This episode explores the financial and technical machinery behind the AI boom. Patel and Nanos argue that AI infrastructure spending is larger than commonly measured, supported by global demand and expanding supply chains, but that growth must eventually slow as capital and physical capacity become constrained. They discuss how future revenue from software, robotics, drug discovery, and other industries could justify today’s investment, while AI shifts value from labor toward capital. The conversation then turns to knowledge work, neocloud contracts and security, NVIDIA’s financing of new capacity, Oracle’s execution risks, Google’s strategic position, and the possibility that frontier labs will keep their strongest models for internal products.