Mercor, a relatively unknown startup based in San Francisco, has quietly reached a $2 billion annualized revenue run rate by connecting human contractors with AI labs. The company, just three years old, processed over $2 million in daily payouts to workers who help train AI models through tasks like data labeling and reinforcement learning from human feedback (RLHF).

In the first half of 2026, Mercor generated $614 million in gross revenue, scaling from a $1 billion run rate to $2 billion within four months. This jump represents a 70% increase compared to its total revenue for 2025. More than 90% of Mercor’s income comes from three major clients: OpenAI, Anthropic, and Google DeepMind, highlighting a heavy concentration on a few large AI players.

About 30,000 contractors work through Mercor, earning an average of $105 per hour. The platform acts as a middleman, coordinating distributed labor for AI training and evaluation. Founded in 2023 by Brendan Foody, Adarsh Hiremath, and Surya Midha, Mercor has raised $486 million and now holds a valuation of $10 billion.

Why Blockchain Projects See Opportunity

While Mercor itself doesn’t operate on blockchain or offer tokens, its model highlights a key challenge for decentralized networks: managing large-scale human labor with minimal middleman fees. Several crypto protocols are developing marketplaces aimed at replacing centralized platforms like Mercor. These projects seek to use token incentives and transparent pricing to coordinate contributors and reduce costs.

The idea is simple. If a startup can coordinate 30,000 contractors, a decentralized protocol could do the same, but with fewer fees and open participation. This approach could attract contractors and AI labs alike by offering competitive pay and more efficient service.

Mercor’s financials provide concrete benchmarks for decentralized AI efforts, proving that the market for human-in-the-loop training is substantial and growing rapidly. Major AI developers are clearly willing to invest heavily in securing reliable human feedback.

The $105 hourly wage sets a reference point. Decentralized labor platforms that can deliver similar quality at lower cost by cutting out middlemen have a real chance to gain market share.

This material is informational and does not constitute financial advice.