The Deployment Layer Is the Most Underfunded Layer in Physical AI
Bilal Zuberi left a $7 billion fund to back physical AI 18 years in. He says the least funded layer is the dirtiest one, putting robots in the field, and a $10 billion fund earns $2 billion in fees.
· 7 min read
Bilal Zuberi, founder and general partner of Red Glass Ventures, interviewed by Turner Novak (The Peel)
Bilal Zuberi is the founder and general partner of Red Glass Ventures, a seed-stage firm that invests in AI and the physical world. He spent about 18 years investing in deep tech, most recently at Lux Capital, where he invested in companies such as Applied Intuition, before leaving to start his own fund of about $150 million. He holds a PhD in science and says he took a pay cut of more than 90% to start Red Glass. His portfolio covers robotics, defense, energy and industrial automation.
Summary
Bilal Zuberi says physical AI is now fundable because a new kind of founder moves fast, the government became a real customer, and probabilistic models let one robot do hundreds of tasks. He argues that the layer investors skip is deployment, where robots work in the messy field and generate the data that models and simulations need. He says founders must build a business with margins, not only a startup, because research labs and hype stop funding themselves. On the industry side, he thinks big-fund incentives, such as $2 billion of fees on a $10 billion fund, push investors to gather assets and restrict small investors from deals. These signals are from an interview that Turner Novak did with Bilal Zuberi on The Peel.
