Just caught something interesting about how a billionaire investor is reshaping his portfolio, and it tells you a lot about where smart money is moving right now.

Bill Ackman's been making some calculated AI bets through Pershing Square Capital. He grabbed Alphabet back in 2023 when everyone was panicking about ChatGPT, then picked up Amazon last year during the selloff. Both moves have worked out pretty well. But here's what's wild—his most recent AI stock purchase is already up 1,650% since going public, and he thinks there's way more upside ahead.

Meanwhile, he just completely exited a position his fund held for years: Hilton Worldwide. Now, this wasn't a bad investment by any stretch. He originally bought in 2018, added more near the start of COVID, and watched the company transform. Hilton grew its loyalty members from 85 million to 243 million. They expanded from 913,000 rooms to over 1.3 million. EBITDA went from $2.1 billion to $3.7 billion. That's solid execution.

But here's the thing—the stock price ran way faster than the fundamentals. Up over 350% since end of 2018. The valuation got stretched: enterprise value tripled, EV-to-EBITDA hit 21.5, and the forward P/E sits at 36. At those levels, even a billionaire investor with a long-term horizon sees diminishing returns ahead. So he took the gains and moved on.

What did he move into? Meta Platforms. According to Ackman, Meta's business model is basically built to benefit from AI integration. The market's been spooked about Meta's massive capex spending—they're talking $115 to $135 billion this year, a 73% jump. But Ackman sees this differently. He thinks it's front-loading infrastructure costs that'll pay off huge.

Look at what's already happening: Meta's AI recommendation algorithm is driving engagement up, ad impressions jumped 18% last quarter, and average ad pricing went up 6%. The real potential? Generative AI could unlock entire new advertising channels—think chatbots in Messenger and WhatsApp, or ads in Meta's own AI assistant. That's serious monetization runway.

The valuation argument is pretty compelling too. At 22 times forward earnings, Meta looks cheap compared to where it could go. Strip out Reality Labs and the core ad business trades at just 18x earnings. For a company with that kind of AI tailwinds and growth potential, that's genuinely attractive.

What this move really signals is that billionaire-level capital is flowing toward AI infrastructure plays that can actually monetize it. Hilton was a good business at a stretched price. Meta's a good business at a reasonable price with exponential AI upside. That's the calculation that matters when you're managing billions.