Mark Zuckerberg spent money he doesn't have on AI capacity he can't explain to build products he won't specify for customers who don't exist yet. Told investors there's a trade-off between selling the AI and keeping it. Failed to mention which side of that trade-off makes money.
The dilemma breaks down like this: Meta built data centers the size of small nations to train AI models that either power Facebook's ad-targeting system or get sold to enterprises as cloud services. Zuckerberg won't say which one. Probably because he doesn't know. Definitely because both options require admitting he's competing with Microsoft and Amazon in a market they've owned for a decade while also admitting Facebook still makes all its money showing you ads for sheets.
Investors wanted clarity on return timelines. Got a philosophy lecture instead. The man who made seventeen billion dollars convincing your aunt to argue about vaccines in a comment section now needs you to believe he's got a coherent enterprise AI strategy. He's asking you to price in revenue from products he might sell to customers he might find using infrastructure he's still building.
Every tech CEO runs this play. Spend cash on capital expenditures. Call it investment. Watch the stock price decide whether you're a visionary or an idiot. Zuckerberg's betting investors will ignore the fact that keeping the AI means competing with Google in search and ChatGPT in everything else, while selling it means competing with Azure and AWS in a margin-crushing knife fight he's a decade late to join.
The trade-off isn't what to sell versus what to keep. It's which failure mode loses less money.
Photo by Shutter Speed on Unsplash
Leave a Comment