JPMorgan raises Zhipu AI target to HK$2,000, cuts MiniMax to HK$300

JPMorgan Chase raised its price target on Zhipu AI to HK$2,000 from HK$1,800 while cutting MiniMax's target to HK$300 from HK$400, citing diverging pricing power under an open-weight model strategy the bank calls "winner takes more."

"Open-weight distribution expands model reach, but monetization still depends on model quality, iteration speed and endpoint reliability," a JPMorgan analyst wrote in a July 8 report. "For leading models, open-weight access can scale usage through cloud service providers and inference platforms while generating revenue via official APIs, licensing and platform SaaS."

The bank maintained its overweight rating on Zhipu AI, the Beijing-based company behind the GLM family of large language models. Zhipu's GLM-5.2, released under an MIT license, has expanded distribution across AWS and Microsoft Azure through Fireworks AI. The new target implies the market has largely priced in Zhipu's year-end US$10 billion annualized revenue guidance, JPMorgan said, with remaining upside contingent on whether open-weight models can scale through external infrastructure rather than relying solely on Zhipu's own GPU compute stack.

For MiniMax, JPMorgan kept its neutral rating but cut the target by 25 percent. The bank cited M3's permanent 50 percent price cut as evidence the model has not achieved sufficient differentiation to command a premium over domestic rivals. "Discounts may support near-term usage, but they reduce confidence in model-driven monetization," the analyst said. In an open-weight environment, models with weaker differentiation face faster price comparison and easier traffic diversion, the bank added.

Zhipu AI shares surged as much as 9.9 percent in Hong Kong trading Tuesday after a separate report said the company is exploring development of a custom ASIC chip for its AI models. The stock has more than doubled this year, giving the company a market capitalization of about HK$700 billion, down from a peak above HK$1 trillion in April.

MiniMax shares have fallen more than 60 percent from their post-listing peak, with a market value of just over HK$100 billion.

JPMorgan flagged financing risk for both companies, noting they are in capital-intensive phases and may require two additional funding rounds in 2026 and 2027. While current cash balances support operations under base-case scenarios, accelerated model iteration or higher-than-expected inference costs could push both firms to seek external capital.

The divergent ratings signal that JPMorgan views model quality as the primary determinant of AI company valuations in China's increasingly competitive open-weight landscape. Investors will watch the next model releases — including Zhipu's GLM-5.5 and GLM-6, along with competing models from Kimi and DeepSeek — as key tests of whether Zhipu can sustain its pricing advantage.

This article is for informational purposes only and does not constitute investment advice.