The United States leads in AI investments by a margin of 23 times, but Kimi K3 is making Washington nervous.
U.S. private investment in artificial intelligence reached $285.9 billion in 2025, 23 times China's $12.4 billion. However, despite the massive funding gap, the recent launch of the open-source Kimi K3 model by China's Moonshot AI has sparked a new wave of debate in Washington over the need for strict restrictions.
According to fresh data from the Stanford AI Index 2026 report, the U.S. private sector continues to dominate the AI race in terms of investment volume. However, China is demonstrating that technological efficiency may matter more than pure capital.
Why Kimi K3 Raises Concerns in Washington
The Kimi K3 model, distributed with open weights, unexpectedly outperformed Western counterparts in programming tests, leading to a collapse in the stock prices of U.S. chip manufacturers. In response, U.S. officials have returned to discussions about adding Chinese labs to the trade blacklist (Entity List) — where Huawei already resides.
Previously, the White House administration had shied away from radical measures in favor of supporting innovation, but the success of Kimi K3 has given new arguments to hardline proponents. AI advisor-at-large David Sacks noted: "This is a key moment for regulation. Closed market leaders want authorities to eliminate their open-source competitors."
The model's popularity is growing rapidly: Moonshot AI suspended new subscriptions just 48 hours after launch and is preparing for an IPO in Hong Kong.
What the Funding Gap Hides
Official data does not account for direct state injections from Beijing. Estimates suggest that from 2000 to 2023, Chinese government funds directed approximately $184 billion into the local AI sector. Thus, the real gap may be significantly smaller.
Financial benefits also explain why American businesses are increasingly choosing Chinese solutions. For example, DeepSeek V4 Pro costs $0.87 per 1 million output tokens, while a similar volume from Anthropic's Claude Fable 5 costs $50. Coinbase CEO Brian Armstrong reported that the exchange's switch to GLM 5.2 and Kimi K2.7 Code halved corporate AI expenses.
The DeepSeek incident in January 2025, which cost Nvidia $589 billion in market capitalization in a single day, showed how vulnerable American giants are to Chinese competition.
A ban likely won't work at all. The weight coefficients of Kimi K3 are already hosted in open repositories, and their complete removal is practically impossible. Restrictions will only raise costs, and the launch of Alibaba Qwen3.8-Max demonstrates that new competitors will keep emerging — regardless of bans.
Expert opinion: While financial indicators favor the U.S., the decisive factor is becoming not the volume of investment, but the real efficiency of the technologies being created. If China continues to release models that are cheaper and sometimes outperform Western counterparts, Washington will have to choose between costly bans and accepting market reality.