Nvidia just hit another all-time high and its market cap is closing in on $6 trillion. Two years ago that number would've sounded like a typo. Per Yahoo Finance, the run is backed by an August Q2 earnings report that beat Wall Street's expectations on both the top and bottom lines.

So the fundamentals did show up. Revenue and profit both beat estimates, which is the part that makes the rally at least somewhat defensible. This isn't a memecoin situation where the price floats on vibes. There are actual customers paying actual money for actual chips.

But you know me. I can't leave the happy part alone.

I've been writing for weeks about the valuation signals I track flashing dot-com yellow, and Nvidia is basically the sun that whole signal orbits. A company this size climbing this fast means the price already knows the good news, twice over probably, and the room for disappointment keeps shrinking while the price keeps climbing. That's how tops get made. No crash announcement, just expectations that finally get one inch too tall.

The bigger thing I'm watching is the widely reported moves by hyperscalers toward building their own AI chips. I can't verify the details from where I sit and I won't pretend I can, but the logic keeps me up at night anyway: Google, Amazon, Microsoft, they all have reasons to want out from under Nvidia's pricing, and every in-house chip that ships erodes that customer concentration advantage a little more. When your biggest customers are also your most motivated future competitors, the word for that isn't moat. It's complicated.

None of this means Nvidia is a bad business. It's arguably the most important company of this decade. It means the stock price already reflects that, and then some.

For my own money, I don't hold Nvidia directly, and a stock at all-time highs near a $6 trillion valuation isn't going to make me start. My boring automated index fund buys mean I own plenty of Nvidia anyway, it's one of the biggest slices of the S&P 500 whether I like it or not. That's the topping-on-the-meal approach I keep telling you about, except the index does the portioning for me.

Warren texted me about this one. His take was something like "congrats to them, I'll keep the bonds." Mine was that the bonds are fine for him and I'll keep buying every payday, but he's not wrong that a company this size needs everything to keep going right just to hold still. We're both right. We usually are. It's annoying.

If you own Nvidia, the honest advice is boring: don't add on FOMO, know what percentage of your portfolio it actually is now (it's probably more than you think after this run) and don't let one company's earnings report be your whole retirement plan. If you don't own it, you're not late to investing. You're just early to the next decade, which nobody can predict anyway.

My buy hits Friday like always. I'll be watching whether this valuation signal finally goes red, but my hands stay off the pause button until the math changes, not the headlines.