A profile of Mercor, which pays about $2M daily to ~30K experts training AI models at $95/hour on average, with roles like radiologists earning up to $375/hour
Tens of thousands of professionals are joining tech start-up Mercor to hone technology in their skills — At first Lola thought the job ads were fake.
Context & Ripple Effects
Mercor began as an AI-assisted hiring marketplace, then expanded into domain-expert model training; its earlier profile described a Scale AI rival already hiring specialists at scale (its expert-training marketplace).
The reported payout pace builds on Mercor’s prior disclosure of roughly $1.5 million in daily contractor payments and $500 million in ARR (its earlier contractor-payment scale), indicating that expert labor is central to its operating model rather than a peripheral service.
First-order effects
- Roughly 30,000 experts gain a paid channel to apply professional judgment to AI training, with particularly scarce specialties commanding materially higher rates.
- Mercor must fund and manage a large, recurring contractor pool; payment rates become a direct lever for attracting specialized training capacity.
Second-order effects
- Rival AI-data and contractor platforms face pressure to compete for qualified professionals on pay, task quality, and access to recurring work.
- For model developers, expert feedback becomes a more visible input cost in improving performance on specialized tasks—an instance of the market for domain-expert training work moving beyond general-purpose labeling.
Third-order effects
- If high-value expert work remains essential to model improvement, AI training supply chains may shift toward marketplaces that aggregate and evaluate professional judgment, not just low-cost annotation labor.
- The economics of advanced AI systems will increasingly hinge on the cost of reliably completing useful tasks, including the human expertise required to create and assess training material.
The trend: AI training is becoming a specialized labor market in which verified domain expertise is priced as a strategic input to model quality.