Mnemonic Labs, a recent Y Combinator investment, is an early-stage AI startup currently listed in Y Combinator Fall 2026 (YC F26). It calls itself "OpenAI for Financial Markets," with a core focus on training foundational models for quantitative trading. Similar to LLM's dominance in NLP with scaling laws, they aim to build "world models" in financial markets that automatically learn new strategies as the market grows. - The company consists of only two people, two brothers. - They built their training pipeline and model architecture in less than 6 weeks. - The current training run uses over 25TB of tokens. Founding Team: Videh Seksaria @videhps MIT Math + CS (graduated at 19); formerly Citadel's youngest Quant (19 years old), also worked on Walleye QR Kalki Seksaria @bayesiankalki MIT Math; previously worked as a Quant at Tower Research Capital, achieving approximately $20 million in annualized returns by age 22. PnL They believe that quantitative trading firms can't achieve the return on investment for developing such basic models. I don't think quantitative trading firms are incapable of doing this; rather, even if they could, they wouldn't offer it as a service. Offering it as a service could actually be an opportunity for model providers. I think YC might be interested in their backgrounds and the market opportunity for offering such services. Official Website: http://mnemoniclabs.ai