The 10-Year Yield Hit Its Highest Level Since 2007 — What the Rate Shock Means for the AI Trade

The AI trade spent the first half of this week setting records. The Nasdaq closed at a record on both Monday and Tuesday, finishing September 22 at 27,244.28, and AMD crossing $1 trillion in market value on Monday capped a five-day run up roughly 24%. Then Wednesday arrived with a different kind of headline: the 10-year Treasury yield spiked above 5.11%, its highest level since 2007, and the rally stopped. The S&P 500 fell 0.8%, the Dow dropped 0.7%, and the Nasdaq — the index that had just posted back-to-back records — gave back 1.1%. Even bitcoin slid more than 3% as the rate shock rippled through every risk asset at once.

Nothing about the AI demand story changed on Wednesday. No chipmaker missed a quarter, no data center deal fell apart, no order book thinned out. What moved was the price of money itself, and that is the part of the AI trade this blog has not yet given its own chapter — the discount rate, the quiet variable underneath every multiple in the sector.

What pushed yields to a 19-year high

Three pressures stacked up in a single session. First, oil turned higher — Brent traded near $98 a barrel — as supply concerns kept energy prices elevated, and energy prices feed directly into inflation expectations. Second, S&P’s gauge of US business activity expanded more than economists expected, a strong-economy reading that sounds like good news but pushes yields up because it implies growth and prices running hotter, not cooler. Third, and most importantly for the rate path: Fed governor Michael Barr called for “further policy adjustments” to bring inflation down, and bets rebuilt that the Federal Reserve would hike rates at least once more this year.

Put together, the market spent Wednesday repricing the entire future: more inflation persistence, more Fed tightening, and a 10-year yield above 5% that had not been seen in nearly two decades. The VIX, the market’s fear gauge, jumped almost 7% off its recent lows. This is the same selloff mechanics as the September 14 chip rout, just with a different trigger — that one started with executives questioning the pace of AI spending; this one started with the bond market questioning the price of everything.

Why a 5% risk-free rate hits AI names hardest

The Nasdaq fell only about a third more than the S&P 500 on the day, but the mechanics behind the move matter more than the size. A stock’s value is the sum of its future cash flows, discounted back to today — and when the discount rate rises, the cash flows furthest in the future lose the most value. High-multiple growth names, the AMDs and Nvidias of the world, are priced on the assumption that enormous profits arrive years from now. A 5% risk-free rate is also a personal hurdle rate: the guaranteed return available in a Treasury is now high enough to compete with the volatility of a chip stock that can swing 5% in a single session, as this month has demonstrated repeatedly. Every investor’s required return quietly moved up with the yield, and valuations built for a cheaper-money world have to adjust.

That does not mean the AI trade is over — the demand evidence remains overwhelming, as this month’s record prints from AMD, Micron, and Nvidia all showed. It means the trade now has two forces acting on it instead of one. The earnings story says buy; the discount rate says be careful about the price you pay. September has been a live demonstration of both: a 5.6% chip rout on the slowdown call, a relief rally, a trillion-dollar milestone, and now a 19-year yield high, all inside three weeks. The direction of the theme and the volatility of the ride are two different questions, and a portfolio has to be built to survive the second while it waits for the first.

The rate-shock checklist for your own positions

A yield shock is a good moment for a cold look at what you actually own, because compressing multiples do not announce themselves — they just show up as a position that stops going up. The first step is seeing your exposure across every brokerage at once, with what you paid for it, instead of logging into each account and adding up tickers by hand. That is exactly what InvestSheet removes: it syncs positions from Robinhood, Fidelity, Schwab, and 35+ other brokerages into a single Google Sheet, so every AI name shows up once with live formulas no matter which account holds it:

With value and cost basis side by side, run the rate-shock checks. Has a runner’s multiple drifted past your single-name cap after a 24% five-day run? Is your combined AI exposure — direct shares plus the tech weight inside any fund — larger than you planned when rates were lower and every multiple expanded? And does your cash allocation acknowledge that a guaranteed 5% now exists as an alternative? Positions sized when the 10-year yielded less look different at 5.11%. Seeing the whole picture in one sheet is the only way to make that comparison honestly — and to decide deliberately whether you are holding the trade or just riding its volatility.

Frequently asked questions