The conversation delves into the limitations of digital computing for AI, highlighting how 80 years of binary abstraction have led to energy-inefficient systems ill-suited for intelligence. Rao argues that analog computing, inspired by the brain’s physics, could drastically reduce power consumption—citing that while the human brain runs on 20 watts, data centers already consume 4% of U.S. electricity. He emphasizes the need for hardware that models time and causality, essential for true AGI, which current transformers lack. The discussion touches on strategic partnerships, particularly with TSMC, and the vision for building one of the largest analog AI chips ever. Rao also reflects on the deep satisfaction of hardware innovation and the importance of empowering high-agency individuals in startups to drive transformative change.