Five months after Meta didle the startup with billions of dollars and named it one of the hottest operators in training data, the glow around Scale AI has begun to fade. Inside the company, it feels like the atmosphere has gone from triumphant to tense, from “we’re the future of AI infrastructure” to “why does everything suddenly feel so difficult? Workers say the Meta deal did not deliver the stability or prestige they had hoped for. Ergo, it just threw fuel on to the pressure, audience pay resentment and slapped a big target on Scale’s back that competitors are now driving toward.
Speak to people near the company and you hear the same refrain: The Meta partnership was supposed to verify Scale’s supremacy, not reveal its fragility. But big deals change expectations. The company that for so long felt scrappy, and unstoppable, suddenly finds itself being assessed by Big Tech metrics. There’s a sense among employees that pay packages appear less competitive than billed, growth projections feel more uncertain and that the culture has adopted an edge as leadership pushes to keep up with the momentum suggested by the Meta deal.
What’s hitting the team hardest, though, is that it feels like market around them is moving faster than expected. Rather than locking in Scale’s edge, the Meta deal brought a ton of attention to it from rivals sniffing opportunity. Start-ups that specialize in synthetic data, annotation automation and cost-saving labeling pipelines are pitching themselves directly to some of Scale’s largest clients. Big enterprise customers who used to write Scale purchase orders by default are now taking meetings with competitors just to see how low prices can go. That loyalty thing which made Scale so much-cooler-than-thou in the early days is not a given any more, and employees feel under some pressure.
Compensation is the biggest source of frustration inside the company. Some people said they hoped Meta-era pay would offer something close to a Meta package or at least the equivalent of a company that is valued as AI’s beating heart. Instead, internal inequalities grew more apparent. Long-time workers were earning less than the new hires. Contractors felt the squeeze as expectations were raised. And across teams, people asked in a hushed tone the same question: if the company is winning billions in deals, why does it not feel like it?
The culture, once fiercely intense in a “we’re building something big” way, now feels fiercely intense in a “we have to defend what we built” way. Employees say strategy meetings are tense, with priorities changing frequently and more and more concern about whether the next year will be a matter of catching up or a showcase for innovation. And innovation is at the heart of the problem. The industry Scale leads is evolving quickly — more automation, cheaper alternatives and clients who ask how much of their human-labeling budget they can cut as AI models get better.
The irony is that Scale AI remains in a phenomenally strong position. It is deeply embedded in government contracts, enterprise AI pipelines, foundation model training and the unseen plumbing of machine learning. But the company is no longer existing in a vacuum. The competition is real now. The expectations are enormous. And internally, the toll of trying to sustain explosive growth is catching up with a team that once appeared energized by it.
Meta’s $14 billion deal did not break Scale — but it pushed the company off a startup cliff before it was prepared. There is an undercurrent of “growing pains” that takes place across everything, employees say: reorganizations, rushed product shifts, narrowing performance expectations and persistent questions about whether the company can keep growing at the same pace as the Valley’s mania for A.I. plugins.
What’s taking place now is a traditional post-deal identity crisis. Scale still has the talent, the client list, and the technical advantage in training data. But inside the office, there’s been a change in mood. The deal intended to be a moment of triumph for the company is now defining this time of reckoning — the realization that staying on top demands more than momentum. It needs coherence and stability; it needs a workforce that feels rewarded, not stretched.
Scale AI will either rebound or slip on what it does next. But for the first time in a long time, the company that powered the AI boom is starting to grapple with what it’s like to be on the other side of it — and inside Scale, you can feel that turn every day.
