There is a recurring mistake in financial markets.
Investors see a very intelligent person with a powerful idea, a deep understanding of a major technological trend and a strong record of making money, and they begin to believe that superior intelligence alone can overcome uncertainty.
It cannot.
Intelligence can improve our ability to understand businesses, technology, economics and human behaviour. It can help us identify patterns that other people miss. It can help us build better models of how businesses and the world operate. It can help us ask better questions. It can help us distinguish signal from noise.
But intelligence cannot provide certainty about a future that has not happened yet.
This distinction becomes particularly important when investing in future-oriented industries such as artificial intelligence, semiconductors, cloud computing, energy infrastructure and other technologies that may transform the global economy.
A Contemporary Case Study
The reported experience of Situational Awareness, the investment firm and hedge fund founded by former OpenAI researcher Leopold Aschenbrenner, provides a striking contemporary case study.
The fund attracted substantial attention after generating spectacular returns and building very large positions around the AI infrastructure theme. By 2026, the firm had become a major participant in the AI investment story. Then the environment changed dramatically.
Recent reporting indicates that the fund suffered severe losses in July 2026 and subsequently sold the vast majority of its public-equity portfolio to Kenneth C. Griffin of Citadel LLC. Reuters reported that heavy losses in technology holdings had battered the fund, while other reporting described margin pressure and a large unwinding of positions.
What the reporting described
The important lesson is not that artificial intelligence is a bad investment theme. It is not even necessarily that the long-term AI thesis was wrong.
The deeper lesson is much more uncomfortable: you can have a very intelligent investor, a powerful thesis, an enormous amount of research and a potentially correct long-term view — and still experience catastrophic investment losses if concentration, leverage, valuation and timing work against you.
That is one of the oldest lessons in financial markets. And it is one that investors repeatedly forget.
The Future Is Not an Exam That Can Be Solved
Investing is sometimes treated as though it were an intellectual competition. The assumption is simple: the smarter investor should make better predictions. The better predictions should produce better investments. Therefore, the investor with the greatest intelligence should eventually win.
Financial markets do not work that way.
Markets are not examinations. There is no answer sheet hidden somewhere in the back of the textbook. The future is continuously being created by millions of companies, over 180 governments, numerous consumers, engineers, scientists, investors, competitors, and entrepreneurs.
Some events can be forecast with reasonable confidence. Many events can’t be forecasted with any meaningful accuracy and precision. And some cannot even be meaningfully predicted until the conditions that produce them begin to emerge.
Consider artificial intelligence. An investor may reasonably believe that AI will become increasingly important to the global economy. That is a perfectly rational investment thesis. But that thesis immediately creates dozens of unanswered questions.
How quickly will AI adoption develop?
How much computing power will be required?
Which architectures will dominate?
Will model costs fall rapidly?
Will open-source models become increasingly competitive?
Which companies will capture the highest margins?
Will chip manufacturers, cloud companies, model developers, power producers, data-centre operators or software companies capture most of the economic value created?
Will customers retain most of the benefits through lower costs?
Will competition eventually drive margins down?
Nobody can answer all of these questions with certainty. An intelligent investor can build a sophisticated framework for thinking about them. But a framework is not a crystal ball.
Being Right About the Future Does Not Guarantee Investment Success
This is perhaps the most important distinction in the entire discussion. There are actually two separate questions.
Will the underlying industry or technology become enormously successful?
Will buying the companies associated with that industry at today’s prices generate attractive future returns?
These are not the same question.
Suppose a company is expected to become a dominant player in a revolutionary industry. Investors become excited. The stock rises from $50 to $100. The market becomes even more optimistic.
Eventually, the company becomes one of the most successful businesses in the world. The original investment thesis turns out to be correct.
Yet the investor who purchased at $400 could still achieve poor returns if the company’s future earnings do not grow sufficiently to justify the price.
This is the difference between business performance and investment performance. A wonderful business can be a poor investment when purchased at an excessive valuation. Conversely, an ordinary business can sometimes be an attractive investment when purchased at a sufficiently low valuation.
For a long-term value investor, the ideal investment is different: a wonderful business purchased at a sensible price. That is the basic idea behind Charlie Munger’s Quality School of Value Investing.
The Market Does Not Reward Intelligence Directly
Markets reward the ownership of assets that ultimately generate attractive economic returns relative to the price paid. Intelligence can help an investor identify those opportunities. But intelligence itself has no direct cash value.
A brilliant investor can lose money. A less intellectually gifted investor can make money.
