Mega-Cap Tech Is the Last Shelter as 5% Yields Break Everything Else

Let's get one thing straight: the stock market is not crashing. But it is doing something arguably more annoying. It's grinding lower in slow motion while pretending everything is fine. The S&P 500 closed at 7,683 on September 28, down 0.77% on the day, the Nasdaq shed 0.92%, and the Dow dropped 0.67%. Four of the last five sessions have been red. And yet the VIX sits at 14.21, down over 4% on the week. The market is calm the way a person is calm when they've already accepted bad news.

The culprit everyone agrees on is yields. The 10-year Treasury climbed from 4.96% on September 21 to 5.17% by September 25, with Bloomberg's rate strategist Ira Jersey flagging a potential buyer strike in Treasuries and warning the Fed could hike at least 100 basis points over the next year. The Iran conflict blocking the Strait of Hormuz is keeping oil elevated near $92.60 a barrel, which feeds inflation expectations, which feeds yield pressure. It's a doom loop wearing business casual.

Here's the rotation that Dan Greenhaus from Solus put plainly in a CNBC interview: the equal-weight S&P 500 peaked exactly when the 10-year yield spiked in response to oil moving from $80 to $100. Small caps, banks, airlines, and discretionary are all getting quietly crushed. What's holding up? Mega-cap tech. $META and $AAPL are being called underestimated AI plays. $NVDA just announced a $150 billion share buyback, the largest in history, backed by projected cash flows of $200 billion this fiscal year rising to $330 billion next year. That's not a company worried about rates eating its lunch.

On the bull side, Kevin Hassett at the NEC is tracking GDP growth at 5.14% real, citing AI productivity gains and capital deepening as structural tailwinds that rate hikes simply can't touch. Bill Ackman made the same point on Seeking Alpha: traditional monetary tightening is a blunt instrument when the inflation is coming from AI infrastructure buildout and energy demand. You can't hike your way out of a compute shortage.

The bear case is that Janus Henderson's Michael Kontopoulos thinks we're somewhere between 3 and 12 months from a cycle top. His warning signs: margin compression, slowing earnings growth, and widening corporate spreads. None of those are flashing red today, but the 10-year at 5.17% is the kind of thing that eventually shows up in corporate balance sheets with a 6-month delay, like a parking ticket you forgot about until your registration gets flagged.

Watch the $GGG angle if you want a sleeper: Graco is a diversified industrial with 23% net margins, doubled free cash flow over five years, and direct exposure to AI data center construction. Street target is $93. It's trading at $77. That gap tends to close eventually, usually when everyone stops staring at the Iran headlines long enough to notice it.