Ex-IMC Semiconductors Quant: AI Is Good Enough to Beat the S&P 500

Lihong Wang spent his career as a discretionary options trader at IMC, one of the world's largest market makers, trading semiconductor volatility on names like Nvidia, AMD, and Broadcom. In this episode, he breaks down how a prop trading desk actually makes money: reading order flow to identify counterparties, trading against structured product hedging flows from Asian banks, warehousing benign risk instead of externalizing it, and why correlation blowups — like the DeepSeek selloff and July's deleveraging cascade — are how options market makers get hurt. He explains why the seat at a top trading firm is one of the most leveraged information positions in markets, how firms like IMC, Jane Street, SIG, and HRT train and allocate quant trader talent, and what separates market making from prediction.Now the founder of Freeport, a YC-backed perpetual futures exchange, Lihong shares his AI investing thesis: why he holds a 2x levered portfolio of 50 stocks across the entire AI and semiconductor supply chain — from TSMC and ASML to optics, connectivity, and memory names like SK Hynix — and why nearly every quant trader he knows is doing the same. We cover how narratives move from private conversations to Twitter to Bloomberg (and where the edge dies), Leopold Aschenbrenner's Situational Awareness fund, whether AI is a dot-com-style bubble, Kelly criterion leverage sizing, the future of perp DEXs like Hyperliquid, and how young people should think about risk, career moats, and personal brand in the age of AI. Essential listening for hedge fund analysts, quants, options traders, and anyone trading the AI supply chain.