Mercor, a company that hires experts to help train AI models, has raised $750 million in a funding round that values it at $20 billion. It says its revenue is now running at $3 billion a year. A year ago, it was under $500 million.
How fast it grew
"Run rate" means the yearly total if the current pace holds. It is not money already earned. Mercor says it crossed $1 billion in February and $2 billion in June. Its previous round, last year, valued it at $10 billion. The company reports 5 million experts. It also says it survived hacks, but the material gives no details.
What Mercor says about AI and work
Mercor says the United States has more than 800 distinct occupations. Each has tasks that AI models have not yet been tested against. That is the market it sees. The company also says it spends 3x as much on AI computing as on employee salaries, and that this pays for itself by making its team better at their jobs.
Mercor adds that if this ratio holds across the economy, it points to double-digit economic growth over five years. That is a company claim, not a proven result.
What is not known yet
- Who invested in the round
- How much experts are paid
- What the hacks involved
If AI firms keep paying experts to train their models, more skilled workers may find paid side work. More jobs may also be measured against AI.
The takeawayMercor's jump to a $20 billion value shows AI firms are paying heavily for human expertise, though its growth and economy claims come from the company itself.