According to a research report, models such as DeepSeek, MINIMAX-WP (00100), Kimi, and GLM are expanding the industry's Total Addressable Market by reducing experimentation and deployment costs. The report values MINIMAX-WP at HK$1,330 per share, based on a projected 2028 price-to-sales ratio of 30 times. A two-year forward P/S ratio is applied to reflect the company's strong revenue growth, with an estimated compound annual growth rate of 128% from 2025 to 2030, and 184% from 2025 to 2028. The investment risk rating for the stock is "Buy."
The report indicates that DeepSeek's V4 model, specifically its V4-Pro version, performs comparably to Claude Sonnet 4.6 on artificial analysis intelligence benchmarks, while being priced significantly lower than GPT-5.5. This highlights a growing divergence between open-weight and closed-state-of-the-art models. The performance gap for open-weight models has narrowed in areas such as coding, agent workflows, and long-context applications, primarily due to cost-competitive model architectures.
It was noted that although DeepSeek's pricing for V4-Pro and V4-Flash is substantially lower than comparable models, the company acknowledges its reasoning capabilities still lag behind closed-source models by approximately three to six months. However, closed-state-of-the-art models maintain a significant competitive advantage in long-term workflows where reliable performance remains critical. As competition among open-weight models intensifies, the broader ecosystem stands to benefit from accelerated enterprise adoption, an expansion of AI's total addressable market, and raised barriers for differentiation among leading models.
Comments