By Brendan Steinhauser
Beijing-based Moonshot AI’s Kimi K3 model is steadily catching up to Anthropic, OpenAI, and other frontier artificial intelligence labs for what seems like a lower cost. Now seeking a new investment at a valuation of up to $50 billion for a year-end Hong Kong IPO, Moonshot is reportedly built on Nvidia’s powerful chips and building momentum by the day.
China’s technological capabilities are now in the top tier of global AI, with massive ramifications for geopolitics and military might. How quickly times change: In May, the federal government concluded that Chinese firm DeepSeek’s newest model lagged about eight months behind the likes of Anthropic’s Opus and OpenAI’s GPT. Within months, Beijing is celebrating historic interest in the Kimi model, whose maker recently halted new subscriptions to manage sky-high demand for Chinese technology. To quote The Telegraph, China’s Moonshot is threatening the West with “full AI communism.”
For those following the Chinese Communist Party, Beijing’s AI moonshot should come as no surprise. For decades, China has identified the United States as its main competitor and has stolen American technology as a result, including AI model theft. Chinese companies have also secured restricted Nvidia chips through extensive smuggling networks, despite years of restrictions on the computing needed to train AI.
If there was ever a time to be hawkish on the China issue, the time is now. Kimi is yet another warning that China is working aggressively to close the AI gap — and succeeding. From Silicon Valley to Washington, D.C., America simply cannot afford to accelerate that effort by supplying Chinese companies like Moonshot with the world’s most advanced chips through global trade. Helping the CCP, whose stated aim is global dominance, is nothing short of a dereliction of duty.
Kimi’s emergence is a reminder for the White House and Congress to double down on export controls, restricting China’s access to American technology. Beijing is doing the same, pursuing its own export controls on AI models and chips, making it all the more important for Washington to respond in kind.
Of course, export controls only work when paired with serious enforcement, tracking, and anti-circumvention measures. This means harnessing location verification and other digital tools to prevent chips from being smuggled or diverted to China. Restricting AI-related exports is the first step. The next step is for leading national security experts and law enforcement officials to identify Chinese smuggling networks and crack down on those seeking to skirt the rules.
Last December, the federal government shut down a major China-linked AI smuggling network operating partly out of Texas. U.S. authorities seized more than $50 million in advanced chips destined for China and other restricted locations, but countless smuggling operations remain in place. From mislabeled freight to front companies and dummy server staging, CCP allies are known to engage in deception to fool the authorities, routing advanced AI chips to Taiwan and other countries before they are re-exported to China.
Proper enforcement sends a message to Beijing that U.S. export controls are not symbolic measures. They must be strong deterrents for the sake of our national security.
Advanced AI is not just a clever new invention. It is not like Instagram or Snapchat; it is national security technology. The same systems that improve commercial productivity can also support military research, cyber operations, intelligence gathering, weapons development, mass surveillance, and more. The People’s Liberation Army of China is currently leveraging AI to generate, augment and fuse increasing quantities of data, speeding up military decision-making and strengthening PLA operations to gain an edge. The emergence of Kimi makes the CCP’s job easier, at the expense of the U.S. military.
The stakes are high, and the mandate for Washington is clear. Track the chips. Close the loopholes. Protect America’s AI advantage — or suffer the consequences.