companies

Numerai

Crowdsourced hedge fund where data scientists worldwide submit machine learning models against encrypted market data and stake crypto on performance.

Numerai was founded in San Francisco in October 2015 by Richard Craib, a South African technologist, around an unusual premise: give data scientists real financial data, but encrypt it so they can't tell what stocks or features they're actually looking at, then have them build models blind and combine the best ones into a meta-model that runs an actual hedge fund. The encryption solves a real problem in quant finance, stopping any single contributor from overfitting to obviously exploitable patterns or reverse-engineering the fund's positions, while still letting thousands of outside data scientists compete to improve the model. Contributors stake Numerai's native token, NMR, on their submissions, earning more NMR when their model performs well and losing the stake when it doesn't, aligning pay with actual predictive skill rather than participation. The company introduced Numerai Signals in 2020 to let contributors bring their own external data rather than working only from Numerai's provided dataset, widening the pool of usable ideas. Numerai has drawn serious institutional backing, including a commitment of up to $500 million from J.P. Morgan Asset Management, and raised a $30 million Series C in late 2025 at a $500 million valuation from investors including Union Square Ventures and Paul Tudor Jones. It remains based in San Francisco with a small, distributed team running the platform behind a global community of contributors.

Founded

2015

Headquarters

San Francisco, United States

Sector

data