Zhipu and MiniMax have turned large models into two different business models.
Zhipu AI and MiniMax recently released their first mid-year reports after going public, showing the commercialization strategy of large models. Zhipu’s revenue in the first half of the year reached 954 million yuan, a year-on-year increase of 399.7%; MiniMax’s revenue was 117 million US dollars, a year-on-year increase of 283.1%. Both companies had adjusted losses of nearly 300 million US dollars, and they have not yet achieved profitability. Zhipu API revenue accounted for 86.5%, with local deployment becoming less important, and the company is shifting to a cloud-based model where fees are charged based on the number of Token calls. The number of Token calls on its MaaS platform increased by more than 40 times since the beginning of the year, and the unit推理 cost decreased by about 80%. MiniMax continues to expand its global presence, with overseas revenue accounting for 60.8%, focusing on reducing unit intelligent costs to increase user base and application scenarios for Agents. Both companies emphasize high investment in research and development, and in the short term, they still rely on accounting factors related to preferred stock conversion to affect their loss figures.