U.S. companies, including cryptocurrency exchange Coinbase, are increasingly adopting Chinese AI models like Moonshot’s Kimi K3 to trim operational costs. This shift, driven by the models' affordability and growing efficiency, challenges the dominance of U.S. tech giants and their more expensive offerings.
San Francisco-based Raffi Krikorian, chief technology officer at Mozilla, switched to Kimi K3 within days of its July 2026 launch. He found it "snappier" than Anthropic's pricier Claude Fable chatbot, which he had used previously. Krikorian had also relied on Z.ai’s GLM-5.2 for routine tasks like managing his calendar and email, highlighting a broader trend among independent software developers.
Capital's Cost-Cutting Drive
Curt Meinhold, a Greensboro, North Carolina-based technology executive and founder of LilyList, prefers DeepSeek for business leads and sales generation. He stated that the vast majority of users don't need the most expensive models like Anthropic’s Mythos or Fable. Meinhold noted the significant cost difference, paying "a handful of cents per million output tokens versus 30 bucks or 40 bucks or 50 bucks" for comparable performance in code and research tasks.
This relentless pursuit of lower operating expenses has propelled Chinese models to the forefront. Data from OpenRouter, an AI model tracking platform, shows Chinese models occupying the top five most popular spots over the past month. Sensor Tower, a market intelligence firm, reported Kimi K3 garnered over 930,000 downloads globally in the week following its July release, a 200% increase. In the U.S. alone, downloads jumped 387% to approximately 86,000, overwhelming Moonshot's capacity and forcing a temporary suspension of new subscriptions.
The State as Enforcer
The rise of Chinese AI models comes amidst American-led restrictions designed to block China from accessing advanced technologies, including cutting-edge AI chips. U.S. Treasury Secretary Scott Bessent has warned of potential new sanctions, ostensibly to protect American intellectual property. On Wednesday, U.S. President Donald Trump’s administration accused Moonshot of using “covert” methods to build K3, allegedly leveraging Anthropic’s Fable. Some U.S. politicians and AI companies, including Anthropic, have leveled accusations of “illicit distillation” against Chinese startups, claims Beijing dismisses as “groundless.” These state actions serve to protect the accumulated wealth and market share of U.S. capital.
However, U.S. tech policies have at times inadvertently aided Chinese competitors. Z.ai released its GLM-5.2 model in mid-June, shortly after Trump administration export controls kept Anthropic’s Fable and Mythos models offline for over two weeks. Anastasios Angelopoulos, co-founder and CEO of Arena, observed that "Restricting an American model can immediately create an opening for a Chinese competitor."
Global Competition and Concentration
Most Chinese AI models are open-source, allowing anyone to examine and build upon them, a stark contrast to the closed-source frontier models from U.S. companies like Anthropic and OpenAI. Lian Jye Su of Omdia anticipates Chinese vendors will use open-source software to expand their global reach. Krikorian noted that "the open frontier is becoming increasingly Chinese-built" as these models approach the capabilities of leading U.S. closed systems. While the U.S. considers further restrictions, a group of American tech firms, including Microsoft, Meta, and Nvidia, signed an open letter backing “open” AI models, a strategic move to manage market dynamics.
In China, state support has fostered rapid AI adoption by people and businesses. Chinese President Xi Jinping championed open-source AI models and global equity at a Shanghai technology summit in the same month, pledging Chinese involvement in raising AI capabilities, particularly in developing nations. Companies like Huawei and Tencent are integrating AI into devices such as smartphones and humanoid robots. This intense domestic competition drives Chinese companies to expand globally, with leading startups raising significant funding, including through public share offerings. Chelsey Tam of Morningstar noted that both the U.S. and China aim to "encourage widespread adoption of their AI ecosystems, while safeguarding technologies that could materially strengthen strategic rivals." The massive investments required for this competition are evident in Z.ai's financial reports, which showed revenue surging 132% to $107 million last year, but net losses also jumped 60% to $694 million, illustrating the speculative nature of capital accumulation in this sector.