If you could only hold one stock for the next 50 years, who would you choose? The average lifespan of companies in the S&P 500 has shrunk from 30 years to 15 years. This means that 90% of your assets might disappear before 2040.

After analyzing 100 years of volatility patterns and moat evolution, I found that most people's understanding of "long-term value" is completely wrong.

1. The truth about long-term returns: profits are the only anchor Short-term stock price fluctuations (within 1 year) are mainly driven by valuation and sentiment, but over a 10-year horizon, 74% of returns are explained solely by profit growth. Dividends are just the dessert, and valuation expansion is an illusion. If you want to hold a company for half a century, you must ensure it has the "business DNA" to adapt to industry evolution.

2. Why do tech giants often fail to last long? The fall of IBM, Nokia, and Oracle offers a common lesson: their moats are built on specific "technological architectures." Once the technological paradigm shifts (such as mobile replacing PCs, AI replacing traditional cloud), those once impregnable technological barriers can instantly become heavy liabilities. This is my core reason for excluding Google: when AI agents replace search boxes, Google's front-end advertising logic will face fundamental upheaval.

3. The commonality of top assets: utilities based on human basic needs Through layers of screening, the final winners must possess "resilience." S&P Global is too stable but lacks elasticity; Costco is limited by physical expansion and retail competition. Only Amazon is evolving from a retail company into a "fundamental operating system for the digital and physical worlds."

50-year logic: the transaction hub Regardless of whether humans buy things via smartphones, AR glasses, or brain-computer interfaces in the future, goods always need to be stored, processed, transported, and delivered. Amazon's true value lies not in that website, but in its unmatched heavy-asset infrastructure. As long as transactions pass through its ecosystem, it can collect commissions. This "physical utility" based logic is much more resilient than the life cycle of pure software companies.

AWS is not just cloud services; it is the computing infrastructure for the future AI era. Just as electricity was essential during the Industrial Revolution, the global demand for computing power and storage in the coming decades will be rigid. This diversified growth engine gives it a strong risk resistance even when retail profits are under pressure.

Many criticize Amazon for sacrificing profits for infrastructure investments, but from a 2026 perspective, this is its deepest moat. Such "inefficient" large-scale investments translate into extremely low marginal costs over the long cycle. When competitors try to catch up, due to capital costs and physical space limitations, the cost of imitation becomes astronomical.

The "dimensionality reduction" of business models The traditional moat is a wall, but Amazon's moat is water. Water is shapeless but ubiquitous. It locks in users through Prime subscriptions, secures the supply chain through logistics, and binds developers through AWS. This full-chain penetration is the underlying logic that allows it to withstand 50-year cycle fluctuations.

Conclusion: To find winners over 50 years, don't look for the most advanced; look for the hardest to replace. In a changing world, bet on those things that "won't change."