Mercor reported $614 million in revenue for the first half of 2026, with 90% of that coming from AI model firms. The company's major clients include OpenAI and Anthropic, underscoring the massive demand for data labeling and training services — a market that decentralized crypto projects are also trying to capture.
The numbers
Mercor's $614 million haul for the first six months of 2026 is a clear sign of how fast the AI training pipeline is growing. The company gets nine out of every ten dollars from AI model companies, meaning the rest of its business — whatever that is — barely moves the needle. OpenAI and Anthropic are the headline clients, but the fact that Mercor can pull in over half a billion dollars in six months suggests the whole sector is scaling hard.
The article frames Mercor's success as validation for the AI training market that decentralized crypto projects aim to serve. If a traditional company can rake in that kind of revenue from data labeling and model training, the argument goes, then there's room for blockchain-based alternatives that promise better data provenance, token incentives, or decentralized compute. The question is whether those crypto projects can actually deliver at Mercor's scale — or if they'll stay niche.
The gold rush
The piece calls Mercor's trajectory an 'infrastructure gold rush' that crypto investors should watch. It's a blunt assessment: the money is flowing into AI training infrastructure, and the winners so far are centralized players like Mercor. But the market is still early, and the crypto side hasn't fully shown its hand. For now, Mercor's numbers are a benchmark — a reminder that the underlying demand is real, even if the decentralized solutions are still finding their footing.
