Kimi K3 is so powerful that the Strategic Lead at OpenAI is starting to discuss how the United States should defend itself

According to Pulse Beat monitoring, OpenAI's Chief Strategic Officer, Dean W. Ball, stated that Kimi K3 is a very powerful model. Its performance in Agent Programming is approaching the best public model in the first quarter of 2026. Such capability cannot be easily explained by distillation.

However, his comments were not just to praise Kimi K3.

Ball's real concern is why China is still willing to open source this level of model and how it will impact the U.S. AI industry.

In his view, the pressure brought by China's open-source models is not just about having another cheap competitor. As long as the open models are strong enough, developers do not have to continue paying high prices for closed-source models. The profits of model providers will be reduced, and investors will also be more cautious. The incentive for American companies to invest billions of dollars in training the next generation of models may weaken as a result.

Open weights can accelerate the spread of technology, but it may also make it increasingly difficult to earn money by training cutting-edge models. Ultimately, model development may have to rely on subsidies from other businesses or government funding, turning AI into a public infrastructure similar to the electricity grid and roads.

Ball speculates that China is willing to do this partly because it has not fully considered the risks of advanced AI. Another reason is that since the U.S. restricted the export of advanced chips, China lacks the computing power to provide inference services to global users. Since it is difficult to monopolize users through APIs, open weights have become a way to expand influence.

He also predicts that the U.S. government will eventually find ways to prevent Chinese models from entering domestic enterprises. The U.S. does not need to directly ban open-source models; it only needs to continuously highlight concerns about backdoors, data security, and compliance risks, and regulated industries such as banks will voluntarily steer clear.

Ball even believes that these warnings do not need particularly strong evidence. Simply creating enough uncertainty can make companies reluctant to adopt them, while not forcing all developers to turn to overseas providers that are harder to regulate.