Huatai Securities: The competition for foundational models is gradually shifting from product exploration to high-value scenarios and commercialization driven by enterprise markets and usage volume.

According to a research report by Huatai Securities, the competition in foundational models is gradually shifting from product exploration to commercialization driven by high-value scenarios, enterprise markets, and usage volume. OpenAI initially relied on a relatively bottom-up product innovation mechanism to establish a large-scale C-end entry through ChatGPT; Anthropic, on the other hand, continues to focus on model capabilities, coding, and enterprise APIs, gradually demonstrating resource concentration advantages as coding scenarios mature. Meanwhile, Zhipu and MiniMax are also progressively converging towards APIs, coding, and agents. Huatai Securities believes that the competitive dimensions in the foundational model industry are expanding from model capabilities and user scale to include high-value token consumption, enterprise penetration, unit economics, and product distribution capabilities.