Chinese artificial intelligence models have surged in global use, with their share of tokens on OpenRouter rising from 6%-13% in February to 57%-67% in the week of September 14. This rapid adoption is driven by lower costs and greater flexibility, directly benefiting corporations seeking to maximize profit margins. American firms like Airbnb, DoorDash, and Shopify have already integrated these Chinese models into their operations.
The global usage of Chinese models jumped from less than 15% to over 54% last week, according to OpenRouter data, led by DeepSeek. This shift highlights a clear capitalist calculation: companies prioritize cost-efficiency over national origin, even amidst escalating geopolitical tensions. Harpreet Arora, head of agentic infrastructure at Vercel, noted that Chinese models are becoming capable enough for more tasks at a much lower cost, making the price difference compelling once a quality bar is met.
Capital's New Frontier
The real beneficiaries of this trend are not just the end-users but the infrastructure providers. Eric Olander, editor in chief of The China-Global South Project, observed that while DeepSeek or Moonshot might not generate significant direct revenue, companies like Huawei and ZTE, which provide support, technology, and services around AI, stand to profit immensely. Chucheng Feng, founding partner of Hutong Research, framed the choice for most users as a trade-off between a model that costs ten times as much but is six months ahead, or one that costs one-tenth as much but is only six months behind—a decision driven purely by economic calculus.
This pursuit of cheaper AI models extends to the so-called "Global South," encompassing 82 countries across Central and South America, Africa, and Asia. Daniel Remler, a senior fellow at the Center for a New American Security, stated that entrepreneurs and governments in regions from Lagos to São Paulo to Jakarta, seeking cheap, open models, will look first to Chinese AI. This expansion represents a new avenue for market capture and technological dependency, solidifying spheres of influence for competing capital blocs.
Imperialist AI Race
Washington and Beijing are locked in a fierce competition to control the narrative and infrastructure of global AI. President Trump declared, "We're leading China in AI...and frankly, I want to keep it that way, because whoever wins AI wins." This statement lays bare the imperialist dimension of the AI race, where technological dominance is equated with global power and capital accumulation. The United States, with its pioneering AI labs, advanced chips, and deep-pocketed investors, maintains an obvious edge, forcing countries worldwide to align with its interests.
In response to this technological rivalry, Trump formed the Pax Silica alliance last year to reduce reliance on China for AI supply chains, drawing in more than two dozen countries and the European Union. Not to be outdone, Chinese leader Xi Jinping launched the World Artificial Intelligence Cooperation Organization in the same year, recruiting 29 countries, including Russia, Indonesia, and Pakistan, to champion an alternative vision of open, accessible models. This dual-alliance system reflects a global struggle for control over the next generation of productive forces.
Accusations of "industrial-scale theft" of American trade secrets have been leveled against China by Washington. Firms like Anthropic and OpenAI claim Chinese AI labs trained their models on outputs from American systems, a process known as distillation. Beijing has rejected these accusations, framing them as attempts to contain China's rise. George Chen, chair of digital practice at The Asia Group, confirmed that Beijing perceives Washington's actions as an effort to contain its technological advancement, shaping future engagement between the two powers.
The State as Enforcer
The state's role in this contest extends beyond economic competition to the instruments of violence. The United States and Russia actively weakened a global UN effort to regulate lethal autonomous weapons. They stripped crucial safeguards from the language, ensuring that the development and deployment of "killer AI" remain largely unchecked. This action directly serves the interests of the military-industrial complex and reinforces the capacity for imperial projection, demonstrating how the state acts to protect and expand the power of accumulated capital, even at the cost of human life.
Amidst these escalating tensions, a summit between Trump and Xi this week produced little beyond a formal dialogue and a bilateral communication channel. The two sides agreed to refer to AI as "super intelligence" or "SI" and to pursue a dialogue on SI. Xi's call to "draw on each other's strengths, not guard against each other" and to "exchange views on its risks and benefits, and jointly prevent the misuse and abuse of AI" offers symbolic concessions that mask the underlying structural competition for global technological and economic hegemony. These dialogues function as a means to manage contradictions without addressing the fundamental drive for capital accumulation and imperial dominance. Two U.S. House committees are now investigating the impact of rising adoption of Chinese models, signaling the state's intent to protect domestic capital interests. Daniel Remler warned that Chinese AI represents "real economic and security risks for the United States," fearing that integration of Chinese models could pull countries into a "Chinese technology sphere of influence that hardens into geopolitical alignment."