Simile AI Just Raised Money to Simulate Every Human on Earth

Simile AI wants to answer a question that sounds almost too big for one company: what if you could test any decision on a stand-in for humanity before you actually make it?

That pitch has pulled in $300 million in under six months and a $2 billion valuation for a company that barely existed a year ago. If you’ve seen the name pop up on TechCrunch, Bloomberg, or your LinkedIn feed and wondered what Simile AI actually does, who’s behind it, and whether the hype holds up, this guide covers all of it.

What Is Simile AI

Simile AI, Inc. is a Palo Alto startup building a foundation model trained to predict how real people respond to products, messages, policies, and decisions. Instead of running a traditional focus group or survey, a company can ask Simile’s AI agents the same questions and get back a modeled answer in minutes.

The company calls its approach “AI simulation of society.” Each simulated person, which Simile refers to internally as a “simile,” is built from structured interviews, past choices, and behavioral data tied to a real individual rather than a generic demographic description like “35-year-old suburban mom.”

Simile’s own language is dramatic on purpose. Its site opens with a comparison to pilots training on simulators instead of real passengers, and surgeons practicing before they operate on real people. The argument is that big corporate and policy decisions get pushed straight into the real world without anything close to that kind of rehearsal. This shift aligns with a broader industry trend toward deploying autonomous enterprise AI agents to automate complex strategic and operational workflows.

Simile AI Company Background and Founding Story

Simile grew directly out of Stanford research rather than a garage or an accelerator. The company’s roots trace back to a 2023 paper called “Generative Agents: Interactive Simulacra of Human Behavior,” which introduced a virtual town nicknamed Smallville.

In that project, 25 AI-driven characters lived out simulated daily routines. They woke up, made coffee, held conversations, formed opinions, and in one memorable case, one of them decided to run for mayor. The paper won Best Paper at UIST 2023 and became one of the most cited AI papers of the year, sparking wide interest in what researchers call agentic AI.

The lead author, Joon Sung Park, had been exploring a specific idea since around 2020, right as GPT-3 was showing that large language models could do things they were never explicitly trained to do. Park noticed that these models had absorbed countless small patterns of human behavior from the text they were trained on. Ask the right question, and a model could produce a surprisingly specific guess at what a particular kind of person would do next.

Park built on that idea with his Stanford advisors, publishing an earlier related project called Social Simulacra in 2022 before Smallville made headlines. He later co-authored a 2024 follow-up study simulating 1,052 real people, replicating their survey responses and behavior with roughly 85% accuracy two weeks after the original interviews.

Simile AI, Inc. was founded in 2025 to turn that research into a commercial product. It spent about seven months in stealth before going public in February 2026.

Who Founded Simile AI

Simile AI’s founders combine two rare things: deep academic credibility in AI research and a background actually turning research into shipped products. The four co-founders are:

• Joon Sung Park, CEO. Park is the Stanford PhD graduate behind the Smallville and Generative Agents research. Multiple profiles describe him as an oil painter turned entrepreneur, which is an unusual detail for an AI founder but one that keeps coming up in interviews about his creative approach to the mission. He frames today’s large language models as the “CPU of intelligence,” good at problems with clean right answers, and describes Simile as building the “GPU of intelligence” instead, a system meant to capture the messiness of human values and preferences.
• Michael Bernstein, Chief Product Officer. Bernstein is a Stanford professor known for work on human-computer interaction and social computing. He co-authored the influential 2015 ImageNet Challenge paper, a project widely credited with helping kick off the modern deep learning boom, alongside Andrej Karpathy and Fei-Fei Li, who are both now Simile investors.
• Percy Liang, Chief Scientist. Liang directs Stanford’s Center for Research on Foundation Models and co-authored the 2021 paper that introduced the term “foundation model” into the AI vocabulary. He also created HELM, a widely used benchmark for evaluating large language models. When Simile calls its product a foundation model for human behavior, that phrase comes from the person who effectively coined it.
• Lainie Yallen, Chief Customer Officer. Yallen is the team’s commercial operator. She studied business at McGill, started her career at Boston Consulting Group, and helped scale companies including Hebbia and Valence before joining Simile to build out sales and customer relationships.

Bain Capital Ventures, one of Simile’s investors, has described this combination directly: Park brings vision, Liang brings deep model expertise, Bernstein turns research into usable product, and Yallen brings the commercial horsepower to close enterprise deals.

Simile AI Funding and Valuation Timeline

Simile’s funding history moved unusually fast, even by AI startup standards.

