Raffi Krikorian, chief technology officer at Mozilla, switched to Chinese AI startup Moonshot's Kimi K3 within days of its July launch. He found it snappier than Anthropic's more expensive Claude Fable chatbot. That single decision by a Silicon Valley executive captures something larger happening across American technology: Chinese AI models are displacing US competitors because they cost less, perform nearly as well, and remain open-source at a moment when US companies are locking theirs behind paywalls.
Krikorian isn't alone. Coinbase and other US companies are switching to Chinese systems to trim costs. Estimates from Sensor Tower showed Kimi had more than 930,000 downloads globally during the week after K3's July release—a 200% increase from the week before. In the US, it was downloaded around 86,000 times, a 387% jump. The model proved so popular that Moonshot had to temporarily suspend new subscriptions after overwhelming demand pushed its capacity close to its limits.
The Economics Driving the Shift
Curt Meinhold, a technology executive in Greensboro, North Carolina, articulated the math plainly. "Most of us, the vast majority of us, 90 plus percent, don't need Anthropic's Mythos or Fable," he said. "We just don't need it, we need something good enough." Meinhold pays a handful of cents per million output tokens for Chinese models versus 30 to 50 dollars for US alternatives. When companies deploy AI "agents"—systems that autonomously conduct multistep, complex tasks—those cost differences compound dramatically.
Based on data over the past month, the top five most popular models on OpenRouter, a platform tracking AI usage, were Chinese. Goldman Sachs wrote in a July research report that Chinese AI models are reaching a "critical stage" for wide adoption, especially as demand for cost-effective systems surges alongside the rise of agentic AI usage.
The latest wave of Chinese competition is built on momentum. Early last year, DeepSeek shook up the US technology industry with an AI model that performed on par with US systems but much more cheaply. Now Z.ai released its GLM-5.2 model in mid-June. In April, DeepSeek rolled out previews of its latest V4 model, challenging OpenAI's GPT and Google's Gemini. In July, China's Alibaba previewed its Qwen3.8 Max AI model. Experts say these latest systems are nearly as intelligent as frontier models from OpenAI and Anthropic.
Open Source as a Competitive Advantage
Most Chinese AI models are open-source, meaning anyone can examine and build on them. That contrasts sharply with frontier models from US companies like Anthropic and OpenAI, which remain closed-source. Lian Jye Su of the technology research firm Omdia said Chinese model vendors are expected to leverage open-source software to promote their global usage and adoption.
Raffi Krikorian observed that "the open frontier is becoming increasingly Chinese-built." That's significant. A group of American tech firms including Microsoft, Meta and Nvidia signed an open letter published Friday backing "open" AI models—suggesting even US tech leaders recognize the strategic appeal of transparency and accessibility in the AI market.
US policy is inadvertently accelerating the shift. When the Trump administration put export controls on Anthropic's Fable and Mythos models in June, keeping them offline for more than two weeks, Chinese competitors rushed to fill the void. Z.ai released its GLM-5.2 model not long after. "Restricting an American model can immediately create an opening for a Chinese competitor," said Anastasios Angelopoulos, co-founder and CEO of Arena, a platform for evaluating AI systems.
The US Response and Its Limits
The Trump administration on Wednesday accused Moonshot of using "covert" but not necessarily illegal methods to build K3 off the back of Anthropic's Fable. US Treasury Secretary Scott Bessent has warned more sanctions could be coming to protect American intellectual property. Some US politicians and AI companies including Anthropic have accused Chinese startups of illicit "distillation" of their models to extract their technologies. Beijing rejects these claims as "groundless."
Yet short of an outright ban, Chinese models are likely to continue appealing to independent software developers in the US and elsewhere. Angelopoulos noted that Chinese AI labs are also putting pressure on one another, meaning "the competition is no longer simply the United States against China."
While Chinese models still lag American AI leaders across their overall, full-range capabilities, the gap is narrowing. Yasir Atalan of the Center for Strategic and International Studies said US AI firms are looking for more options to reduce pricing, such as cheaper alternatives of competitive models.
In China, intense competition at home is driving companies to expand globally. Chinese tech companies like Huawei and Tencent are embedding AI in devices such as smartphones, AI glasses and humanoid robots. At a flagship technology summit in Shanghai, Chinese President Xi Jinping championed open-source AI models and the importance of promoting greater global equity, while pledging Chinese involvement in raising AI capabilities especially in developing nations. Leading Chinese AI startups are now raising more funding to support that global effort, including through public share offerings.
Sustainability Questions Loom
Huge investments in AI have raised concerns over the sustainability of Chinese startups. Z.ai reported its revenue surged 132% to 724 million yuan ($107 million) last year, but net loss jumped 60% to 4.7 billion yuan ($694 million). That pattern of explosive growth paired with deepening losses mirrors challenges facing US AI companies as well.
Still, Krikorian's recommendation stands: "I would highly recommend anyone doing any serious AI load to at least evaluate it."
Why This Matters:
The rise of cheaper Chinese AI models reveals a fundamental tension in US technology policy. While the Trump administration pursues restrictions on Chinese competitors, those very restrictions can create market openings for the companies they aim to constrain. More broadly, the shift highlights how market concentration among expensive US AI providers—Anthropic, OpenAI—creates conditions where more affordable alternatives gain traction, especially among smaller companies and developers who can't absorb high costs. The dominance of open-source Chinese models also raises questions about who controls AI infrastructure globally and whether closed, proprietary systems from US companies will remain the default. For workers and small businesses relying on AI tools, the availability of cheaper options matters materially to their ability to compete. For regulators, the speed of Chinese adoption in the US suggests that trade restrictions alone won't determine technological outcomes—market forces, pricing, and accessibility may prove more decisive than policy.