The October CPI print is out at 8:30 ET. Consensus sits at 0.3 percent headline and 0.2 percent core month-over-month. No year-over-year figures will be published because of last year’s data gap. Markets are already pricing an 80–85 percent chance of a December rate hike. The dates I keep are on the 2026 economic calendar.

The circulating story is the usual one: a hotter core number reinforces the hike path, lifts yields, and pressures long-duration tech. A softer number does the opposite. That framing is not wrong. It is also not the interesting part.

The interesting part is what the print cannot change.

High-voltage transformers still sit at lead times above 160 weeks. Morgan Stanley’s revision of U.S. data-center power demand still leaves a shortfall of roughly 32 GW through 2028 even after behind-the-meter generation. AWS cleared its top-line number while trailing twelve-month free cash flow turned to an outflow of $42.4 billion. Hardware continues to arrive faster than the grid can energize the sites. Dark GPUs remain inventory until the power and cooling arrive.

A core print of 0.3 percent or higher will likely pressure tech multiples through the rate channel. A print of 0.2 percent or lower may give them a temporary bid. Neither outcome shortens the transformer queue, fills the gigawatt gap, or turns stranded compute into billable capacity. The capital-cost effect is real at the margin. It is secondary to the physical constraint.

That is why the portfolio continues to hold both the hyperscalers and the energy producers that own or control the input those hyperscalers must buy. The 70/30 framework was built for exactly this split: the software and chips get priced on rates and narrative; the grid sets the pace either way.

Today’s number will move the duration trade. It will not move the grid.