Simile AI startup raises $200M at $2B valuation to simulate consumers with digital twins

Simile Secures $200M Series B Funding at $2B Valuation

Simile, a Palo Alto-based AI startup, has closed a $200 million Series B funding round at a $2 billion valuation, marking a meteoric rise just five months after emerging from stealth with a $100 million Series A 1

1

. Greenoaks led the latest round, with participation from Index Ventures, Hanabi, Bain Capital Ventures, A*, Factory, Definition, and CVS Health Ventures 1

1

. The rapid fundraising trajectory positions Simile among the fastest-growing members of the AI unicorn club, reflecting investor appetite for tools that promise to transform how companies conduct market research.

AI Agentic Twins Transform Market Research

The company offers AI agentic twins—AI-generated replicas of real consumers that organizations can survey instead of conducting traditional market research 3

3

. These synthetic users enable companies to test reactions to new products, evaluate brand perception, and predict consumer behavior faster and cheaper than conventional methods 2

2

. Simile's technology allows clients to simulate human behavior by tapping into its growing database of AI agents, with the stated goal of modeling all 8 billion people on Earth 4

4

.

Founded by Joon Sung Park, a 32-year-old Stanford PhD graduate, Simile emerged from his 2023 research paper on AI agents that simulate human behavior in interactive environments 2

2

. Park's dissertation project, called "Smallville," featured AI agents carrying on simulated human lives, right down to holding parties 1

1

. His co-founders include Stanford computer science professors Michael Bernstein and Percy Liang, who helped coin the term "foundation model" 2

2

.

CVS Health and Major Clients Deploy AI Digital Twin Service

CVS Health became one of Simile's marquee customers, using 400,000 agentic twins last fall to address prescription adherence challenges 2

2

. Millions of patients never fill their prescriptions or fail to finish them, and CVS Health sought to understand how to improve medication compliance 4

4

. Deloitte and Wealthfront have also joined Simile's client roster, leveraging AI-driven simulation for product research and consumer insights 3

3

.

The company built its initial model through two-hour interviews with 1,000 participants selected to represent the broader population 2

2

. Simile then partnered with Gallup to supplement its data with input from a few million people worldwide 2

2

. According to the company's head of product, Mihika Kapoor, Simile's responses now achieve 85 percent to 99 percent accuracy, depending on the scenario and population surveyed 2

2

.

Pricing Model and Confidence Model Development

Simile generates revenue based partly on the number of agents it models for customers, ranging from hundreds of thousands to millions of agents 2

2

. "The larger the panel, the higher the price," Park explained 3

3

. The Series B funding will help develop what the company calls "a first-of-its-kind confidence model that predicts the accuracy of every simulation" 4

4

. This addresses a critical question for enterprise buyers: whether AI stand-ins truly capture how humans behave.

Contested Premise and Growing Competition

The premise remains contested, with critics questioning whether synthetic data can genuinely simulate human behavior driven by both emotions and reason 1

1

. The accuracy figures are Simile's own, and the company relies on real human data to manufacture synthetic users 3

3

. Synthetic data also carries risks, as it can leak sensitive signals through inference and replicate historical biases, embedding them behind layers of abstraction that make them harder to detect and audit 4

4

.

Simile faces competition from startups including Helm, Artificial Societies, Rehearsals, and Aaru, which raised a Series A in December at a $1 billion valuation 1

1

. Park envisions expanding beyond survey-based market research to develop "dynamic simulations" where AI agents are measured over time, such as testing products over the course of a week 2

2

. Financial institutions are already replacing live customers with AI-generated stand-ins to test new products like credit cards, with U.S. Bank, JPMorganChase, NatWest, Monzo, and Santander using synthetic data ecosystems 4

4

.