Friends,

When I was a kid (Stoffel, here), my parents and I had a deal. They would pay for the first three years of college. Year Four was entirely my responsibility. After that, I told them I didn't want to have to borrow any money from them. If, however, I needed a co-signer for rent or a car loan, I would be willing to ask.

In the end, it all worked out fine. A combination of loan forgiveness for my years of teaching, locking in a low interest rate, and living below my means helped me clear the debt just fine.

But there's no way I could've done it on my own. In fact, my parents couldn't have done it on their own, either.

Back then, the school I went to had a $30,000 sticker price (it's MUCH higher today). Scholarships helped lower that figure. But I've since found out that an inheritance from my grandfather also helped cover the cost.

And most importantly, no one ever asked to be paid back. For that, I'm insanely grateful.

Broadcom (and NVIDIA) as the (grand)parents

Of course, there's a connection to investing here. Over the past month, two of the world's most important technology companies have announced huge deals.

• NVIDIA: The company announced it was partnering with major financial institutions like Apollo and Blackstone to arrange up to $500 billion to help fund AI's buildout.
• Broadcom: The custom-silicon provider is considering lending its creditworthiness for up to $100 billion in financing for similar projects -- with similar financiers like Apollo and Blackstone participating.

On one hand, I hear stories like this and think of what AI bears like Michael Burry (below) would have to say.

Cassandra Unchained @michaeljburry Well, you could have heard it first, months ago, 2025 even Like compression’s threat now, to be covered in 2027 by @WSJ and @CNBC 4:31 AM • Aug 17, 2026 329 Retweets 2209 Likes Read 159 replies

​

On the other hand, it's also unreasonable to think that a company like Anthropic -- which likely had $1 billion in revenue last year and zero profits -- could afford to build the multi-billion-dollar data centers (and all the chips in them) necessary to meet the demand for AI compute.

Yes, we realize Anthropic's run rate has soared this year. But that would be akin to telling a recent high school grad that they were on their own in paying for four years at Harvey-Mudd College in California -- the most expensive in the U.S. The estimated price tag for four years: nearing half-a-million dollars!

Where's the line?

Of course, this is the actual fate that several students face every year. But they also have the option of going to cheaper schools, or going right into the workforce.

Anthropic and OpenAI have no such option: the cost for chips and data centers are what they are. There's no way around it. So to help them get off the ground, players like NVIDIA, Broadcom, Apollo, and Blackstone (among others) are stepping in to fill the gap.

Unlike most parents, they aren't doing it out of the goodness of their hearts. They believe it's in their own best financial interests to do so.

But that doesn't mean the move isn't without risks. If demand slows or open models commoditize Anthropic / OpenAI, it could create real trouble for everyone involved.

The hardest part is knowing where the line is. I drew a line in the sand when it came to borrowing money after college. Imagine, however, that I made terrible financial choices that threatened my future.

• Would I have held that line?
• Would my parents have held it?
• How far would it have been healthy for them to bend?
• Will I hold that line when I have to pay for my own kids' college?

There's no easy answer to any of these questions. And while helping your loved ones is fundamentally different than funding a buildout in a capitalist economy, these decisions rhyme with each other.

When it comes to your own portfolio, it's wise to focus on the most important factors you can control: investing in wide moat businesses with solid balance sheets and reasonable valuations. You should also assure that you aren't over-exposed to a trade that gets riskier over time.

That doesn't mean go out and sell your exposure to the "AI Trade", it just means you need to monitor how much you devote to it. In the long run, the goal isn't to have the best returns in the world -- it's likely to have the freedom to offer your kids (and grandkids) the same type of deal I was lucky enough to get.

Wishing you investing success,

Brian Feroldi, Brian Stoffel, & Brian Withers Long-Term Mindset

One simple graphic

One piece of timeless content

There are discussions in D.C. about ending the capital gains tax. We don't tread into the political scene, but this change could significantly impact U.S. investors, so we wanted to share Graham Stephan's overview of the situation in Removing the capital gains tax.

One resource

Hyperscalers are planning over $600B in capital expenses this year for AI buildouts, rising to $1T next year. Bulls cheer; bears are calling a bubble. The folks at Market Sentiment modeled out the economics in Is the capex sustainable?​

One Stock Dive

A Stock Simplifier user asked if we could create a "one-page" overview of a company after a stock was rated.

We thought that was a great idea. So we built it.

Here is Feroldi's one-page overview of Alphabet (NASDAQ: GOOGL).

To make a one-page overview of any stock, simply rate each category, and the download button will appear.

One quote

👋 This newsletter was...

​🧠🧠🧠🧠🧠 Awesome!​

​🧠🧠🧠 It was OK​

​🧠 Do better​