20VC: Fireworks, Open-Source AI & Why Token Costs Could Drop 10X

Open-source AI may not converge into one dominant model—instead, Lin Qiao argues intelligence will be distributed across specialized systems. In this condensed summary of 20VC with Harry Stebbings and Fireworks founder Lin Qiao, you’ll hear why enterprise AI is shifting from one-size-fits-all models to customized inference, routing, and tuning for private data, coding, legal, recruiting, and more. The summary distills the full-length conversation into a few minutes, highlighting Qiao’s forecast that token costs could fall 10x over the next three years while usage explodes 100x as companies move from experimentation to production. You’ll also learn why product-market fit in the AI era is no longer the same as scalable deployment, and how infrastructure, data centers, and chips fit into the new economics of AI. Listen now to get the key ideas in minutes.