A sophisticated economist can make a terrible investment. A relatively ordinary business owner can sometimes make an excellent investment simply because they understand what they are buying, pay a reasonable price, and hold it for a very long time.
This is one reason Warren Buffett and Charlie Munger repeatedly emphasised temperament, discipline and avoiding unnecessary mistakes. Investing requires a combination of intelligence, knowledge, judgment, discipline, temperament, and patience.
Intelligence is only one component. And there are situations in which greater intelligence can even create a new danger.
The Danger of Becoming Too Confident in Your Own Model
A highly intelligent person can construct a very sophisticated explanation for why something will happen. The explanation may be internally consistent. It may be supported by hundreds of pages of research. It may contain economic data, technical analysis, industry knowledge, and historical comparisons. It may even be substantially correct.
But a model can only be as reliable as its assumptions.
Consider a simple valuation model. Suppose an investor estimates that a company will grow free cash flow by 25% annually for the next ten years. The investor then assumes strong margins, low capital requirements and a high terminal valuation. The spreadsheet produces an attractive intrinsic value. The investor becomes confident.
Suddenly, the valuation can change dramatically. The problem is not mathematics. The mathematics may be completely correct. The problem is that the future inputs are uncertain.
This is why precision in a financial model should never be confused with certainty about the future. A spreadsheet can calculate an answer to the assumptions you give it. It cannot guarantee that those assumptions will occur.
The Illusion of Precision
This is one of the most dangerous psychological traps in modern investing. Suppose an analyst produces a valuation of $437.82 per share. That number looks precise. But perhaps the underlying assumptions are:
$437.82 per share
27% revenue growth · 43% operating margins · 14% discount rate · 4% terminal growth · 30× terminal earnings multiple
There is nothing inherently wrong with making these assumptions. The problem begins when the investor forgets that they are assumptions.
The output looks scientific. The uncertainty remains enormous.
A valuation of $437.82 can create the psychological impression that the investor knows the company’s future value to within a few cents. In reality, the economically meaningful question may be whether the intrinsic value is somewhere between $250 and $600. That is a completely different level of certainty.
This is why I believe the best investors should become increasingly comfortable with ranges rather than precise predictions. The question becomes: what range of outcomes is reasonably possible, and what price gives me enough protection if reality turns out worse than expected?
That is a far more useful question.
Concentration Increases the Importance of Being Right
Concentration is not automatically bad. If an investor has deep knowledge of a business, understands its economics and can tolerate volatility, a concentrated position can generate excellent returns.
But concentration changes the mathematics of being wrong.
Suppose you own 30 companies. One company suffers a permanent competitive setback. The damage may be manageable. Now imagine that 60% or 70% of your portfolio depends upon one industry continuing to perform exceptionally well. The situation changes.
You are no longer making one forecast. You are making a series of interconnected forecasts — industry growth, technological progress, competitive structure, market share, margins, capital expenditure, regulation, financing conditions, and valuation multiples.
If several of these assumptions move against you simultaneously, the portfolio can suffer enormous damage. That is the hidden cost of concentration. Concentration increases the consequences of being wrong.
Leverage Changes the Game Completely
Then comes leverage. Leverage is powerful because it magnifies returns. If you invest $100 and earn 20%, you make $20. If you borrow another $100 and invest $200, a 20% return produces $40 before financing costs.
The same mechanism works in reverse.
The mathematics becomes even more dangerous when lenders or prime brokers can require additional collateral. At that point, the investor may be forced to sell.
This creates a critical distinction: an unleveraged investor can wait for the market to recover. A leveraged investor may be forced to sell before recovery occurs.
That difference can turn temporary volatility into permanent capital loss. Decades of research on leverage and financial markets have also shown how margin constraints can amplify market declines because falling prices can trigger forced selling, which causes further price declines.
This is why leverage is fundamentally different from ordinary investment risk. Without leverage, a temporary decline may be survivable. With leverage, the same decline can become existential.
The Path Matters, Not Merely the Destination
This is another reason forecasts can be misleading. Imagine that an investor correctly predicts: “Company X will be a multi-bagger within five years.” That does not mean the investor will necessarily earn a good return. What happens during those five years matters.
If the investor is unleveraged and psychologically prepared, they may survive. If the investor is leveraged, they may be forced to sell.
The company eventually reaches the predicted value, if it is based upon reasonable and realistic assumptions. The forecast was correct. The leveraged investor still lost money.
This is called path dependency. Investment outcomes depend not only on where an asset eventually ends up but also on the path it takes to get there. This is particularly important for investors who use borrowed money.