For three years, an Israeli startup called 42AI has been building what it calls a "behavioral world model": a system that does not simply extrapolate from what a person has done before, but tries to infer the underlying tendencies that will shape what they do next.
On Monday the company came out of stealth with $62.5 million in seed funding and a customer list of ten large financial brokerages, where its model is used to anticipate how individual traders will react to volatile markets — for liquidity management, margin requirements and hedging exposure.
The pitch is deliberately grand. "Behavior is not deterministic; it's constantly probabilistic," CEO and co-founder Amit Gal-Or told TechCrunch. His co-founder and CTO, Ehud Ezra, put it more starkly: "In a world model, every individual is the world."
The company says it builds those individual worlds from a wide range of signals: psych assessments and behavioral scenarios run on thousands of people, government and biobank datasets covering more than two million more, and customer data that has since grown to cover almost 30 million people. In a brokerage deployment that means KYC steps, onboarding, support calls — including sentiment and voice — money transfers, and then trading behavior read against market conditions.
42AI says the system does not need personally identifiable information and that "the model never knows who the person is." It does, however, build individual behavioral profiles from linked customer activity, and the distinction between a de-identified profile and a profile of a specific person is doing a lot of work in that sentence.
The performance claims deserve the same care. The 98 percent accuracy figure the company cites is a measure of whether trading flows fall inside a predicted range, not whether individual trades were called correctly, Gal-Or acknowledged. For predicting the direction and volume of trades, he said the model holds an F1 score of 0.8 to 0.85. The system also degrades when it misses context — a trader acting on a friend's advice, for instance.
Commercially, the numbers are early but real: more than "seven figures" in revenue, the founders say, with customers paying either a flat fee or a share of the value created. One brokerage generated roughly $30 million in additional value through dynamic hedging, according to the company. The seed round came from the founders and Infinity Group over three years and was not entirely external; Infinity was founded by Gal-Or's father, Amir Gal-Or, who is also 42AI's chairman and a co-founder.
The obvious question is what stops a brokerage that can predict its customers' next move from trading against them. Ezra called it "a valid concern." The founders pointed to regulation, monitoring and an internal ethics committee, but did not explain what specifically would prevent a customer of theirs from using the model to gain an advantage over the people it profiles. Gal-Or's answer was that the company is "part of the solution."
42AI is now looking beyond finance, at gaming, fraud prevention, cybersecurity, and eventually robots and AI agents that need to model the humans around them. Each new domain needs its own rules and context, Ezra said. The company plans to raise more money.
