⚠ AI Commoditizing the StackModerate threat

The subtler and more interesting threat to Microsoft is that artificial intelligence, which the company has embraced so eagerly, could over time loosen the very lock-in that is its greatest asset. Microsoft's moat rests on switching costs — the enormous difficulty of moving off its software, its file formats, its trained-in habits. But if AI makes it dramatically easier to build new software, to translate between formats, to automate the migrations that were once forbiddingly complex, then the walls that took decades to build could, in principle, become easier to climb.

Consider what the deep moats actually depend on. Learned-skill lock-in assumes it is costly to retrain a workforce onto new tools — but AI assistants can flatten a learning curve. Data gravity assumes it is hard to move and reformat years of files — but AI can make translation between systems trivial. The compatibility network assumes everyone must use the same programs to exchange work — but AI intermediaries could bridge incompatible tools invisibly. None of this is certain, but each is a plausible way that a technology Microsoft is championing could quietly corrode the friction that keeps its customers captive.

The reasons for calm are substantial, and they are why Microsoft has so far been a winner from AI rather than a victim. The company is not standing still; it is racing to put AI at the center of its own products, so that if the technology does reshape how software works, Microsoft owns the reshaping rather than being disrupted by it. Its distribution, its enterprise relationships, and its cloud give it enormous advantages in deploying AI, and lower switching costs cut both ways — they make it easier to win new customers as well as to lose old ones.

A long-term owner should file this as the most important thing to watch, precisely because it is the one force that could weaken the moat from the inside rather than chip at it from outside. It is not a present danger — today AI is deepening Microsoft's franchises, not dissolving them — but it is the scenario that most deserves ongoing attention, because a switching-cost moat in a world where switching gets easy is a moat whose foundations are quietly shifting. The prudent stance is confidence tempered by vigilance: Microsoft is winning the AI transition so far — Azure grew 41% and passed $100 billion, and the backlog grew 84% — but it is paying for the victory in a currency it never used to spend, with capital expenditure now taking 63% of operating cash flow against 38% two years ago¹. That is the shape of the danger. A switching-cost moat in a world where switching gets easy is a moat whose foundations are quietly shifting, and the bill for holding it is already visible in free cash flow.