Guozhen AIGlobal AI field notes and model intelligence

Realtime AI News

Hong Kong-Listed AI Firms Zhipu and MiniMax Release Internal Letters on Same Day as Lockup Expiry Nears

Zhipu AI and MiniMax, two major Chinese AI model companies listed in Hong Kong, released internal letters from their founders on the same day outlining strategic directions. The letters arrived during the sensitive IPO lockup expiration period, and the two stocks showed diverging price movements.

Published
港股大模型双雄智谱与MiniMax同日发内部信,解禁期股价走势分化
Image source: zhipuai.cn

Zhipu AI and MiniMax, the two prominent Chinese large language model companies listed on the Hong Kong Stock Exchange, issued internal letters from their founders on the same day, drawing widespread market attention. Both founders addressed the competitive landscape, product strategy, and organizational priorities in their communications.

Zhipu founder Tang Jie outlined the company's strategic focus areas, including deepening AI model R&D, accelerating commercialization, and expanding international reach. MiniMax's founder also detailed the company's next-phase development roadmap in their letter. Media reports interpreted both letters as efforts to align internal morale and clarify direction during the post-IPO lockup period.

The timing was significant: both stocks were approaching the expiration of their IPO lockup periods, during which early investors and employees become eligible to sell shares. Following the letters, Zhipu and MiniMax shares showed diverging price trajectories, reflecting differentiated market assessments of the two companies' prospects.

As the leading AI model companies listed in Hong Kong, Zhipu and MiniMax serve as key vehicles for global investors to gain exposure to China's AI sector. The coordinated timing of their internal letters sparked discussion about the competitive dynamics between the two firms.

Analysts noted that while internal letters are routine corporate communication, their publication during the sensitive lockup window served to stabilize market expectations. The diverging stock performance suggests investors are moving from a sector-wide approach to individual company assessment in the Chinese AI model space.

Why it matters

The simultaneous internal letters from Zhipu and MiniMax highlight how Chinese AI model companies are navigating post-IPO transitions, while diverging stock performance signals that investors are increasingly making company-specific rather than sector-wide bets.

智谱MiniMax港股大模型内部信
Back to realtime news

Nearby Updates

All

07/12, 16:46

Apple Sues OpenAI, Sparking Three-Way Public Feud: Musk Accuses Altman of Fraud, Altman Fires Back on 'Space Data Center'

Apple has filed a lawsuit against OpenAI, igniting a high-profile public feud involving Elon Musk and Sam Altman. Musk accused Altman of fraudulent behavior while Altman retorted by mocking Musk's touted space data center project.

07/12, 14:29

NVIDIA RTX Spark Demo: Runs Alan Wake II, Full Unreal Engine City, and a Self-Coding AI Agent

NVIDIA's compact AI workstation, the RTX Spark, demonstrated remarkable versatility in a public showcase, running the AAA game Alan Wake II, a complete Unreal Engine city scene, and an AI agent capable of writing its own code. The demo underscores how far desktop-class AI computing has come in handling diverse high-intensity workloads.

07/12, 14:02

Anthropic Responds to Complaints of Claude Getting 'Dumber' — Says the Model Isn't at Fault

After a wave of user complaints that Claude's responses have declined in quality, Anthropic has officially responded, stating the model itself is not underperforming. The company suggested the perceived degradation may stem from changes in system prompts, safety filters, or context handling mechanisms.

07/12, 11:57

StepFun Unveils Step Edge On-Device Model Family for Smartphones and Automotive AI

Chinese AI company StepFun (阶跃星辰) launched its Step Edge family of on-device AI models on July 12, enabling AI inference directly on smartphones and in-vehicle systems. The release marks a strategic expansion from cloud-based models to edge computing across multiple terminal categories.