Chinese artificial intelligence laboratories are aggressively deploying open-weight models, yet this strategy is generating severe financial strain and structural market distortions. Unlike traditional open-source software, which benefits from near-zero marginal distribution costs, open-weight AI requires expensive compute infrastructure for every inference request. This fundamental economic mismatch is driving substantial losses for Chinese AI labs while redirecting revenue toward cloud infrastructure providers. Financial data from recent quarters underscores the challenge. Zhipu AI, the public laboratory behind the GLM 5.2 model, reported nearly $500 million in losses against $107 million in revenue, triggering a share price decline exceeding 40 percent over the past month. MiniMax faces similar pressure, with $250 million in annual losses on $79 million in revenue and a stock drop surpassing 50 percent. Moonshot AI recently suspended new customer onboarding following the launch of its Kimi K3 open-weight model due to severe computing capacity constraints. Industry analysts note that unlike open-source ventures that monetize through enterprise support and customization, open-weight AI labs primarily generate revenue by hosting models. However, inference workloads naturally flow to operators with the most efficient infrastructure, typically major cloud vendors rather than model creators. The economic dynamics of open-weight distribution further complicate profitability. By releasing trained parameters to the public, Chinese labs enable third parties to download, fine-tune, and deploy models independently. Corporate customers frequently bypass domestic inference services in favor of Western cloud platforms like Amazon, Microsoft, and Google for data security compliance, or rely on specialized inference providers that rent capacity from hyperscalers. Consequently, the laboratories that invest hundreds of millions in research and training capture minimal ongoing revenue, while infrastructure providers secure the bulk of the profits. Despite the financial strain, the open-weight approach aligns with explicit national strategy. Chinese leadership has increasingly advocated for open technology as a means to accelerate commercialization and disrupt established markets. By flooding the ecosystem with capable, low-cost alternatives, Chinese laboratories aim to compress margins for American competitors like OpenAI and Anthropic, effectively commoditizing frontier AI capabilities. Regulatory and policy directives from Beijing suggest that domestic AI firms are expected to prioritize market penetration and strategic technological sovereignty over near-term profitability. Market reaction reflects this divergence between financial fundamentals and strategic positioning. While Zhipu and MiniMax shares have contracted significantly, Alibaba has appreciated by approximately 13 percent over the same period, benefiting from its dual role as a cloud infrastructure provider and model host. Analysts project that the open-weight model will intensify pricing competition and prolong the path to sustainable profitability for independent Chinese AI labs. Nevertheless, the strategy appears designed to secure long-term geopolitical and industrial advantages, transforming artificial intelligence into a widely accessible utility while positioning China as a dominant force in the global AI supply chain.