• February 2026: $100 million Series A. Simile raised $100 million in new funding to build a model designed to predict human behavior, including helping companies anticipate questions likely to be asked on earnings calls. The financing was led by Index Ventures, with participation from Bain Capital Ventures, A* and Hanabi Capital. Notable AI figures, including Fei-Fei Li and Andrej Karpathy, also contributed to the round. Simile did not disclose a valuation for this round: AI Startup Simile Secures Funding to Help Companies Anticipate Customer Choices – Bloomberg +2.
• July 2026: $200 million Series B at a $2 billion valuation. Just five months later, Simile closed a much larger round. Greenoaks led the B round, with participation from Index, Hanabi, Bain Capital Ventures, A*, Factory, Definition, and CVS Health Ventures. CVS is also one of Simile’s marquee customers.

That brought Simile’s total funding to roughly $300 million across two rounds, giving it official unicorn status. Some data providers put the cumulative total slightly higher, around $310 million, once smaller investor allocations are factored in.

Here’s a quick side-by-side of the two rounds:

For context on how fast that is, most startups take years between a Series A and a $2 billion valuation. Simile did it in less than half a year, which tells you a lot about how much appetite investors currently have for anything touching AI agents and simulation.

Is Simile AI Publicly Traded

No. Simile AI is a private company and does not have a stock ticker. There is no way to buy shares of Simile on a public exchange like the NYSE or Nasdaq.

If you’ve searched for a “Simile AI stock” or “Simile AI ticker symbol,” you’re not going to find one, because the company hasn’t filed for an IPO and has given no public indication that one is coming soon. The only way to hold equity in Simile right now is through venture capital investment, employee stock grants, or secondary markets that trade private shares among accredited investors, none of which are accessible to typical retail investors.

What Simile AI Actually Sells

Simile’s core product is a simulation platform built around what it calls a foundation model for human behavior. Companies use it to run large numbers of virtual interviews or tests against modeled populations instead of recruiting real participants for every question.

According to reporting on Simile’s process, the company built its initial model using two-hour interviews with 1,000 carefully selected participants meant to represent the broader population. It then tested the model’s predicted answers against how those same real people actually responded, checking for accuracy before scaling up.

Simile has since partnered with Gallup and other data partners to expand its pool of consented interview data to a few million people worldwide. The company reports accuracy rates between 85% and 99%, depending on the scenario and population being tested, according to its head of product.

Reported use cases include:

• Rehearsing corporate earnings calls by predicting the questions analysts are likely to ask
• Modeling potential litigation outcomes before a case goes to trial
• Testing marketing messages and product concepts before launch
• Screening research questions and customer journey ideas before running a real study or pilot
• Helping healthcare and insurance companies test things like medication adherence programs

Simile prices its product based on the number of simulated agents a customer runs, and larger simulated panels cost more. The company has said it wants to eventually move beyond one-time surveys toward “dynamic simulations” that track how a modeled population’s opinion might shift over the course of days or weeks, not just at a single point in time.

Simile AI’s Customers

Simile has been fairly open about who is testing its platform, which is unusual for such a young company. Named customers and partners reported across multiple sources include:

• CVS Health, which is both a customer and a Series B investor through CVS Health Ventures
• Wealthfront
• Gallup, the polling and analytics firm
• Suntory, the Japanese beverage company
• Telstra
• Deloitte
• Banco Itaú
• Garnett Station Partners

CVS Health has been the most visible reference customer. The Wall Street Journal covered how CVS is using Simile’s technology to model customer behavior, and CVS has reportedly fed some of its own historical survey data, CRM records, and A/B test results back into Simile’s system for calibration, pushing reproduction accuracy on some known outcomes as high as 95% in internal testing.

How Simile AI Compares to Competitors

Simile is not the only company chasing this idea, and the space has gotten crowded fast. The most frequently mentioned competitors include:

• Aaru. A New York-based startup founded in March 2024 that builds a simulation engine called Lumen. Aaru raised a Series A in December 2025 led by Redpoint Ventures at a headline valuation around $1 billion, though the actual blended valuation across investors landed lower due to a multi-tier deal structure. Aaru’s client list includes Accenture, EY, and Interpublic Group, and the company has drawn attention for using its model to forecast election outcomes.
• Artificial Societies. A Y Combinator-backed company that runs a reported half a million AI personas for research purposes.
• CulturePulse. Another social simulation startup frequently named alongside Simile and Aaru in coverage of this niche.

The bigger picture is that the traditional market research industry, estimated at around $120 billion globally, is watching this space closely. While Simile AI targets human behavior and market simulation, other specialized autonomous systems, such as Bland AI’s in-house voice agent technology, are tackling dedicated customer interaction channels in parallel. Simile’s own team has acknowledged the competitive pressure directly. Because the foundational research behind generative agents was published openly, with thousands of academic citations, any well-funded team can study the same blueprint Simile used to get started. The real test is execution and enterprise sales, not just having the best research pedigree.

