Yujing Qian
Head of Engineering
GMI Cloud
Presentation Title:
Token Economics and the Collapsing Bubble of Closed-Source Lock-In
Presentation Summary:
Closed-source AI models deliver strong performance today, but their economics don’t scale. Builders are already spending tens of thousands monthly on inference, with token costs 8–10x higher than open-source alternatives for only marginal benchmark gains.
This talk explores how reliance on closed-source providers is the real AI bubble—and why forward-looking companies are insulating themselves with two strategies: model agnosticism and open-source + reinforcement learning (RL).
About | Yujing Qian
Yujing Qian is the Head of Engineering at GMI Cloud, where he leads the development of cutting-edge GPU cloud infrastructure to help companies accelerate their AI workloads at scale.
Previously, he served as an engineering leader at Mineral, an Alphabet company, building AI-powered solutions for sustainable agriculture.
With deep expertise in software engineering, AI, and advanced research — including work in quantum cryptography and multiple patents — Yujing combines technical excellence with a passion for solving real-world problems.