42AI Raises $62.5M to Forecast Trader Behavior During Market Swings

Israeli artificial intelligence startup 42AI unveiled itself today with $62.5 million in funding aimed at developing what the company describes as a behavioral world model. The technology targets brokerages seeking to anticipate how individual traders will respond when markets experience volatility.

The company's software calculates the probability of different choices a person might make under specific circumstances. According to co-founder and Chief Executive Amit Gal-Or, "Behavior is not deterministic; it's constantly probabilistic." 42AI converts personal information into mathematical representations of personality characteristics, then applies a transformer-based world model to determine outcome probabilities. A separate language model subsequently identifies which personality traits and contextual factors influenced each prediction.

Brokerages using 42AI can forecast what each client is likely to trade within the next 24 hours, encompassing the asset type, direction and quantity. When clients appear poised to execute identical trades, risk departments can establish hedges before the position accumulates on their balance sheet. The system draws input from know-your-customer documentation completed during account setup, as well as call frequency and vocal sentiment analysis from client interactions.

The startup is currently working with ten major brokerages and operates on either a flat-fee or revenue-sharing model. Gal-Or disclosed that revenue has surpassed seven figures but declined to specify an exact amount. According to the company, one brokerage generated roughly $30 million in incremental value through dynamic hedging. 42AI also reports 98% accuracy on trades, which Gal-Or clarified means trading flows fall within predicted ranges rather than individual trades being correctly predicted.

The underlying model required three years of development. Initial research involved thousands of individuals completing psychological evaluations, expanding to 2 million participants when the team incorporated government and biobank datasets. Current customer data encompasses approximately 30 million people. Chief Technology Officer Ehud Ezra and three co-authors published a paper in February detailing their Large Behavioral Model, which modifies an 8 billion-parameter version of Meta Platforms Inc.'s Llama 3.1 to forecast individual strategic decisions based on comprehensive psychometric profiles.

Privacy and Ethical Considerations

Gal-Or emphasized that the system operates without personally identifiable information and that "the model never knows who the person is." However, individual profiles are still constructed from connected account activity. When asked whether a brokerage might weaponize predictions against its own clients, Ezra acknowledged it as "a valid concern." The founders cited regulatory oversight, internal monitoring and an ethics committee as safeguards against misuse.

Future Expansion

42AI is investigating opportunities in gaming, fraud detection and cybersecurity, with longer-term ambitions to help robots and AI agents comprehend human behavior. The model's effectiveness diminishes when external factors go unobserved, such as traders acting on recommendations from colleagues. Gal-Or noted, "There's a natural ceiling because you really don't have access to everything that a person experiences."

The $62.5 million funding round, characterized by 42AI as seed capital, originated from the founders and Infinity Group over a three-year period. Co-founder and chairman Amir Gal-Or established Infinity Group. Dan Shechtman, recipient of the 2011 Nobel Prize in Chemistry, serves as an adviser to the research team. The company intends to pursue additional funding.