원문 보기
Hello investors. The markets are sending a clear signal. Artificial intelligence is no longer just driving expectations. It is moving capital, reshaping infrastructure, and pushing major stock indexes to new highs. The S&P 500 recently reached another record supported by strong gains in major technology companies.
Nvidia rose nearly 4% while Meta gained around 15% for the week after reports that it was improving the cost structure of its artificial intelligence operations. That detail matters. Investors are no longer rewarding companies simply for spending heavily on AI. They increasingly want evidence of efficiency, stronger margins, and measurable financial returns.
The impact is also spreading beyond software. Anthropic signed a $19 billion agreement linked to a new data center project in Kentucky. The 20-year contract involves a facility capable of supporting approximately 400 megawatts with a first power delivery expected in the second half of 2027. This reflects a broader shift.
The AI race is becoming a physical infrastructure race. Data centers, electricity generation, transmission capacity, cooling systems, memory, and storage are now essential parts of the technology boom. But while the market celebrates this expansion, Michael Burry is betting against one of its most popular segments.
As always, this video is not intended to provide financial advice. The goal is to share objective information so you can do your own research and make decisions with better context. Keep in mind that all investments involve risk. Burry is not predicting the collapse of artificial intelligence or a broad financial crisis.
His wager is more specific. He has taken short positions in Micron and the SOXX semiconductor ETF while going long Microsoft. His thesis is simple. Sell the companies whose valuations depend on an unusually favorable cycle and by a company that may benefit from AI through a more durable business model.
To understand the trade, we first need to understand the enthusiasm surrounding memory companies. In 2026, Micron, Samsung, SK Hynix, and SanDisk have delivered extraordinary returns. The market's argument is persuasive. AI systems require not only GPUs, but also high bandwidth memory and enormous amounts of data center storage.
Supply is concentrated among only a few manufacturers. Demand has grown faster than production capacity and much of the industry's future output has already been committed. This shortage has driven prices and profits sharply higher. Analysts are overwhelmingly bullish on Micron and the company has also received strong political support in the United States.
That combination, scarcity, investor enthusiasm, analyst optimism, and the political endorsement has created one of the strongest narratives in the market. And it is precisely when the consensus appears nearly unanimous that Burry has decided to bet against it. His argument against Micron has two parts.
The historical quality of the business and the valuation investors are currently paying. First, the business. Burry does not deny that demand for memory is real or that manufacturers are earning exceptional profits. His point is that strong demand does not automatically transform a cyclical company into a high-quality long-term investment.
Over the past 42 years, Micron's median return on invested capital has been approximately 4% while its cost of capital has exceeded 10%. In simple terms, the company has often earned less on its investments than it paid to finance them. Micron has also produced negative free cash flow in nearly half of the quarters examined.
That means the company has frequently consumed more cash than it generated after funding the factories and equipment required to remain competitive. This is the central weakness of the memory business. When supply is limited, prices and profits rise rapidly. Those profits encourage companies to build factories and expand capacity.
But new plants take years to become operational. By the time they begin producing, supply may exceed demand, causing prices, margins, and earnings to collapse. The cycle is familiar: scarcity, rising prices, capacity expansion, oversupply, and falling profits. Micron's stock market history reflects that pattern.
Over 42 years, its shares have experienced 34 declines of more than 30%, 18 declines of more than 50%, and seven drawdowns exceeding 70%. Supporters argue that this time is different because AI creates a structural source of demand, and high-bandwidth memory capacity is already sold out. Burry accepts that demand may remain strong for years, but he argues that every shortage cycle comes with a reason why the current product is supposedly unique, strategic, and permanently scarce.
Today, that product is HBM. The second part of his thesis is valuation. A cyclical company can still be an excellent investment when purchased cheaply. The problem is that Micron is no longer cheap. According to Burry, the company is trading at more than 12 times sales compared with a historical median of approximately 2.2 times.
That places the stock at the highest valuation in its 42-year history. Some investors point to Micron's price-to-earnings ratio and argue that it does not look excessive. But the PE ratio can be misleading for cyclical companies. At the peak of the cycle, profits are temporarily inflated, making the stock appear cheaper than it really is.
If earnings later collapse, that apparently attractive valuation disappears. Samsung recently offered a useful warning. The company reported exceptional preliminary results, including a 19-fold increase in operating profit. Yet, its shares fell because the market had expected even more. When a company reports extraordinary news and the stock still declines, the issue is often not the result itself.
It is the amount of perfection already reflected in the price. This matters beyond Micron. Estimates cited in the analysis suggest that Micron alone could represent approximately 16% of the S&P 500's projected earnings growth in 2026. More broadly, semiconductor and AI infrastructure companies could account for roughly half of expected market earnings growth.
In other words, the S&P 500 increasingly depends on the AI, memory, and semiconductor boom continuing. If the cycle turns, the impact may extend far beyond a single company and weaken one of the main forces supporting index performance. Why might investors be underestimating that risk? Scarcity makes shortages appear permanent.
Social proof reinforces the belief that AI-related companies are obvious winners. Analyst recommendations and political endorsements add authority. And after years of enthusiasm, every new rise in demand is interpreted as confirmation that the original thesis was correct. None of those factors proves that Micron is overvalued.
But together, they can make investors less willing to question whether current profits are sustainable. This is also why Burry's position deserves more attention than some of his previous predictions. Many of his earlier warnings depended on forecasting recessions, inflation, central bank decisions, or broad market crashes.
This trade is narrower and based on company-specific data. He sees a deeply cyclical business, a long record of weak returns on capital, and severe drawdowns, and the highest valuation in its history. He does not need AI to fail or the economy to enter a recession. He only needs the memory cycle to behave as it has in the past.
At the same time, he is buying Microsoft. That makes his strategy less a bet against artificial intelligence than a rotation within it. Away from companies whose profits depend on scarcity and toward a business that may benefit from AI through software, cloud services, and a more stable economic model.
If you find this information useful, consider liking the video and subscribing to the channel. It helps us continue sharing relevant market insights with people who take their personal finances seriously. There is, however, a credible case against Burry. AI demand could remain strong enough to absorb new capacity for much longer than historical cycles suggest.
Technical constraints may limit production growth, allowing memory manufacturers to sustain high margins. If Micron begins generating returns above its cost of capital for years rather than quarters, its historical record may become less relevant. That outcome is possible, but the current valuation leaves little room for error.
The higher the expectations embedded in the share price, the more damaging even a small disappointment can become. So, whom should investors believe? Neither Donald Trump nor Michael Burry should be treated as an authority whose opinion replaces analysis. The answer lies in the data. Return on capital, free cash flow, capital expenditure, valuation, production capacity, and the company's position within the cycle.
No one can know exactly when the memory boom will end. The euphoria may continue much longer than short sellers expect. But Burry's trade is not irrational. It is built around a specific contradiction. A company with a long history of cyclicality and capital destruction is being valued as though today's extraordinary profits were permanent.
That is the opportunity he believes the market has created. For investors, the lesson is broader. The goal is not to predict every market movement or determine exactly how long the enthusiasm will last. It is to understand what an asset is worth, identify the assumptions built into its price, and recognize when expectations may may moved too far away from economic reality.
We hope this information was useful. We will see you again very soon with more financial news and market insights. Thank you very much for watching.
Comments 0