42AI raises $62.5 million for behavior model

The Israeli startup says its system can estimate how people may act in new situations and is already being used by financial brokerages.

42AI has emerged from stealth with a striking claim and a sizable funding round. The Israeli startup says it is building a “behavioral world model” that can estimate how people are likely to act in new situations, and it has raised $62.5 million in seed funding. The company says the system is already in use at a small number of financial brokerages, where it helps forecast how traders may react in volatile markets. That combination makes the launch notable — and puts questions about profiling, privacy and misuse front and center.

42AI says its model estimates decisions, not just past behavior

42AI says its approach starts from the idea that behavior is probabilistic rather than fixed. CEO and co-founder Amit Gal-Or argues that people carry underlying psychological patterns that continue to shape decisions even as circumstances change.

Co-founder and CTO Ehud Ezra described the model as a system that can show the probability of a person making a certain decision. In his view, people may change over time, but some tendencies still appear in predictable ways.

To build those predictions, the startup says it looks at how people respond to events that may seem unrelated to the final decision. For an investor or trader, that can include KYC checks, onboarding, support calls, money transfers and stock trades. The company also says it examines signals such as sentiment, voice, and the number of interactions a customer has with support or sales teams.

Those observations are turned into mathematical representations of traits and fed into a transformer-based world model. 42AI says it also uses a separate language model to generate explanations for its predictions. As IT-PUB News notes, the company is pitching something broader than standard analytics based only on historical actions.

Ezra said the system is trained to preserve a person’s underlying tendencies during simulations. If the model starts treating someone who is usually risk-averse as a risk-taker, he said, the training target is considered off.

Brokerages are the first place 42AI is deploying the system

The startup says its first live deployments are in financial brokerages. Gal-Or said 42AI is working with 10 large brokerages that use the system to manage liquidity, margin requirements and hedging exposure.

According to the company, the model analyzes existing customer records, including transactions and support calls. Gal-Or said it does not need personally identifiable information and that “the model never knows who the person is.”

Even so, the company’s description suggests the system still builds individual behavioral profiles from linked customer activity. That distinction matters. The model is not simply processing anonymous market data; it is trying to infer how specific people are likely to behave from the traces they leave behind.

42AI says the tool is useful in fast-moving trading environments because it helps predict how individual traders may respond under stress or volatility. The company also says it is onboarding more customers.

42AI's accuracy claims come with clear caveats

42AI says its model reaches 98% accuracy in predicting market trades, but Gal-Or said that does not mean every trade is predicted correctly. Instead, he said, the figure refers more to whether trading flows fall within a predicted range.

The company also says the model consistently achieves an F1 score of 0.8 to 0.85 for predicting trading direction and volume. In simple terms, that measures how well a model balances correct predictions against false alarms and missed events.

Gal-Or also acknowledged the system’s limits. He said there is a natural ceiling because no model can capture everything a person experiences. Ezra added that accuracy drops when the system misses important context, such as advice from a friend or colleague that changes a trader’s decision.

That caveat matters because 42AI is presenting the system as a tool for probabilities, not certainty. Still, the company is making a substantial claim: that enough personal and behavioral data can reveal patterns strong enough to guide decisions in finance.

The funding came from founders and Infinity Group

42AI says the $62.5 million seed round came from the founders and Infinity Group over three years, and was not entirely raised from outside investors. Gal-Or said Infinity was founded by his father, Amir Gal-Or, who is also 42AI’s chairman and a co-founder.

The company says it has already generated more than “seven figures” in revenue, though it did not provide a precise number. Customers pay either a flat fee or a percentage of the value the platform helps them gain. Gal-Or also claimed that one brokerage generated about $30 million in extra value through dynamic hedging.

42AI now plans to raise more money. Beyond finance, it is exploring possible uses in gaming, fraud prevention and cybersecurity. It also wants, eventually, to help robots and AI agents understand people. But Ezra said the model would need to be trained on the rules and context of each new domain, so any expansion beyond finance is not immediate.

Questions about privacy and misuse are still unresolved

The most sensitive part of 42AI’s pitch may be less about the model itself than what customers could do with it. If a brokerage can anticipate how customers may behave, that could create an advantage that does not belong to those customers. That is why the company’s claims raise immediate questions about profiling and misuse.

Ezra called that a valid concern when asked about it. The founders pointed to regulations, monitoring and an internal ethics committee, but they did not explain what would prevent customers from using the system to gain an edge over the people being profiled.

Gal-Or said he believes the company is part of the solution rather than the problem. For now, 42AI is trying to turn an ambitious idea — predicting human behavior — into a commercial product, starting with finance and carrying unresolved ethical risks with it.