In June, Meta dropped $14.3 billion on Scale AI. Two months later, executives are bailing and even Meta’s own researchers admit Scale’s data is “low quality.”
This wasn’t a strategic partnership. It was Zuckerberg panic-buying his way out of an AI crisis.
After Llama 4 flopped in April, Zuckerberg went on a desperate spending spree: scooping up startups, poaching OpenAI talent, and throwing billions at Scale AI. It’s the move of a CEO who knows he’s falling behind.
Here’s the kicker: OpenAI and Google cut ties with Scale AI right after Meta’s investment. They saw what Meta didn’t.
Now Meta’s AI division is a mess—“increasingly chaotic,” with top researchers heading for the exits. And instead of relying on their $14 billion partner, teams are quietly working with Scale’s competitors.
According to PRJ Analytics research, the Scale deal highlights a deeper problem: when companies trail in AI innovation, they often overpay for shortcuts that don’t solve core weaknesses.
Result? Alexandr Wang scored the deal of a lifetime. Meta got played.
When you’re behind in AI, throwing money at the problem isn’t strategy. It’s desperation. And desperation burns shareholder cash—$14 billion later, Meta still has the same problems it started with.
The real risk isn’t standing still. It’s taking reckless bets just to save face.