Nvidia CEO Jensen Huang recently revealed alarming insights about the AI talent gap between China and the U.S. During a private dinner in Taipei, he highlighted that China boasts one million AI workers, while America has only about 20,000. Huang criticized U.S. export controls on semiconductors, arguing they are unintentionally accelerating China’s AI development instead of hindering it. Since the imposition of these controls in 2022, Nvidia's revenue from China has plummeted from $17 billion to near-zero for advanced products, costing the company an estimated $15 billion in lost sales. Huang’s claims extend to Huawei's progress in AI chip development. He suggested that Huawei's Ascend 910C processor is closing the performance gap with Nvidia, reaching 88-92% of Nvidia's capabilities. While independent research indicates the Ascend 910C achieves around 60% of Nvidia's H100 performance, Huawei is ramping up production significantly. By 2026, Huawei plans to manufacture 600,000 units of the Ascend 910C, despite challenges with memory production that may limit overall output. Huang also warned that by 2027, China could possess more AI computing power than the rest of the world combined, fueled by aggressive government support and localization policies. Chinese intelligent computing power is projected to grow from 260 EFLOPS in 2022 to 1,117 EFLOPS by 2027. Huang’s remarks, though meant to be private, became public and led to market concerns, prompting him to retract some statements. His ongoing criticism of U.S. policy reflects a broader strategic dilemma: restricting China’s access to advanced AI technology may backfire by prompting stronger domestic development in China, undermining U.S. leverage in the tech space.