
The AI industry is undergoing a structural shift as open-weights models increasingly capture token usage share from closed-source frontier models. While companies like OpenAI and Anthropic maintain dominance in revenue by providing high-performance, state-of-the-art solutions for complex tasks, open-weights alternatives—many originating from China—offer superior cost-efficiency, privacy, and customization for enterprise applications. By fine-tuning these models on proprietary data, businesses can build long-term intellectual equity rather than merely renting intelligence from closed-source platforms. However, this trend introduces significant geopolitical risks, as widespread reliance on Chinese-developed open-weights models could lead to long-term dependency on foreign infrastructure and hardware. Ultimately, the market is bifurcating into a high-margin, closed-source frontier and a high-volume, open-source ecosystem, forcing enterprises to carefully balance immediate cost savings against long-term strategic sovereignty.
Sign in to continue reading, translating and more.
Open full episode in Podwise