The Risk Is Indeed High
The stories of all great new technologies in the past have played out the same way, for the same reasons. High risk combined with great uncertainty is inherent to these new technology companies. Looking back at their performance in similar situations, you'll find that even revolutionary companies that ultimately succeeded in the long run (like Microsoft and Apple) got beaten to a pulp at similar junctures. And in the present moment of a new technology company's emergence (not in hindsight), it's simply not easy to tell who will succeed and who will fail; IBM is an example. Looking at all these cases laid out, you'll understand that high uncertainty about the future is the nature of new technology companies.
For example, they either invest too much or too little. The reason is: not investing enough guarantees failure, but they can't predict the future precisely, so they can't know if they're overinvesting. Both over- and under-investing come at a cost.
They also can't accurately foresee all the changes that will affect them, including exogenous ones—monetary tightening, war, dramatic tax changes. So they all go through huge ups and downs, first exciting investors, then terrifying the faint of heart and washing them out, thereby amplifying market volatility. Digging a layer deeper: these new technologies and companies, which disrupted their predecessors, will ultimately be disrupted by newer technologies and companies themselves, in ways that are unthinkable now. We have to consider whether the same thing could happen to today's companies. The impact of quantum computing is one of the 'known knowns'. What about those not yet imagined?
What about the risk from competitors? For example, China is producing and distributing AI technology, and Chinese policymakers have a completely different view of the economy and AI. We are in a new technology war, and leaders of each country believe they must win. From China's perspective, because AI has huge productivity dividends and can raise the overall standard of living, it should be provided free or at low cost to the public. In their view, profit is less important; the overall benefits from many people using these new technologies are what matter. I expect them to compete in the international market just as they have with products like cars, solar panels, and batteries.
The current situation resembles many historical moments that can give us lessons. I can't help but think of the late Dutch Empire and the early British Empire, the period when Britain surpassed the Netherlands in shipbuilding and other important industries. Also, geopolitical conflict surrounding Taiwan, which should at least make us consider the possibility: could China use 'preventing chips from leaving Taiwan' as a tool in geopolitical competition? AI stocks face other risks, such as the risk of wealth taxes and other taxes rising—which could force people who have a lot of wealth tied up in these stocks to sell; or rising anti-AI sentiment, which could impose restrictions on companies' expansion.
I could list a bunch of other things to worry about, and I could also list an equally long list of great AI opportunities I'd like to bet on. I'm not saying how these risks will turn out, nor am I saying you shouldn't buy AI companies. I'm just saying that there is a large amount of concentrated risk in the market—that's indisputable—and you should know how to play in such a situation. Based on my research of all similar cases, and out of logic, I am confident that the risk is high, and the best way to play this situation is:
Diversification Is Good
You probably know my mantra is diversification. My 'Holy Grail of Investing' is to try to hold 15 good, uncorrelated, risk-balanced positions. To put it another way:
"A well-diversified portfolio of good bets will outperform a concentrated bet (it has a higher return-to-risk ratio and can be engineered to produce better returns at the same level of risk). The more risk is concentrated in one part 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, assume a bet has a return-to-risk ratio of 0.3 (e.g., 6% return with 18% standard deviation, which is typically considered the level for stocks), and compare holding 5, 10, 15 uncorrelated bets: I can achieve the same 6% return, but the risk measured by standard deviation drops to 8%, 6%, and 5%, respectively. That is, 15 good uncorrelated investments can improve my return-to-risk ratio by 4.3 times (from 0.3 to 1.29). You can also add leverage if you wish, achieving much higher returns at the same level of risk. This is a fact.
My confidence comes from backtesting, from the real returns I've delivered over more than 50 years of investing, and from the largely solid logic: a well-diversified set of bets, adjusted to one's desired level of volatility, will produce much better returns over time than the concentrated bets most investors prefer. To be more specific, good diversification can give you a better risk-return ratio than any concentrated bet; adjusting it to your desired risk level allows you to achieve higher returns at that risk level than with any other approach.
