US vs. China: Contrasting Approaches to AI Growth

The united States and China have two very different strategies for developing and exporting their artificial intelligence. For Silicon Valley in its new iteration, AIs are concurrently an instrument of technological colonialism and a matter of profit and intellectual property. The development of technologies is concentrated in the hands of a few interconnected companies. US models are almost all proprietary and cannot be freely reused; licenses and rules of use are stringent and subject to arbitrary modification.

A report by the Council on Foreign Relations states that the United States relies on its advantage in hardware production and algorithm development, and its ability to attract talent with economic capital.Seven companies (nvidia, Microsoft, Google, Amazon, Meta, Apple and Tesla) control the entire US artificial intelligence supply chain: they design the chips, manage the cloud, develop the models, and sell the applications. They compete, but are vertically integrated, each building its own ecosystem with startups, models, and cloud services, while also maintaining numerous interconnections. Microsoft initially allied almost exclusively with OpenAI, but OpenAI then reached an agreement with Amazon; Google and Amazon have invested billions in Anthropic, the developer of the claude models, and so on.

The United States also attempts to maintain its lead through technological and economic barriers, such as limiting the export of advanced chips to prevent their use by Chinese companies and the government. Artificial intelligence is also viewed from a national security viewpoint, justifying the production of proprietary and closed models.

China, conversely, has made diffusion and open source (at least partial) a strategic lever, including the release of models like DeepSeek or Qwen that can be reused by anyone, even US companies; simplified and free access; international cooperation with developing countries; the creation of a global ecosystem around its models; and the establishment of a sustainable development of artificial intelligence.