AI's honeymoon period appears to be over.
On July 9th, the Hong Kong-listed AI large language model company MINIMAX-WP (HKEX: 00100) faced its first post-IPO lock-up expiry. Under pressure from the share release, the company's stock price plummeted over 20% intraday, hitting a nearly six-month low. As of 2:20 PM, MINIMAX-WP shares were trading at HK$289.6, with its total market capitalization falling below HK$100 billion to approximately HK$90.8 billion.
MINIMAX-WP listed on the Hong Kong Stock Exchange on January 9th this year. On its debut, the stock surged nearly 110% above its issue price, closing with a market cap exceeding HK$105 billion. In March, the share price reached a high of HK$1,330, pushing its market value past HK$410 billion.
However, in just a few months, the company's stock price has been on a sustained downward trend. To date, compared to its historical peak, MINIMAX-WP's market value has evaporated by approximately HK$320 billion.
Just a day earlier, on July 8th, fellow AI star KNOWLEDGE ATLAS (HKEX: 02513) also saw its first lock-up expiry, yet its stock surged over 13% on the day. In contrast, MINIMAX-WP faced significantly greater volatility. Why did these two leading AI model companies experience such divergent market movements on their respective expiry days? Market consensus points to shareholder structure and the scale of the share release as key factors.
According to analyses by institutions like CICC, the first tranche of unlocked shares for MINIMAX-WP amounted to 146 million shares, representing a substantial 63% of its Hong Kong-listed share capital, with a relatively high proportion held by financial investors. Industry observers believe financial investors have stronger motivation and fewer constraints to exit, implying significant selling pressure post-expiry.
In comparison, the proportion of shares unlocked for KNOWLEDGE ATLAS on July 8th was relatively smaller. Furthermore, the largest block of unlocked shares was held by cornerstone investors with state-owned background. Multiple institutional investors, accounting for nearly 70% of the unlocked cornerstone shares, have explicitly stated their intention to hold long-term.
Previously, MINIMAX-WP's two core strategic shareholders, Alibaba and miHoYo, publicly expressed long-term confidence in the company's development and committed to holding their shares without selling. The founding team of MINIMAX-WP also voluntarily agreed to a 12-month lock-up period, meaning this expiry does not involve shares held by founders or employees. However, these factors were insufficient to offset the pressure from prevailing market sentiment.
Over the past six months, MINIMAX-WP's stock performance has been notably weaker than that of KNOWLEDGE ATLAS. The market cap gap between the two companies now exceeds HK$700 billion. Beyond shareholder structure, this disparity is closely linked to model competitiveness and commercialization expectations.
On June 1st, MINIMAX-WP launched its new flagship model, M3. Market reception post-launch was mixed, and its pricing strategy sparked controversy among developers, leading to a stock drop of over 15% that day.
On July 6th, Citi downgraded its target price for MINIMAX-WP to HK$533, attributing the recent stock correction to the lukewarm market reaction to the M3 model launched in early June. The bank noted that negative sentiment, coupled with uncertainties around user retention and monetization strategy, would likely pressure the stock in the short term.
Financially, MINIMAX-WP remains in a high-investment phase. Its financial report released in March showed that for 2025, the company's revenue was approximately $79.04 million, a year-on-year increase of 159%. However, its annual loss widened by 302% to $1.87 billion, with an adjusted net loss of $250 million.
Nonetheless, some institutions remain optimistic about the company's future commercialization. In a research report dated July 3rd, Goldman Sachs stated that MINIMAX-WP's earnings call conveyed strong commercialization signals, with management expressing confidence in achieving an Annual Recurring Revenue (ARR) target of $1 billion by the end of 2026. The report highlighted that the core catalyst lies in the Chinese AI large model industry reaching an inflection point in its pricing war, with industry pricing returning to rationality as competitors like DeepSeek announced price increases for peak usage.
For MINIMAX-WP, the first lock-up expiry may represent a short-term stress test. What will ultimately determine whether the company's valuation can recover is the future iteration of its model's capabilities and its commercial performance.