Simile raises $200M Series B at $2B valuation to survey AI 'twins' of consumers

Simile , a year-old Palo Alto startup, has raised over $200 million in a Series B at a $2 billion post-money valuation to build "agentic twins" — AI stand-ins for real consumers that companies can survey instead of running traditional market research. Greenoaks led the round, with Index Ventures doubling down and returning backers Hanabi, Bain Capital Ventures , A*, Factory, and CVS Health Ventures participating. Definition joined as a new investor.

The raise comes fast. The startup raised $100 million less than six months ago, and says revenue has grown 5x since launching five months ago.

Simile's pitch is that it can predict how people behave by simulating them, letting firms test a product or brand faster and cheaper than a conventional study. Its stated mission is to model all eight billion people on earth.

Simile built a foundation model for human behavior that has run tens of millions of simulations for Fortune 100 enterprises. It started with two-hour interviews of 1,000 representative participants, then teamed with Gallup to add input from a few million more. The company claims 85% to 99% accuracy, depending on the group, and has trained a confidence model to predict the accuracy of each simulation.

The customers are real even if the respondents are not. Last autumn, CVS Health used 400,000 agentic twins to work out how to get patients to take their medicine; Deloitte and finance app Wealthfront are clients too. Pricing scales with the crowd. "The larger the panel, the higher the price," said chief executive Joon Sung Park.

Park, 32, built Simile from a 2023 Stanford paper on AI agents that mimic human life. His co-founders include Percy Liang, who helped coin the term "foundation model." The team has expanded from Palo Alto to more than 50 employees worldwide.

The premise is contested. The accuracy figures are Simile's own, and whether AI stand-ins truly capture human behaviour is exactly what critics doubt. Rivals including Helm, Artificial Societies, and Aaru are chasing the same idea.

Simile leans on real human data to manufacture the synthetic kind, on a web already flooding with AI agents. Its promise to model humanity is the sort enterprise buyers will judge on results.