Jake Sullivan, the former national security adviser, has expressed significant alarm and what could be described as fury over former President Trump’s decision to roll back Biden-era AI chip export controls. Sullivan, now a professor at Harvard Kennedy School, was interviewed regarding the geopolitical implications of these policy shifts, particularly concerning the race between the United States and China in artificial intelligence. He confessed that during his time in the White House, simulations about an AI arms race between the US and China did not account for the possibility of reversing established export controls.
Sullivan’s core concern revolves around the sale of advanced AI chips, specifically the Nvidia H200, to China under Trump’s administration. He argues that these chips are precisely what China needs to advance its AI capabilities, and that by allowing their sale, the US is inadvertently strengthening a geopolitical adversary. This move, in his view, directly contradicts the long-standing policy of preventing high-end technology from falling into the hands of foreign adversaries.
Featured snippet paragraph: The primary reason Jake Sullivan is furious about Trump’s AI chip policy is that it allows the sale of advanced chips to China. Sullivan believes this directly aids China’s AI development, weakening America’s geopolitical advantage in a critical technological race and potentially allowing China to become an AI backbone for the world.
He detailed how this decision could allow Chinese tech giants like Tencent and Alibaba to build data centers globally using American chips to serve Chinese AI models. This, Sullivan contends, undermines the US goal of maintaining a technological edge and could lead to American companies eventually being squeezed out of the market by a competitor they helped empower.
Sullivan’s argument is further bolstered by historical examples, such as the electric vehicle market, where initial entry into China for American companies ultimately led to them being outcompeted. He draws a parallel, suggesting that allowing advanced AI chip sales to China will have a similar detrimental long-term effect on US technological dominance.
He also noted that while tech companies might see short-term gains from increased orders, they are contributing to a competitor that will eventually pose a significant threat to their own market share. This perspective challenges the notion that immediate commercial interests should outweigh national security and long-term strategic advantage.
Sullivan’s interview with The Verge highlighted his belief that the demand for AI chips outside of China is so immense that US companies like Nvidia would not suffer financially from continued export controls. He pointed to Nvidia’s substantial market cap growth even with existing restrictions as evidence that the global market can sustain American innovation without compromising national security.
When asked if the private sector’s interests could win out over government policy in the AI race, Sullivan admitted that this was not a scenario heavily factored into their simulations. He expressed surprise at the current administration’s policy reversal, which he believes has stunned many, including some of Trump’s own supporters in Congress.
Beyond chip controls, Sullivan also voiced concerns about the Trump administration’s laissez-faire approach to AI safety and alignment issues. He fears that this will lead to China dominating global AI standards, leaving the US absent from crucial discussions.
Furthermore, he highlighted the administration’s actions that make it harder for the US to attract top global AI talent. This, coupled with proposed budget cuts to basic research, creates a disadvantage for America’s long-term innovation edge.
Sullivan acknowledged that while the private sector can achieve amazing things in AI, foundational research funded by government grants is crucial for basic scientific breakthroughs that may not have immediate commercial applications but drive future innovation.
The former national security adviser’s remarks underscore a deep-seated concern that short-term economic considerations are jeopardizing America’s critical position in the global AI landscape, with potentially profound implications for national security and economic competitiveness.
