Chinese AI Firms Struggle to Profit From Advanced Models as Open-Source Strategy Squeezes Revenue
Updated
Updated · The New York Times · Jul 27
Chinese AI Firms Struggle to Profit From Advanced Models as Open-Source Strategy Squeezes Revenue
3 articles · Updated · The New York Times · Jul 27
Summary
Chinese AI companies are releasing models that approach the world’s best systems, yet still cannot turn them into sustainable profits.
Alibaba and ByteDance have started charging for top-tier model access, while startups such as DeepSeek and Moonshot AI have raised billions to keep funding development.
The economics remain punishing because state-of-the-art AI requires constant spending on advanced chips, testing capacity and data-center infrastructure to serve users at scale.
China’s open-source and open-weight approach speeds industrywide progress, but it also makes monetization harder by spreading gains across competitors instead of concentrating revenue.
The profitability problem is not unique to China: OpenAI, Anthropic and Google are also spending more on AI than they currently earn from it.