China's 'open' AI is a terrible business, and nothing like open-source software (www.businessinsider.com)

China's AI landscape is facing significant financial challenges as open-weight AI models, such as GLM 5.2 from Zhipu and Kimi K3 from Moonshot, struggle to generate profits despite impressive technological capabilities. Zhipu reported a staggering loss of nearly $500 million against a revenue of $107 million last year, while MiniMax recorded a loss of $250 million on just $79 million in revenue. Analysts point out that the fundamental differences between open-source software and open-weight AI models lie in their economic structures; unlike software, which can scale with minimal additional costs, running AI models incurs substantial infrastructure expenses with each new user. The implications for the AI/ML community are profound, as Chinese firms adopt an open-weight strategy to challenge established players like OpenAI and Anthropic, potentially commoditizing advanced AI capabilities. While this may push competitors to lower their prices, the Chinese labs may not see immediate financial gains, as the revenue model for hosting and inference is not favorable. This strategic push for "openness," aligned with Chinese governmental priorities, highlights a shift in the competitive landscape, raising questions about long-term profitability for AI startups and the sustainability of their business models in an aggressive market.