According to an article on The Next Platform, recent concerns have arisen that open weight AI models from China, such as Alibaba Qwen, DeepSeek R1, V3, and V4, and Moonshot Kimi K3, are approaching the capability of closed models from the US like OpenAI GPT, Anthropic Claude, and Google Gemini. This could lead American and European companies to choose Chinese open-weight models over more expensive closed-weight models, potentially reducing revenue for US AI model makers and making it insufficient to cover their enormous AI hardware rental and acquisition costs. There are trillions of dollars of investment in datacenters and hardware planned over the next two years. Annualized revenue run rates are reported as about $25 billion for OpenAI (up around 2X from January) and $47 billion for Anthropic in May (up 5.2X from December 2025). However, the article notes that ARR is not actual revenue, and it examines new data from Gartner on actual spending on AI models and platforms, which is a more restricted view than a previous Gartner dataset from January that included AI PCs, AI servers, and AI-enhanced applications for 2025 with forecasts for 2026 and 2027.