AI startup 42AI raises $62.5 million for predictive trading model
Israeli AI startup 42AI has exited stealth with $62.5 million in seed funding to develop a behavioral world model that predicts individual trading decisions for brokerages, claiming 98% accuracy.
Israeli AI startup 42AI has raised $62.5 million in seed funding. The company, which emerged from stealth on November 2, 2026, secured the capital from its founders and the investment firm Infinity Group, which was founded by Amir Gal-Or. 42AI plans to raise more money, though its current valuation is not disclosed.
Predictive Accuracy and Model Innovation
42AI claims its behavioral world model achieves 98% accuracy in predicting market trades, defined as trading flows falling within a predicted range. For predicting trading direction and volume, the model reportedly achieves F1 scores between 0.8 and 0.85.
The model is designed to predict individual decision probabilities. It is built on psychological, neuroscientific, and genomic data from millions of individuals. Training data includes psychological assessments of thousands of people, government and biobank datasets covering over 2 million people, and customer datasets covering almost 30 million people. The system converts observations into mathematical representations of traits and uses a transformer-based world model with a separate language model for explanations. Co-founder and CTO Ehud Ezra stated, "In a world model, every individual is the world."
The company acknowledges a natural ceiling to the model's accuracy due to limited access to all personal experiences. Amit Gal-Or, the CEO, emphasized the probabilistic nature of the approach: "Behavior is not deterministic; it’s constantly probabilistic."
Client Applications and Real-World Impact
42AI's initial deployment is at financial brokerages. The model analyzes customer interactions, including KYC processes, onboarding, support calls, and trading activity, to predict how individual traders will respond to volatile conditions. It does not use personally identifiable information.
The startup works with 10 large brokerages, using the predictions to manage liquidity, margin requirements, and hedging exposure. The model runs on customers' own records of transactions and support calls, creating closed data loops. One brokerage client reported generating about $30 million in additional value through dynamic hedging using 42AI's technology. The company states it is selling risk management, not a chatbot.
| Metric | Claimed Performance |
|---|---|
| Prediction Accuracy | 98% for market trades within a predicted range |
| F1 Score (Direction & Volume) | 0.8 to 0.85 |
| Prediction Horizon | Up to 24 hours ahead by instrument, direction, and volume |
Ethical Considerations and Future Expansion
42AI acknowledges ethical concerns about using predictive models for competitive advantage in trading. The company states it is part of the solution to such ethical risks.
Financially, 42AI already generates more than seven figures in annual revenue. The recent funding round reflected insider conviction, with the CEO noting it was not entirely raised from external investors.
The company is exploring applications for its behavioral modeling technology beyond finance. Future domains include gaming, fraud prevention, cybersecurity, and AI agents. 42AI plans to expand into these new areas while continuing to serve brokerages with its risk management solutions.
