Can cheaper Chinese AI models win adoption in developing countries?
Core argument: The performance gap between America's top AI model and its Chinese counterpart narrowed from 9.3% in January 2024 to 1.7% in February 2025.
Chinese AI models are rapidly catching up in quality compared to their American competitors. In January 2024, the performance gap between America’s top AI model and that of its Chinese counterpart was 9.3%. By February 2025, that gap had narrowed to 1.7%. Chinese AI companies, which tend to rely on less expensive and open-source models, could be able to strategically price their software licenses at lower costs than American AI companies running on closed models. This price differential may lead to in-roads on the part of Chinese AI companies with developing countries, building ecosystem dependency on Chinese AI models and software that need to be just “good enough” to deliver more basic automation of services in sectors like education, healthcare, and public administration. Chinese company DeepSeek has priced usage of their models 20-40 times less expensive than OpenAI’s ChatGPT.

