I have been no fan of Meta Platforms (META) in the past.
I owned the stock briefly, selling out during the 2022 “tech wreck,” when it fell to a low below $100 per share.
Now the stock is selling for nearly $550 per share, and it’s the best bet I can see for 2025.
Let me try to break down why.
Buy the Jockey
I have often said I “buy the jockey” not the horse. That is, a great CEO will outperform a bad one, especially in a crisis.
It’s time for me to admit that Mark Zuckerberg is a great CEO. His big calls have shown true vision.
The first is one I’ve praised for years, the 2012 decision to go all-in on the cloud. While huge companies like AT&T and IBM dithered, Zuckerberg put all his chips in, even before he had chips to make the bet. The result is a network of 33 data centers, almost entirely in the U.S. and Europe, and enough fiber capacity to serve the most remote African village at broadband speeds.
The second call is one I found stupid at the time, his “metaverse.” Turns out that, as with the cloud he was just ahead of the game. (Better than a little late.) It’s clear now that AI is the metaverse, and Zuckerberg’s investment in its concepts was not wasted.
Zuckerberg is now one of only two Big Tech founders still running his company, Jensen Huang being the other. And while Huang is approaching retirement age, Zuckerberg is just 40. He has several turns of the wheel ahead of him.
LLaMa
The second reason to buy Meta is LLaMa software, the primary alternative to Nvidia’s CUDA tools.
I won’t pretend LLaMa is open source. It’s not. It doesn’t allow the transfer of LLaMa knowledge into anything but LLaMa, for one thing. Its license is not approved by the Open Source Initiative (OSI).
But you can download LLaMa 3.1. You can use it. You can even modify it. It’s also globally available. That last is important because Asian and African entrepreneurs, whose capital may consist of a phone and an Internet connection, aren’t using CUDA. Facebook’s cloud network makes LLaMa accessible to them, and LLaMa’s license makes it useful.
That’s enough. The AI Era, unlike the cloud era that came before, is built on proprietary advantage. LLaMa gives the little guy a shot. Nothing else does right now.
The market arc of LLaMa will be like that of any open source project. It starts out way behind the Nvidia alternative but, with more hands on the code, it gradually catches up and surpasses it. The Internet runs on open source Linux, not on Microsoft Windows.
One thing that’s clear to me, as I turn 70, is that youth is the fuel of the future. America, Europe and China lack youth. Asia and Africa have youth. There, Facebook represents the free Internet. It’s not poisoned by paywalls the way the western web is.
Another point Americans now may understand. Freedom is relative.
Global reach is what makes Meta competitive with Google as an ad network. The same thing will make LLaMa competitive as an AI model builder.
The Bottom Line
Meta will bring in something north of $160 billion this year, with one third of that becoming net income. Operating cash flow should come in at around $90 billion. It’s a bargain.
It’s going to be a bumpy night. We’re entering a period of great uncertainty, and I certainly see disaster ahead. In that environment I’d like to be riding alongside someone with 20 years of experience in the cloud and $1.4 trillion in market cap.
