Simile bags $200M at $2B, five months after $100M Series A, to predict what humans will do before AI gets it wrong

Simile has raised $200 million in Series B funding, reaching a $2 billion valuation just five months after closing its $100 million Series A.
Founded by Stanford PhD graduate Joon Sung Park alongside Michael Bernstein, Percy Liang and Lainie Yallen, the company develops foundation models that simulate human behaviour for businesses.
Its models predict human behaviour for CVS Health, Wealthfront, Deloitte and Gallup before they act.

Most AI startups want to help people make things faster. Simile wants to tell them whether those things will work before they exist. The Stanford spinout and Palo Alto startup has closed $200 million in Series B funding at a $2 billion valuation, five months after a $100 million Series A that launched the company into the public eye.

The round was led by Greenoaks, the San Francisco-based long-term growth fund whose portfolio includes Stripe, Figma, Nubank, and Brex, with Index Ventures doubling down on its stake and Hanabi, Bain Capital Ventures, A*, Factory, and CVS Health Ventures all returning. Definition joined as a new investor.

“With AI, anyone can now create a product, a campaign, a policy, or a script… The hard question is no longer whether you can create something, but rather what to create, for whom, and how to bring it to life. Our mission is to simulate all eight billion people on earth, accurately and honestly,” said Joon Sung Park, co-founder and CEO.

From generative AI to behavioural AI

The global market research and consumer insights industry is valued at approximately $80 billion and is being disrupted by AI-driven synthetic research tools. Most of that spend still goes toward tools that generate content. Simile is betting enterprises will pay just as much, or more, to know in advance whether that content lands.

Joon Sung Park, a Stanford PhD graduate who completed his doctorate advised by Michael Bernstein and Percy Liang, founded Simile after conducting research in which 25 AI agents lived simulated lives in a virtual town called Smallville, holding conversations and forming routines without human scripting. That paper – Generative Agents: Interactive Simulacra of Human Behavior – won the Best Paper Award at UIST 2023 and became the foundation for a company that envisions behavioural prediction as the next frontier of enterprise AI spend.

The startup, founded in late 2025 and launched publicly in February 2026, works by training foundation models that simulate how customers, patients, employees or citizens are likely to respond to a given decision, then running the scenario at scale before the company commits real budget. For example, a healthcare insurer could simulate patients’ reactions to a new benefits plan before rolling it out nationally, rather than finding out through a costly pilot.

The models have run tens of millions of simulations for Fortune 100 enterprises, and Simile recently launched a product that lets organisations act on those predictions directly rather than treating them as background research. CVS Health, Wealthfront, Deloitte and Gallup already use the platform.

Simile was co-founded by Joon Sung Park (CEO), Michael Bernstein (Stanford HCI professor), Percy Liang (director of Stanford’s Center for Research on Foundation Models), and Lainie Yallen. All four co-founders worked together at Stanford before spinning the company out. The Series A also drew notable angel investors including former Tesla AI head Andrej Karpathy, World Labs CEO Fei-Fei Li, and Adam D’Angelo.

The next enterprise frontier

Its closest rivals are chasing a different problem. Glean raised $260 million at a $4.6 billion valuation to help employees search enterprise knowledge. Harvey raised $200 million at an $11 billion valuation to automate legal workflows, and Writer is valued at $1.9 billion for its enterprise generative AI platform.

All three help organisations produce content or answers faster; Simile is wagering that knowing in advance whether that content works is the bigger prize.

Revenue has grown fivefold since Simile’s launch in February 2026, and headcount has expanded from a handful of researchers to more than 50 people globally. The company has also trained a confidence model that scores the reliability of each simulation, a direct answer to the standard criticism of synthetic-user research: that simulated humans are only as good as their calibration.

The new capital will fund training for Simile’s core foundation models, expand its simulation compute infrastructure, and grow commercial engineering teams across healthcare, financial services, and media. It brings Simile’s total funding to more than $300 million in under 6 months, one of the steepest early funding trajectories among enterprise AI startups in 2026, and a signal that behavioural AI is graduating from academic curiosity to commercial infrastructure.