Because I've made this method public, it has now become my 'no longer so secret' recipe for success. But I rarely encounter people who think about investment strategy this way—that is, few people think about portfolio construction, about how a well-structured, well-diversified portfolio of bets would perform compared to a concentrated position in a few stocks in a great transformative industry. Most people only think about whether these stocks, this industry, will go up, and how to bet. The performance gap between those who think about portfolio construction and those who don't is huge. I'll have a chance to explain this more fully later.
Based on all of the above, in my view, pondering how to play the current hand should lead one to ask oneself: How large should my concentrated positions be? And then, diversify.
Returns Appear Low
High risk is an indisputable fact. Next, I'll give you an opinion that could be wrong: expected returns are low. This judgment comes from my analysis of valuations and the readings of my bubble indicators—over the next 5 to 10 years, the real returns on stocks look to be about -5% to -10%, though there is considerable uncertainty around these numbers. In my view, these stocks are very long-duration assets with significant risk because it's difficult to reliably see the distant future, and they appear both expensive and in shaky hands.
A Question from My Research Team
At a recent meeting, someone on the team asked me: Why do you think the market's current configuration is wrong? How do you know that the lack of diversification in the market today isn't for good reasons—for example, some investors believe the expected returns on AI stocks are very high; for example, when an industry accounts for such a high proportion of market capitalization, such concentration in an index is natural; or, when there is extreme enthusiasm for an industry, many investors buy these stocks without doing the smart, reliable math to figure out what future earnings will be and how those earnings should be priced?
My Answer
Prices rise for all sorts of reasons, not all of them good. Some investors watch prices and push them up because they think prices are attractive relative to fundamentals; some hold these stocks because they are convinced it's a great new technology and take rising prices as confirmation that 'these are good stocks'; and other investors hold index exposure, passively placing heavy weight on these stocks. In my opinion, you can agonize over these issues, trying to figure out what to do; or you can admit that you don't need to agonize at all because you simply don't have enough information to bet confidently. You can perfectly well say, 'I don't know enough, I won't take this bet.' And then actually not bet.
What gets people into trouble is the thought that 'I must have an opinion, and my opinion is worth something,' when the reality is more likely that you simply can't form a view reliable enough to bet on. (Note: To be clear, I'm not recommending not betting—you can't avoid betting anyway because you always have to put your money into some investment, or into cash, and most people think cash has the lowest risk, when it's actually the worst long-term investment. I recommend understanding how to diversify bets well, even if you have no tactical view on which markets are good or bad. The approach is: when you have no confident tactical judgments, hold a balanced, strategic asset allocation portfolio. But that's a topic for later.)
So I believe that knowing what you don't know, and thus deciding when not to bet, is as important as knowing what you do know and betting accordingly.
To put it more simply, I believe in this principle: since it's usually difficult to have enough information to justify concentrated bets, the best approach is to form a diversified portfolio consisting only of your most confident, uncorrelated bets, and then engineer that portfolio to your desired risk level. That's my 'Holy Grail of Investing.'
Right now, facing the current hand, I don't think anyone can know clearly enough what will happen next in this technology-driven market to make a large, concentrated bet. In my view, avoiding concentration and staying diversified is the best way to deal with this 'not knowing.' I know this contradicts the theories you read in textbooks—textbooks basically say markets are efficient, so you should 'just believe in the market.'
To summarize: We currently have an unusually concentrated market centered around a revolutionary new technology, which should precisely remind us not to confuse excitement about the new technology with the attractiveness of new technology stocks, and then throw caution to the wind to hold a bunch of high-risk, highly correlated concentrated bets—especially when we could achieve similarly attractive returns at much lower risk through smart diversification.
P.S. I won't share my holdings or tactical judgments with you, because I don't want to be your investment advisor. But I will soon share some key perspectives behind these judgments, including the readings and logic behind my bubble indicators.
