Analyst says the U.S. AI race is unlikely to slow, with safety rules potentially reinforcing top labs’ edge

Citrini analyst Jukan, citing a research note from Tianfeng Securities, said the U.S. government needs to preserve its lead in artificial intelligence, making it difficult for the country to truly slow down once it is in an AI race. In his view, recent calls from Anthropic and OpenAI to slow AI development should not be read only as safety advocacy. He said those calls may also reflect the difficulty of pausing competition and the possibility that tighter safety regulation could strengthen the position of leading laboratories. Jukan said the public case for slowing AI centers on the idea that safety testing, operational monitoring, and third-party verification are struggling to keep up with the speed of model iteration. He added that, in the short term, such messaging could weigh on sentiment around AI-related stocks and lower expectations for next-generation models. He also outlined another possibility: the industry may still be constructive on AI over the long run, while seeking to delay the next wave of large research spending, prioritize commercialization of existing products, and ease pressure from infrastructure and capital expenditures. He described the AI race as a prisoner’s dilemma in which every party may want to slow down, but none is willing to stop first for fear of losing advantages in technology, customers, and financing. Jukan also said that as model launches require expensive evaluation, certification, and ongoing audit costs, larger labs are better positioned to absorb those fixed expenses, potentially raising barriers for smaller teams.