Meta Deal Opens Opportunity for Surge
Surge AI
The Meta deal turned Scale from a neutral infrastructure vendor into a strategically conflicted supplier, and that matters because frontier labs hand over sensitive prompts, model outputs, and evaluation workflows to these vendors. Once Meta bought 49% of Scale in June 2025 and hired away CEO Alexandr Wang, major customers including OpenAI and Google pulled back, which opened a lane for Surge to win work on the simpler pitch of being independent, high quality, and focused on human data rather than platform bundling.
-
Scale had real leverage before this shock. It reached $1.5B ARR in 2024 and sold a bundled stack of labeling, data management, evaluation, and deployment tools, which let it discount RLHF work to hold accounts. That same breadth made neutrality more important once Meta became an owner.
-
Surge is built for the opposite positioning. It runs a managed marketplace with about 50,000 vetted contractors, lets engineers specify narrow skill filters like language plus domain expertise, and monitors output with agreement scores and worker trust ratings. That makes the switch feel like a quality upgrade, not just a vendor change.
-
The reshuffle did not just help Surge. Other independent vendors also saw work reenter the market, which shows this was a trust event across the category, not a one off account loss. In practice, labs still use outside vendors for second opinions, red teaming, and specialized participant pools even when they have internal annotators.
Going forward, neutrality becomes a product feature in human data. Surge is well positioned if it can turn this moment into long term workflows in evals, red teaming, and safety testing, while Scale faces pressure to prove that a vendor tied to one major lab can still serve the rest of the market without conflicts.