42AI launches with $62.5M in funding

Israeli startup 42AI has emerged from stealth with sixty two point five million dollars in seed funding to develop a behavioral world model that predicts human decision-making.

The model predicts trader behavior

42AI says its system can forecast how individual traders will respond to volatile market conditions by analyzing their underlying psychological tendencies rather than extrapolating past patterns.

How the model works

The system examines how people react to seemingly unrelated events like support calls, money transfers and stock trades, converting observations into mathematical trait representations fed into a transformer-based model.

CEO on psychological tendencies

Amit Gal-Or, 42AI's CEO and co-founder, argues that behavior is probabilistic, not deterministic. People have underlying psychological tendencies that influence decisions even when circumstances change.

“Behavior is not deterministic; it's constantly probabilistic.”

CTO on probability distributions

Co-founder and CTO Ehud Ezra says a person's tendencies play out to a probability distribution even as circumstances change. The model shows the likelihood of someone making a certain decision.

“A week from now, you might be a different person, but your tendencies will always play out in certain areas to a probability or distribution.”

Three years of research and data

42AI spent three years developing the approach, combining psychological research with neuroscience and genetics. The company studied thousands of people via assessments and scenarios before adding government and biobank datasets.

Dataset covers 30 million people

42AI's customer datasets now cover almost thirty million people. The model analyzes existing records including transactions and support calls without requiring personally identifiable information.

Trading prediction accuracy claims

42AI claims ninety eight percent accuracy in predicting market trades, though this measures whether trading flows fall within a predicted range rather than predicting each trade correctly.

F1 score and model limitations

The model achieves an F1 score of zero point eight to zero point eighty five for predicting trading direction and volume. Accuracy suffers when the system lacks important situational context.

Working with ten brokerages

42AI is working with ten large brokerages who use the system to manage liquidity, margin requirements and hedging exposure. The company generates more than seven figures in revenue.

Funding structure and future plans

The sixty two point five million seed funding came from the founders and Infinity Group over three years and was not entirely raised from external investors. 42AI plans to raise more money.

Ethical concerns and next steps

Beyond finance, 42AI is exploring applications in gaming, fraud prevention and cybersecurity. The founders acknowledged concerns about brokerages using the technology to gain advantage over customers.