US CPI can be very misleading as a metric, sometimes. Not because the figures are cooked, but because it is subject to being swayed if one thing has a big price change.

As an investor, I'm interested in monetary inflation: how much general purpose purchasing power is being retained by the currency?

An example: imagine that peanuts make up 10% of the CPI basket. A blight wipes out 90% of the peanut crop, and peanut prices go up by a factor of 10. This would drive up a CPI figure, but would be meaningless in terms of the value of a dollar. It's not a problem with the buck retaining its value, it's a problem with peanuts. (or eggs, as we saw recently)

Conveniently, some nice folks at the New York Fed have a way to look at what's happening in terms of the buying power of the dollar, but ignoring idiosyncratic price changes which are specific to a given product or service. They start with the assumption that the price change of every category is equal to the sum of a change which is specific to that category, plus a single global figure. By minimizing errors and doing a bit of math, they come up with N+1 factors for the N product categories: one for each category, and one price to rule them all. The do this because their theory (and observation) is that the "common to everything" rate is pretty persistent, so it's a good input to policymakers.

This is called Multivariate Core Trend (MCT) Inflation, and can be found here.
newyorkfed.org: Multivariate Core Trend Inflation
Click on the graph and you can see the latest figures. There is also a tab which shows a breakdown of the single figure into goods, services, and housing. That tab is a bit hard to understand at first. This is by FAR my preferred inflation metric, to be used judiciously in combination with the US dollar index.

I mention all this because
* The underlying monetary (not product specific) MCT inflation in the US was remarkably steady from July 2023 to August 2025 at around 2.8%, give or take a very small margin.
* It recently start to fall, sharply, now for three months running. As of the November data released last week, the year-on-year rate was down to 2.05%. For the first time this is back down to pre-2020 levels. Interestingly, the figures show inflation slowing among all three sections: goods, services, and housing. The housing section of the triad now has falling prices, which is actually consistent with some of the statements coming from Washington. I'm not a fan of the current team, but credit where credit is due. This post isn't about the *reasons* that the inflation rate is falling.

My suggestion:
Unless and until you have any better information, these figures strongly suggestion that one should probably pencil in only 2% US inflation in the next year as a baseline assumption. While being prepared, as always, for the notion that certain specific things might get a lot more (or less) expensive.

Jim