Source: Plutus21 Research, Vanguard, SEC API
This chart is doing a lot of work in our thinking right now. For a decade, the median hardware company in the Vanguard Total World universe traded at a persistent discount to the median everything else: roughly 5 to 7 times sales against 9 to 12 times for the rest of the index. That discount existed for good reasons. Hardware is cyclical, capital intensive, and carries gross margins half those of the software and services businesses sitting above it in the chart. The market priced that fundamental difference correctly for ten years.
In the latest quarter, the relationship inverted. The median hardware multiple jumped from 9.9 times sales to 13.4 times in a single quarter, a 35% rerating, while the rest of the universe barely moved. For the first time in the sample, the typical hardware company is more expensive than the typical company of any other kind.
The methodology matters here. This is a cross-sectional median, not a cap-weighted average. Nobody can wave this away as Nvidia distorting the math. The entire cohort has been re-rated. The market is no longer paying a premium for one extraordinary business; it is paying a premium for the category, including its marginal members. And because the median hardware company converts a dollar of revenue into far less gross profit than the median software or services company, the inversion on a gross profit basis is considerably more extreme than the chart shows. Investors are now paying more per dollar of sales for the lower-quality dollar.
The intellectually honest counterargument is that margins and durability in hardware are structurally improving: AI demand turns cyclical capex into a multi-year build, pricing power has shifted toward the picks and shovels, and the old discount deserved to close. Some of that is true. But "the discount deserved to close" and "the relationship deserved to invert in one quarter" are very different claims, and the second one requires believing that a decade of cycle behavior ended precisely when the crowd arrived.
"What the wise man does in the beginning, the fool does in the end." — Howard Marks
So where is the opportunity? It sits on the quiet side of the chart. The blue line, the rest of the universe, is flat to down over the past year while capital crowds into the orange one. Application layer software, digital businesses, and the services companies that will deploy this technology are being starved of flows not because their prospects dimmed but because the rotation into hardware is mechanical: index concentration rises, passive flows follow market cap, and active managers chase the benchmark. That is a flows phenomenon, not a fundamentals phenomenon, and flows phenomena mean revert. A return merely to parity, never mind the historical discount, implies a substantial relative repricing between the two lines.
Our general view is to reduce infrastructure and platform exposure and increase application exposure systematically over two years. Many investors overestimate their ability to time the turn, so a longer-term systematic rotation averages out of one side and into the other, and positions ahead of the flows we expect as the broader market comes to the same conclusion.
In the short and medium term, this is a positioning and flows exercise. In the long term, it is an earnings exercise, and earnings have a way of accruing to whoever sells the output rather than whoever sells the equipment.
