If this holds, Mercor will have doubled its valuation in roughly nine months — a pace that says more about investor appetite for anything touching AI training data than it does about any newly disclosed revenue milestone. The company built its name matching vetted human experts (engineers, doctors, lawyers) to AI labs hungry for high-quality RLHF and evaluation data, a business model that’s become one of the hottest corners of the data economy as foundation model shops fight over scarce, expensive human judgment.
Mercor’s climb mirrors Scale AI’s and Surge AI’s — all riding the same thesis that as pretraining data dries up, the money moves downstream to labeling, evaluation, and expert feedback. But a jump from $10 billion to $20 billion on unverified terms deserves the standard skepticism: markups like this often reflect FOMO among a handful of AI-adjacent funds rather than audited multiples of revenue.
Doubling a valuation in nine months isn’t a growth story yet — it’s a bet that the labeling gold rush still has room to run.
Worth watching: who’s actually writing the check, whether Mercor discloses real ARR to back the number, and whether rivals in the human-data-labeling space see similar markups follow.
A new $20 billion valuation would be a giant step up from the $10 billion valuation it reached in October.
