TAIPEI, May 21 – Nvidia’s Chief Executive Officer, Jensen Huang, voiced strong support for President Donald Trump’s decision to reverse certain U.S. restrictions on AI chip exports to China, calling the original controls ineffective and harmful to American business.
Speaking at the Computex tech conference in Taipei, Huang sharply criticized the existing export policies initiated under President Joe Biden, which aimed to limit China’s access to advanced AI chips by categorizing countries into three tiers, with China being entirely blocked from obtaining high-end semiconductors.
“The export restrictions, in the end, didn’t achieve their intended goals,” Huang stated, emphasizing that the reasoning behind these measures has not held up under practical conditions. “The core beliefs that shaped the AI diffusion rule from the start have clearly shown themselves to be deeply mistaken.”
The AI chip controls, introduced as a measure to restrict China’s progress in AI technology and related sectors, have led to unintended consequences. Rather than curbing development, Huang noted, the controls have only pushed Chinese companies to double down on their domestic production efforts and accelerate the creation of their own supply chains, free from reliance on U.S. companies.
The result? U.S. firms, especially Nvidia (NVDA.O), have lost a major share of a once-dominant market. Huang shared that Nvidia’s market share in China has plummeted from 95% to 50% since the start of the Biden administration. This steep decline has cost American companies billions in potential revenue, he added.
Despite the U.S. restrictions, China has continued to push forward with AI research and development. Huang emphasized that building and training AI models requires significant infrastructure and investment. The demand for high-performance chips hasn’t diminished—it’s simply shifted away from American companies like Nvidia to Chinese alternatives, such as those provided by Huawei.
“Research continues regardless of restrictions,” Huang said. “And research requires capital—lots of it—for AI infrastructure. If we’re not there, someone else will be.”
Huang praised the Trump administration’s approach to the issue, arguing that it reflected a better understanding of the global nature of technology and trade. He welcomed the shift in strategy, where Trump’s team is reportedly considering abandoning the tiered export restrictions in favor of a licensing system based on government-to-government agreements.
“President Trump realizes it’s exactly the wrong goal to block exports in such a sweeping manner,” Huang said. “We need to ensure U.S. companies stay competitive, not box them out of critical markets.”
Nvidia has also felt the impact of the policy’s consequences. The company reported it would take a financial hit of $5.5 billion due to limitations imposed on the sale of its H20 AI chip to China. Huang recently estimated that the total impact on H20-related revenue could rise to $15 billion, underscoring the scale of the damage to one of the company’s best-selling products.
Even amid the setbacks, Nvidia remains adaptive. The firm is reportedly working on a modified version of its Blackwell AI chip that incorporates slower memory components. This change would allow it to meet export compliance while still serving global demand for AI technology.
Competition in China is fierce, according to Huang, who acknowledged that local developers have risen quickly to fill the void left by U.S. companies. “The competitive landscape in China is extremely challenging,” he acknowledged. “Many there would prefer we never return.”
As China’s AI market is projected to reach approximately $50 billion next year, the potential rewards are substantial. For Nvidia, re-entering and reclaiming ground in that market could be a major revenue opportunity.
At the Computex conference, Nvidia also unveiled a series of new innovations and business initiatives designed to expand its $130 billion revenue stream. The announcements showcased Nvidia’s intent to continue leading the AI chip industry, regardless of geopolitical and regulatory headwinds.
Meanwhile, tensions between Washington and Beijing remain high. In a recent statement, China criticized the United States for issuing new guidance discouraging American firms from using Chinese AI chips, such as Huawei’s Ascend series. Beijing called for an immediate end to what it described as discriminatory practices and warned of retaliatory measures if the U.S. continues policies seen as damaging to Chinese interests.
The U.S. government has long expressed concerns over China’s ambitions in AI, citing national security risks and fears of military application. However, critics argue that the current approach is backfiring, causing U.S. companies to suffer and inadvertently accelerating China’s independence in the tech sector.
As global competition in artificial intelligence heats up, Huang’s comments highlight a growing divide over how best to manage international technology trade. For Nvidia and many others in Silicon Valley, the priority is clear: policies must support innovation, not stifle it.