Simile just raised $200M to simulate all eight billion people

Signals Inbox·July 30, 2026·World Models

Simile raised $200 million at a $2 billion valuation to build AI versions of human populations. The bigger signal is the speed: it raised $100 million only five months earlier, meaning investors have put roughly $300 million behind the idea that companies can test decisions on synthetic people instead of waiting for real surveys.

Q1What actually happened?

Simile officially announced a $200 million Series B led by Greenoaks at a $2 billion valuation. Index Ventures, CVS Health Ventures, Bain Capital Ventures, Hanabi, A*, and Definition also participated. The company says its mission is to simulate all eight billion people on Earth accurately.

Q2Why is the timing so striking?

Because Simile raised a $100 million Series A only five months earlier. It has now collected roughly $300 million in half a year and reportedly doubled its valuation. That is an unusually aggressive funding curve for a company that only recently emerged from stealth.

Q3What does Simile actually simulate?

It creates AI twins based on data from real people. Companies can ask these agents questions, test messages, compare products, or simulate how groups may react to a decision. CVS has used twins calibrated with 2.9 million answers from more than 400,000 people, while Gallup has worked with over 1,000 digital twins.

Q4Does it really have eight billion digital people?

Not yet. Eight billion is the mission, not the current user count. Simile is trying to build models that can generate believable people and populations at any scale. The tension is whether a smaller amount of real human data can produce agents accurate enough to represent people who were never interviewed.

Q5Why would companies use fake customers?

Traditional surveys and focus groups can take weeks or months, and every new question can cost more money. AI twins can be questioned repeatedly and almost instantly. Simile says customers pay from roughly $150,000 to several million dollars per year, so this is already being sold as enterprise infrastructure, not a research demo.

Q6Can synthetic people replace real research?

Not safely on their own. CVS reported up to 95% accuracy when its twins were tested against findings it already knew, but that does not guarantee the same accuracy for new products, elections, crises, or unfamiliar groups. The agents still need fresh human data, careful calibration, and real-world checks or they may simply reproduce convincing mistakes at massive scale.

Q7So what is the real bet?

Simile is betting that predicting people becomes a software problem. If it works, companies could rehearse product launches, policy changes, healthcare messages, and major investments before exposing real people or real money. If it fails, investors may have spent $300 million building an extremely confident focus group made of bots.