Jensen Huang – TPU competition, why we should sell chips to China, & Nvidia’s supply chain moat

In this deep-dive conversation, NVIDIA CEO Jensen Huang discusses the company's strategic moat in the AI era, emphasizing that NVIDIA's value lies not in hardware alone but in its unparalleled ecosystem of software, tools, and partnerships. He argues that while competitors like Google's TPU are optimized for specific tasks like matrix multiplication, NVIDIA's programmable GPU architecture enables rapid innovation across diverse AI workloads—from scientific computing to novel model architectures—making it indispensable for both hyperscalers and startups. Huang defends NVIDIA's supply chain dominance, explaining how massive purchase commitments and long-term partnerships with TSMC, SK Hynix, and others create a self-reinforcing cycle of innovation and scale. He also addresses geopolitical tensions around chip exports to China, asserting that denying Chinese firms access to advanced chips is counterproductive: China already has vast manufacturing capacity, energy, and talent, and restricting exports risks ceding technological leadership to a fragmented, less open global AI ecosystem. Instead, he advocates for open collaboration, arguing that American leadership comes from ecosystem strength, not isolation. Finally, Huang reflects on NVIDIA's future, emphasizing that even without AI, the company’s mission to accelerate computing across domains remains central to its identity.