The AI bubble's circular funding problem

Warren Buffett believed that if you know what you're doing, you should concentrate your holdings in a handful of excellent businesses.

The idea makes sense on paper. The lion's share of returns comes from a small fraction of winning stocks. But there's a catch.

Concentrated investing comes with real risks.

Higher volatility. Vulnerability to specific company disasters. The potential for massive losses when you're wrong. I've been thinking about this lately as I watch my Nvidia holdings climb. And I've made a decision.

I've placed orders to sell half my NVDA position when it moves ~8% in either direction.

Up or down, I'm taking profits.

I will probably put the amount into Berkshire Hathaway Class B shares for "slightly safer diversification" yet with some growth opportunity. I'm thinking about Nassim Taleb's Barbell Strategy

Recent news revealed a circular funding pattern that looks uncomfortably familiar.

Nvidia invests $100B in OpenAI. OpenAI buys $300B in cloud computing from Oracle. Oracle spends $40B buying Nvidia chips for its data centers.

It's a web of deals. A house of cards where a single failure could trigger panic.

Some call it vendor financing, others a quid pro quo arrangement. I feel it resembles the dot-com bubble.

The numbers don't make sense anymore.

Nvidia is valued at $4.6 trillion. That's more than the entire pharmaceutical industry.

OpenAI has $12B in revenue but is valued at $500B. As a half-trillion-dollar company, its profit is deeply negative. Sam Altman recently released Sora, which feels like an attempt at a TikTok clone for memes. He mentioned ChatGPT might include adult content by year's end. It makes me wonder if desperation is setting in.

I saw a Video: The State of the Industry is freaking Me Out - by Hank Green.

Videos like The AI Bubble's Dirty Secret Revealed showing up on my YT newsfeed make me nervous.

This is where Buffett's wisdom meets reality. Yes, concentration with Nvidia can lead to superior returns. But it also requires knowing when you're no longer investing based on fundamentals.

When a company is valued at more than an entire global industry, perhaps that's not knowledge. That's speculation.

When circular funding patterns emerge, that's not innovation. That's a warning sign.

I learned this lesson during the dot com bubble, when I was completely concentrated in Tech Stocks.

Concentrated investors who failed to diversify lost everything. They believed they knew what they were doing right up until they didn't.

The downsides of concentrated investing aren't just theoretical. They're real.

Increased volatility. Overconfidence. Behavioral biases like anchoring and a false sense of security.

I'm reducing my exposure, not because I think Nvidia will fail tomorrow. But because the risk no longer matches the reward. I'd rather take profits now than hold through what increasingly looks like a bubble.

Buffett was right about one thing. If you know what you're doing, concentration makes sense. But (not) knowing what you're doing also means knowing when to walk away.

Why only trim HALF?

A new Channel about the dark side of AI that I discovered to and subscribed to at 3am on Monday:

I'm a glass is half full type of Optimist, and love the promise of AI democratization. An example of this is small new content creators like Winters Mirage are using AI to create amazing futuristic SciFi and music Videos.

So the reason I am divesting only half of NVDA is - I don't know if the near future will be utopian or dystopian. I havent a clue if the market will keep going UP or come crashing DOWN.

I find life and truth is usually no completely black or white; it's shades of gray...

What do you think about the AI bubble? Drop a comment and let me know your thoughts.