Reddit and Critic Reactions to Simile AI

Not everyone is convinced synthetic users are ready to replace real ones, and that skepticism shows up clearly in tech and product management communities online.

Discussions on Reddit’s ProductManagement and machine learning communities have repeatedly raised the same concern: synthetic responses are predictions based on text patterns, not actual observations of a real person doing something. If a synthetic user tool generates a response for a “senior IT manager with 15 years of experience,” it is producing text that statistically matches how people with that description tend to write, not tapping into any specific IT manager’s real, lived opinion.

Some UX researchers have gone further, arguing that using generative AI to stand in for real participants in design research misrepresents what design research is supposed to do in the first place. Critics point to a handful of recurring risks with synthetic panels:

• Answers that look plausible but are simply wrong
• Shallow qualitative depth, since a synthetic respondent has no lived experience to draw from
• Reinforcement of biases already baked into the training data
• Artificially low variability compared to how real people actually differ from one another

Even outlets that covered Simile’s $2 billion valuation approvingly included pushback. One widely cited line from TechCrunch’s own reporting called Simile’s stated goal of eventually simulating “all eight billion people on earth, accurately and honestly” preposterous, on the logic that the entire reason companies do market research is that real people are unpredictable.

Simile’s counter-argument, echoed by its investors, is that the platform isn’t meant to fully replace human research. It’s positioned as a way to screen out bad ideas early and decide which questions are worth the time and cost of a real study, not as a permanent substitute for talking to actual customers.

Simile AI Careers and Company Culture

Simile is actively hiring, and its job postings give a decent window into how the company talks about itself internally. Roles posted through Index Ventures’ startup jobs board and other listings include Applied Research Engineer, Member of Technical Staff, Technical Recruiter, and GTM Recruiter, most based on-site in Palo Alto.

Job listings repeatedly use the same framing found on Simile’s public site, describing the mission as helping companies “preview the effect before triggering the cause” across use cases like earnings call rehearsals, litigation modeling, and policy testing.

A few practical details worth knowing if you’re considering applying:

• Most current roles are listed as full-time and on-site in Palo Alto, California, rather than remote
• Equity grants are offered for eligible roles, subject to board approval
• Benefits mentioned include medical, dental, and vision coverage plus flexible time off
• The company states a 90-day waiting period before reconsidering a candidate who previously applied for the same role
• Recruiting language emphasizes hiring for “Applied AI” talent and describes Simile as positioning itself as a “world-class AI lab” rather than a typical enterprise software company.

Given that Simile only has around 50 total employees as of its most recent profile data, joining now would mean working at a company that is still small relative to its valuation, with a lot of ground still to cover in scaling both the research and the go-to-market side of the business.

Frequently Asked Questions

Is Simile AI the same as ChatGPT or other general AI chatbots?

No. Simile is not a general-purpose chatbot. It’s a specialized platform that models how specific groups of real people would respond to a given scenario, built from structured interview and behavioral data rather than open-ended conversation.

Who is the CEO of Simile AI?

Joon Sung Park is the co-founder and CEO. He led the original Smallville and Generative Agents research at Stanford before starting the company.

How much is Simile AI worth?

Simile was valued at $2 billion following its Series B round in July 2026, based on a $200 million raise led by Greenoaks.

Does Simile AI have a stock ticker?

No. Simile is privately held and has no public stock symbol. There’s currently no confirmed IPO timeline.

What does the word “simile” have to do with AI simulation?

A simile is a figure of speech that compares one thing to another using “like” or “as.” The company’s name plays on that idea, since each of its AI agents is meant to stand in for, or resemble, a specific real person.

The Bottom Line

Simile AI sits at an interesting intersection of legitimate, peer-reviewed research and some genuinely bold marketing claims. The Smallville and Generative Agents work behind it is real, well-documented, and has already influenced how the broader AI field thinks about agent memory and planning. The founding team’s academic pedigree is about as strong as it gets in this space.

At the same time, a $2 billion valuation for a one-year-old company with roughly 50 employees is a bet on execution, not just research quality, especially with well-funded rivals like Aaru chasing the same market. Whether synthetic populations end up meaningfully changing how companies do research, or whether this turns into another AI category that overpromises and underdelivers, is still an open question that will play out over the next few years rather than the next few months.

For now, Simile is one of the clearest examples of Stanford AI research turning into a fast-moving, well-funded startup, and it’s worth watching regardless of which side of the synthetic user debate you land on.