Author: Ray Dalio

Translation: Shenchao TechFlow

Shenchao Guide: Ray Dalio, founder of Bridgewater, published an investment note on X, assessing the current market dominated by a few AI giants. His judgment is firm: the risk is high, which is a fact, and the return is low, which is an opinion—real returns on US stocks over the next 5 to 10 years could fall between -5% and -10%. He does not advise you not to buy AI; he advises you not to place all your chips on AI. This is his "holy grail" of investing, summarized from over 50 years of experience, now shared publicly with everyone.

Investment Principles: How to Play This Hand

This note discusses how to play the investment game in the current situation.

You can imagine it like bridge, poker, backgammon, or chess. When it’s your turn to act, there’s a computer beside you helping you assess the situation and give advice. For me, investing feels like this. Regardless of whether you have this computer at hand, I think you should ask yourself one question: given the cards on the table, how should I proceed (in other words, what characteristics does the market currently have, and what forces are affecting it).

I have played this game for a long time. At this stage, my goal is to pass down my strategies, and going further, to build a platform where various people can explore investment topics, in any way they choose—learning, reflecting on how they might have acted, and doing it well. I believe there is a right and wrong way to handle the hand you have. So when you encounter a specific situation, you should ask yourself: "How should I bet in this situation?" and you need to provide a reliable answer.

Next, I want to discuss what I think the market looks like now and what I believe should be done (which I am currently doing).

How to Play This Hand Now

What are the key conditions today, and how should we bet on them?

In my view—likely shared by everyone—we are currently in a market dominated by a tiny number of companies concentrated in a sector with astonishing new technologies (mostly AI), which steer the market's direction. These companies account for a high proportion of market value and have a huge impact on the market and the economy. Whenever this happens, the new tech sector gathers immense excitement, uncertainty, and volatility, which then transmits to the global stock market. Therefore, the ups and downs and uncertainties of this sector are significant.

Besides this, there are several other equally important large variables, which I refer to as the "Five Great Forces": first, what is happening with debt and currency; second, what is happening with political and social issues (which can greatly impact taxation and other politically driven market factors); third, the impact of geopolitics on 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 bet based on these conditions, while I also consider what to bet on myself.

When considering how to bet, the most important question to ask and answer clearly is: do you want to a) bet on new technologies with a greater weight than what is implied by the market index (such as the S&P 500), overweighting this sector or the few companies you believe are the best; b) maintain a weight similar to the index; or c) diversify away from this concentration?

Nearly everyone wants to buy the best assets and is striving to do so, while this new technology seems to be changing almost everything. But history tells us that at this stage of the cycle, betting a high proportion on a few leading companies that produce this technology has led to failure for the vast majority. There is logic behind this, as it has played out like this every time before. AI is indeed unique, but many unique new technologies in history can serve as comparisons. You should look at them. If you choose to ignore them, you need to provide a solid reason to explain why this time is different.

The Risk is Indeed High

The stories of all great new technologies have played out in the same way and from the same logic. High risk combined with enormous uncertainty is an inherent attribute of these new tech companies. Looking back at their performances in similar situations, you will find that even those revolutionary companies that eventually succeeded long-term (like Microsoft and Apple) were often battered at similar moments along the way. Moreover, at the early stages of new tech companies (not looking at it in hindsight), it is not easy to judge who will succeed and who will fail—IBM is a prime example. If you open up these cases for examination, you will understand that the future of new tech companies is highly uncertain; that is their nature.

As an example, they either invest too much or too little. The reason is that under-investing guarantees a loss, but they cannot precisely foresee the future, so they cannot know if they have over-invested. Both over-investing and under-investing come with costs.

They also cannot accurately predict all the changes that will impact them, including external ones—tightening of monetary policy, wars, drastic changes in taxation. Therefore, they will experience significant ups and downs, first exciting investors and then scaring off the timid, washing them out and amplifying market volatility. To go deeper: these new technologies and companies that once disrupted predecessors will ultimately be disrupted by newer technologies and companies in ways we currently cannot imagine. We must consider whether similar fates could befall today's companies. The impact of quantum computing is an example of a "known known." What about those that have yet to be envisioned?

What about the risks posed by competitors? For instance, China is producing and distributing AI technology, and its policymakers have a completely different view of the economy and AI. We are in a new technological war where national leaders believe they must win. From China's perspective, AI should be provided for free or at a low cost to everyone because it has immense productivity benefits and can uplift living standards. They view profits as less important, focusing instead on the overall benefits that many people would gain from using these new technologies. I estimate they will enter the international market competition just as they have with cars, solar panels, and batteries.

This situation is reminiscent of many historical moments that can teach us lessons. I cannot help but think of the late Dutch Empire and the early British Empire, where Britain surpassed the Netherlands in shipbuilding and other important industries. Moreover, in the geopolitical conflicts surrounding Taiwan, one possibility to consider is whether China could use "preventing chips from exiting Taiwan" as a tool in their geopolitical games. AI stocks face other risks as well, such as the risk of wealth taxes and other tax increases—this might force those who have heavily invested in these stocks to have to sell; and anti-AI sentiment could generate restrictions on company expansion.

