Listen to the audio version of this article (generated by AI).

It started on a Thursday.

The 10-year Treasury ticked up a few basis points. Nothing alarming. Yields move every day. Traders noted it, moved on, closed their books.

Friday it moved again. A little more this time. The 30-year followed. By the close, the long end of the curve had shifted enough to register on the dashboards but not enough to interrupt anyone’s weekend.

The financial press mostly wrote it off as routine repricing, while the talking heads on CNBC spent their segments on earnings.

By Monday, yields kept inching higher. On Tuesday it accelerated. By midweek the 10-year was pressing against levels we hadn’t seen since the Great Financial Crisis, the 30-year was inching toward 6%, and the AI trade (the entire AI Boom, the engine that has carried this market for two years) began to wobble.

Somewhere in the Eccles Building, we can assume a newly confirmed Fed chair was watching his bond market send him a message he could not afford to misread.

It’s rare that a breaking point announces itself… What it does instead is slowly accumulate. First with a tick higher, day by day. Then with a consumer who is bruised but not broken, an earnings call where comparable sales rise a “functional” 0.4%. Each detail is dismissible on its own, yet each is another basis point of pressure on a system where the center is about to give way. Most investors don’t realize this…

But this is the most important macro moment of the AI bull market.

Get it right, and the path to new highs reopens. Get it wrong, and the Achilles heel of the entire trade gets exposed.

Every basis point higher in the 10- and 30-year puts downward pressure on growth stock valuations through pure discount math, while simultaneously squeezing the American consumer who, whether they realize it or not, is funding the entire AI capex cycle.

Inflation just printed 3.8% in April. May is tracking 4.2 to 4.3%. Over the past six months, month-over-month inflation has averaged plus 0.4%. Run that math forward and we’re looking at 5.2% inflation by November. No sane Fed chair cuts rates against a 4 to 5% inflation backdrop. And yet that’s exactly what Kevin Warsh — Trump’s pick to replace Jerome Powell — was installed to do.

The best thing Warsh could do to end this yield spike is hike rates. Or at minimum, deliver hawkish commentary that signals he understands the assignment. Below, I’ll break down why the bond vigilantes are sending Warsh a message, the precise 10-year yield level that separates a buyable dip from a serious problem, and the two stocks I’m pounding the table on once these yield jitters pass.

Click the video below to watch this week’s episode of Being Exponential, where I break this all down:

The Market Is Testing the New Fed Chair

The timing isn’t coincidental. Warsh was confirmed on a Wednesday. The yield spike started Thursday. Continued Friday. And accelerated into this week.

The bond market is looking at a Fed chair installed specifically to cut rates, staring at a macro backdrop where inflation could hit 5% by year-end, and saying: “Drop the rate cut act. Be an adult in the room.”

This is the bond vigilantes at work. And the only way to get them to stand down is for Warsh to prove he’s serious about fighting inflation. Hawkish commentary alone might do it. An actual rate hike would absolutely do it. You bring up the short end to save the long end.

There’s also a structural factor at play. Warsh is famously anti-QE. He views quantitative easing as a wealth shift mechanism. With QE off the table as a tool, the long end of the curve loses a key buyer of last resort — which justifies structurally higher long-term rates and structurally lower P/E multiples.

The 5% Line in the Sand

Here’s why I’m not panicking. Most of the gains in this market aren’t being driven by multiple expansion. They’re being driven by earnings growth. Memory stocks are trading at single-digit P/E multiples. Semiconductors driving the majority of market gains are at reasonable valuations. The S&P 500’s forward three-year EPS CAGR sits at roughly 16%.

That math works — even with the 10-year at 4.5 to 5%. We’re talking a 5 to 10% pullback, then a resumption of the AI trade.

But break above 5% on the 10-year and things change fast. That’s when Main Street breaks. Look at Home Depot’s earnings this week: the consumer is still fixing leaky toilets but not splurging on $75,000 kitchen remodels. Push yields higher and that bruised consumer becomes a broken one. Meta’s ad business slows. Google’s ad business slows. Amazon’s retail business slows. The hyperscalers have fewer dollars to fund their AI capex commitments. And that 16% forward EPS CAGR? It collapses to 8%, maybe 5%.

That’s the Achilles heel.

Two Stocks for When the Storm Clears

Two names I’m pounding the table on include potentially the best play on Google’s custom silicon TPU buildout, especially with Google’s new Blackstone-backed NeoCloud entering the hyperscaler arena. And the space-economy ETF positioned for the SpaceX IPO, which just got upsized to a $2 trillion valuation.

Tap into this week’s episode of Being Exponential for the full breakdown — including how this plays out and the rapid-fire calls I couldn’t fit here. Also, be sure to subscribe to Being Exponential on X (formerly Twitter) for more exclusive content.