Generational CadenceNarrow moat

Nvidia's cadence is the metronome of the whole industry: it ships new generations of chips faster and better than anyone else, so that being merely one cycle behind is a real and painful handicap.¹ A customer training frontier models against well-funded rivals cannot afford last year's hardware, so the biggest buyers keep upgrading almost as fast as Nvidia can build — and the pace itself becomes a barrier, because to compete a rival must sustain a matching rhythm indefinitely.

The cadence is a moat because it compounds pressure on competitors. A rival that finally catches this generation discovers Nvidia has already launched the next, so the target never sits still long enough to be hit. Matching Nvidia once is hard; matching it every eighteen months, generation after generation, on power, performance, and software support all at once, is a far taller order that has broken more than one well-funded challenger.

Sustaining the pace requires a war chest and an organization few can assemble — enormous R&D budgets, privileged access to the leading foundry's newest process, deep relationships across the supply chain, and the accumulated institutional knowledge of having done it many times before. Each of these is itself a barrier, and together they mean the cadence is not just fast but expensively, structurally fast.

Yet a careful owner marks this as a narrower moat than the software threads, because a cadence is a lead that must be re-won every single generation, on the merits, against everyone. It is a race, not a toll: lose one step — a delayed launch, a flawed design, a supply stumble — and the advantage can narrow quickly. Nvidia has run the race superbly for years, but a race, unlike a habit, can be won by a rival who simply runs faster for a while.