Narrative Drives Everything As China’s AI Newcomers Enter An Era Of Extreme Volatility, Retail Investors Flood In
China’s emerging AI equities have become among the most volatile segments in Asian markets. Propelled by an intense AI narrative, retail capital has assumed a dominant role in price formation while institutional ownership remains sparse, a combination that has pushed many related stocks into extreme volatility ranges.
Using 90‑day annualized volatility as a metric, half of the ten most volatile Asian stocks with market capitalizations above USD 10 billion are recent listings of Chinese AI companies. Moore Threads surged more than 700 percent within five trading days before subsequently losing roughly half of that gain, and MiniMax Group Inc. has recorded cumulative gains exceeding 500 percent since its January listing. As some issuers prepare for inclusion in the Shanghai‑Shenzhen‑Hong Kong Stock Connect mechanism, mainland investors will gain easier access; given the domestic preference for momentum trading, such participation could further amplify price swings and raise material risks for investors chasing rallies.
Institutional participation in these names is exceptionally limited, leaving prices largely driven by sentiment and capital flows rather than fundamentals. Exchange disclosures show that institutions required to report holdings account for only about 9.3 percent of MiniMax’s shares, with the remainder held by retail or non‑reportable investors. Among the five most volatile stocks in Asia, average institutional ownership is approximately 13 percent, compared with roughly 50 percent for established large caps such as Tencent Holdings and Alibaba Group. The weak institutional base helps explain extreme intraday moves: in March, MiniMax’s average daily intraday range reached about 14 percentage points, Zhipu’s daily high‑low spread approached 13 points, while Alibaba’s comparable figure was only 3.6 points.
The prevailing investment narrative centers on a rapid expansion in demand for “tokens,” the basic units that measure AI model data processing and underpin text generation and related outputs. Market participants have increasingly positioned token‑service providers such as MiniMax and Zhipu as primary beneficiaries of this trend. Bloomberg Intelligence projects that revenue growth for MiniMax and Zhipu could exceed 150 percent year‑on‑year by 2028, driven by direct monetization of token usage. MiniMax’s latest model, MiniMax M2.7, ranked third on OpenRouter’s list of popular models, surpassing offerings such as DeepSeek and models from Alibaba.
Despite promising product metrics, substantial operating losses remain a significant concern. MiniMax reported a net loss of USD 1.9 billion for the prior year, and Zhipu’s 2025 net loss widened by 60 percent year‑on‑year. Nevertheless, market participants have continued to bid these shares higher: Zhipu rallied as much as 35 percent intraday on Wednesday and MiniMax rose about 13 percent, reflecting investor willingness to look past current losses in anticipation of future token demand and pricing power.
Market professionals caution against indiscriminate momentum chasing. Jasmine Duan, Senior Investment Strategist at RBC Wealth Management Asia, observes that valuations for many AI newcomers are effectively predicated on their roadmaps and warns that unproven business models can change rapidly, increasing the risk of chasing rallies. Luo Jing, Investment Director at Value Partners Group, emphasizes that compute infrastructure remains a decisive factor in China’s AI landscape and that established industry leaders retain strategic advantages over the long term. Daniel So, Senior Trading Strategist at Goldhorse Capital Management, notes that clients tend to favor shorter holding periods for names such as MiniMax and Zhipu given persistent uncertainty about long‑term profitability; he characterizes these stocks as high‑potential but low‑certainty, and therefore likely to remain highly volatile.
