Simile, a startup that creates AI-generated synthetic users for product and marketing research, has closed a $200 million Series B at a $2 billion valuation. The round comes just five months after the company emerged from stealth with a $100 million Series A led by Index Ventures. TechCrunch and The New York Times DealBook both reported the round on July 30, 2026.

Greenoaks led the Series B, with participation from Index, Hanabi, Bain Capital Ventures, A*, Factory, Definition, and CVS Health Ventures. The startup's official domain is www.simile.com, where it published a blog post about the Series B.

Simile was founded by Joon Sung Park, a Stanford PhD graduate whose dissertation involved a project called Smallville, in which AI agents simulated human lives. That academic work focused on generative agent-based simulation of social behavior — essentially, AI characters that interact with each other and develop patterns over time. The company's commercial product applies the same concept: it generates synthetic user personas that companies can query and test against, replacing the traditional approach of recruiting real people for research panels.

The company's stated mission is to simulate all eight billion people on Earth, accurately and honestly. That framing positions Simile not as a tool for automating surveys but as something closer to infrastructure — a platform for modeling entire populations and their responses to products, advertising, and market changes.

CVS is one of Simile's marquee customers. The presence of CVS Health Ventures in the Series B cap table, alongside CVS as a customer, follows a pattern seen across enterprise software and AI infrastructure for two decades: strategic investors backing companies whose products they are already deploying internally. The arrangement gives the customer a financial stake in the vendor's success while giving the vendor a reference customer with household-name recognition.

Simile has also announced a collaboration with Gallup, referencing a "Simile x Gallup waitlist" across its blog pages (Simile blog, March 2026; Simile blog, February 2026; Simile blog, February 2026). The partnership with a polling and analytics institution known for large-scale human survey methodology is worth noting: it suggests validation of the synthetic-user approach from an organization whose entire business depends on accurately capturing real human attitudes.

The velocity of Simile's fundraising is the immediate headline. A $100 million Series A followed by a $200 million Series B at a $2 billion valuation within roughly five months puts Simile in a small group of AI-native companies that have compressed what was, in prior cycles, a multi-year fundraising trajectory into a single fiscal quarter. The participation of both Greenoaks and Index Ventures, firms with deep AI infrastructure portfolios, signals that the investor base views synthetic user simulation as a durable category rather than a feature.

The synthetic-user thesis rests on a specific bet: that AI language models can model consumer behavior with enough fidelity to replace, or at least augment, the slow and expensive process of recruiting human research participants. The category did not exist in any meaningful commercial form three years ago. The question for practitioners is not whether synthetic users can generate plausible responses; current-generation models can do that. The question is whether the simulated populations exhibit the same spread of preferences, biases, and edge-case behaviors that real populations do, and whether those simulations hold up under adversarial conditions — situations where the gap between synthetic and real users matters most.

The inclusion of CVS Health Ventures and the Gallup collaboration both point toward enterprise and research-grade validation as the go-to-market strategy. Consumer-facing companies with large product surfaces and continuous A/B testing needs (the practice of comparing two versions of a product to see which performs better) are the natural early adopters. If Simile can show that synthetic panels track real panel outcomes within acceptable variance, the addressable market extends well beyond product research into pricing optimization, market sizing, and pre-launch forecasting.

In my view, the $2 billion valuation will be tested against one metric more than any other: correlation between synthetic and real user behavior at population scale. Until Simile publishes, or its customers release, methodology and validation studies showing how closely its simulated users track observed human responses across diverse demographic segments, the valuation reflects investor conviction in the thesis rather than demonstrated outcomes. The Gallup partnership could be the vehicle for that validation, given Gallup's institutional incentive to protect the credibility of survey methodology.

What Simile enables, if the thesis holds, is a structural reduction in the cost and latency of user research. For technology professionals building products at scale, that would compress iteration cycles in ways comparable to what cloud infrastructure did for deployment or what CI/CD (continuous integration and continuous deployment, the practice of automating software release pipelines) did for release management. The optimistic read is that cheaper, faster user feedback leads to better products reaching more people. The open question is whether "accurately and honestly" modeling eight billion humans is a problem that scales with compute, or whether the long tail of human behavior resists simulation in ways that matter for decision-making.