I could list a whole host of concerning issues for you, and I could also provide a similarly long list of promising AI investment opportunities in which I want to bet. I am not saying how these risks will unfold, nor am I arguing against buying AI companies. I am simply stating that there are concentrated risks in the market, which is indisputable, and you should be aware of how to play in this situation. Based on my research on all similar cases and out of logic, I am confident: the risk is high, and the best play in this situation is:

Diversify Well

You probably know that my mantra is diversification. My "holy grail" of investing is to aim for holding 15 good positions that are uncorrelated and bear a balanced risk. In other words:

"A portfolio made up of good bets that is well diversified will outperform a concentrated bet (it has a better risk-return ratio and can engineer better returns at the same risk levels). The more risk is concentrated in a specific market segment, the more you should diversify, especially when the market is driven by a revolutionary new technology that inherently brings massive uncertainty."

This is not an opinion; it is a mathematical certainty. For example, assume the risk-return ratio of a single bet is 0.3 (say, a 6% return with an 18% standard deviation, which is generally considered stock-like levels), and compare holding 5, 10, and 15 uncorrelated bets: I can achieve the same 6% return, but the risk measured by the standard deviation drops to 8%, 6%, and 5%, respectively. That is, 15 good uncorrelated investments can raise my risk-return ratio by 4.3 times (from 0.3 to 1.29). If you wish, you can also add leverage to obtain significantly higher returns at the same risk. This is a fact.

The confidence I have relies on backtesting, on the real returns delivered throughout my over 50-year investment career, and on logic that has a high probability of holding: diversifying the bets well and adjusting to the desired level of volatility will yield returns over the long term that are far better than the concentrated bets preferred by most investors. To be more specific, good diversification provides a better risk-return ratio than any concentrated bet; and by adjusting it to the desired risk level, you can obtain higher returns at that risk compared to any other approach.

Because I have made this methodology public, it has now become my "less secret" path to success. However, I rarely encounter people who think about investing strategies in this way—meaning few people consider portfolio construction, or think about how a well-structured, well-diversified bet portfolio would perform compared to concentrated positions in a few stocks in a great transformative industry. Most people's focus is merely on 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 do not is enormous. I will find opportunities to explain how to do this better in more detail later.

Based on all of the above, I think that studying how to play this hand should lead one to ask themselves: how large should my concentrated positions be, and then diversify.

Expected Returns Look Very Low

The high risk is an undeniable fact. Next, I am going to give you a viewpoint that might be wrong: the expected returns are very low. This judgment comes from my analysis of valuations and readings from my bubble indicators—future actual returns on stocks appear to be around -5% to -10% over the next 5 to 10 years, but the uncertainty around these figures is significant. In my view, these stocks are long-duration assets, with high risk, because it is difficult to reliably see the distant future, and they appear to be both expensive and held in uncertain hands.

A Question from My Research Team

In a recent meeting, someone on my team asked me: on what basis do you think the current market configuration is wrong? How do you know the market today lacks diversification is not for some good reason—such as some investors believing the expected returns for AI stocks will be very high, or that when an industry occupies such a high proportion of market value, it is natural to see this kind of concentration in an index, or that when everyone is extremely enthusiastic about an industry, many investors will buy these stocks without intelligently and reliably calculating what future profits will be and how those profits should be priced?

My Answer

Price increases can be attributed to various reasons, not all of which are good reasons. Some investors may focus on prices and push them upward because they perceive them as attractive relative to the fundamentals; some hold these stocks because they believe this is a great new technology and see rising stock prices as confirmation of "this is a good stock"; and others hold index exposure, passively applying a heavy weight to these stocks. In my view, you can grapple with these questions, trying to figure out what you want to do; or you can acknowledge that you don’t need to worry, because you simply do not have enough information to confidently bet. You can easily say: "I do not know enough, so I won’t place this bet." And then really don’t place it.

What leads people to get trapped is the idea "I must form a viewpoint, and my viewpoint must be worth something," when the reality is more likely that you cannot form a sufficiently reliable viewpoint on which to bet. (Note: to clarify, I am not suggesting against placing bets—after all, you cannot avoid betting, as you need to allocate money into some investment or keep it in cash; and most people think cash carries the least risk, when in fact it is the worst long-term investment. My advice is that even if you have no tactical judgment about which market is good or bad, you should understand how to diversify your bets well. The approach is: when you have no confident tactical judgment, hold a balanced, strategic-level asset allocation portfolio. But this is a topic for another time.)

Therefore, I believe that understanding what you do not know and deciding when not to bet is just as important as knowing what you do know and thus deciding to bet.

To put it more simply, I adhere to this principle: since it is generally difficult to gather sufficient information to justify a concentrated bet, the best practice is to only compose a diversified portfolio of your most confident and uncorrelated bets and then engineer this portfolio to the desired risk level. This is my "holy grail" of investing.

At this moment, in facing this hand, I do not believe 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 maintaining diversification is the best strategy for dealing with this "not knowing". I am aware this contradicts the theories you read in textbooks—the textbooks generally assert that markets are efficient, so you "can just trust the market".

To summarize: we currently have an extremely concentrated market surrounding a revolutionary new technology, which should remind us not to confuse excitement for new technologies with the attractiveness of new technology stocks, and then throw caution to the wind by holding a pile of high-risk, highly correlated concentrated bets—especially when we could achieve equally attractive returns with much lower risk through smart diversification.

Appendix: I will not share my positions or tactical judgments with you because I do not want to be your investment advisor. However, I will soon share some key perspectives behind these judgments, including my bubble indicator readings and the logic behind them.