Let's demythologize Nvidia's extraordinary recent history with an extremely oversimplified outline:

• First, Nvidia understood, promoted, and bet on the growth of gaming-related demand for GPUs and developed the technological and logistical expertise to profit from it.
• Then they were surprised by the extraordinary good luck of crypto-related demand for GPUs which their technological and logistical expertise — built as a bet on something else, but built well — allowed them to handsomely profit from.
• And then they were surprised once again by the extraordinary good luck of hyperscale generative AI-related demand for GPUs which their technological and logistical expertise allowed them to turn into flabbergasting amounts of money.

This isn't a recap of a long and carefully executed master plan: they got majorly lucky at least twice, at the right time, and at scale. But it's also not a story of luck alone. It's also not just a story of lucky opportunities skillfully exploited.

What makes Nvidia's tale interesting is that to be able to exploit those opportunities they had already to be at the Pareto frontier of that technology (understood in the widest sense including logistics, capital structure, deep engineering talent, suppliers, etc). "Being great at GPUs" was indeed valuable in the pre-generative AI, pre-crypto world; in 2024 it might be the most valuable capability in the world. (How long this will remain the case is outside the scope of this article.)

In financial modeling terms technological excellence, and more generally excellence in any activity, comes with a sometimes very out of the money call option on its centrality to the overall economy. It's not just that the more important an activity is to the economy the more valuable your skills are. The lesson of quick technological shifts is that the centrality of a technology or skill can jump exponentially almost overnight and that it's in the nature of winner-takes-most fast-turnaround markets that being at the critical frontier before it becomes critical can have a nonlinear effect on how much of the suddenly oversized cake you can get.

The relative value of technologies, knowledge, and skills can shift quickly over time. Competitive markets give oversized rewards to those at the capability frontier. Those are platitudes in our understanding of the contemporary economy. Less commonly acknowledged is their implication: being really good at what you do has an intrinsic value above what you can currently get from it because it's what allows you to take advantage of changes in the world that make it suddenly more important.

Nvidia's lesson for any company aiming at nonlinear outcomes is clear. Markets rarely underestimate the value of a capability in the current context but they aren't good at predicting what will become valuable later. If you think something will be more important in the future than it is now then invest more into getting great at it than it seems worth investing now. Today's "unnecessarily good" is tomorrow's core competitive capability.