The market reaction was immediate and brutal. On July 9, shares opened at HK$359.8 and quickly came under heavy selling pressure. By the afternoon, the stock hit an intraday low of HK$290, down roughly 20%, before closing at HK$297.4, a 17.98% plunge . Trading volume surged to roughly 12 million shares, about six times the daily average . The company's market capitalization fell to roughly HK$92.8 billion ($11.8 billion) .

This selloff was not a complete surprise. Analysts and media had been warning for weeks. MiniMax stock had already fallen roughly 53% in the first two weeks of June as the lock-up expiry approached . By July 3, the stock was down 72% from its all-time high . Caixin reported on June 12 that shares in both MiniMax and rival Zhipu AI had fallen by about half over two weeks ahead of their respective lock-up expirations .

Crucially, the lock-up expiry did not involve MiniMax's founding team or employees, who had voluntarily set a 12-month lock-up period . The shares being freed belonged to pre-IPO cornerstone investors, including Alibaba, which held roughly 13% of the company as the largest single external shareholder . Alibaba and miHoYo (with about 5.24%) had expressed long-term support, but the market still braced for a selloff .

The Deal Structure: A Two-Tranche Rescue

On the evening of July 9, hours after the market closed, MiniMax launched a simultaneous two-tranche capital raise :

• Share placement: 35.6 million new Class A shares at HK$268 each, raising about HK$9.54 billion ($1.22 billion) . The placement price represented a 9.89% discount to the last closing price of HK$297.40 .
• Zero-coupon convertible bonds: HK$6.5 billion ($830 million) in zero-coupon, secured convertible bonds due July 2027, with an initial conversion price of HK$335 per share (a 25% premium to the placement price) .
• Total raise: HK$16.04 billion ($2.05 billion) .
• Bookrunners: Morgan Stanley and UBS acted as joint placement agents .
• Oversubscription: Despite the stock's collapse, the deal was reported to be over seven times oversubscribed, reflecting strong institutional demand from sovereign wealth funds, long-only institutions, top-tier Chinese funds, and multi-strategy hedge funds across Asia, Europe, and the U.S. .

The speed of the execution was remarkable: the company disclosed the lock-up expiry on July 8, watched the stock fall 18% on July 9, and launched the capital raise that same evening, with the deal announced before the Hong Kong market opened on July 10 .

Use of Proceeds: Betting on AGI Infrastructure

Per the Hong Kong exchange filing, proceeds were allocated as follows :

• ~80% for AI infrastructure and model R&D, including a planned 2.7 trillion-parameter model .
• ~10% for accelerating commercialization of the "Harness" AI agent product line .
• ~10% for working capital and general corporate purposes .

This allocation reflects MiniMax's aggressive bet on scale. The company has been one of China's most ambitious AI startups, and the 2.7 trillion-parameter model would be among the largest ever attempted .

Customer Growth: From 200K to 1 Million in Six Months

The capital raise came as MiniMax disclosed explosive customer growth. By the end of June 2026, the company's global enterprise and developer customer base exceeded 1 million, up roughly fivefold from ~200,000 at the end of 2025 . Consumer products reached about 300 million global users .

This growth was fueled by successful products like Talkie, MiniMax Code, and the open platform. By February 2026, annualized recurring revenue (ARR) had already crossed $150 million, with over 70% of revenue coming from international markets .

CEO Yan Junjie's Pledge: No Salary Until AGI

On July 10, along with the capital raise announcement, CEO Yan Junjie issued an internal memo that became a major media story :

• He will forgo his salary until the company achieves artificial general intelligence (AGI) .
• He will transfer 5% of his personal equity—4% for core team incentives and 1% to an open-source developer community fund .

Yan, 36, who also serves as board chairman and CTO, framed this as a long-term commitment: "Effective today, and until the day we achieve AGI, I will no longer accept any salary from the company" . The SCMP noted the move came as MiniMax's stock had plunged 80% from its peak .

Contrast with Zhipu AI: State Capital vs. Market-Driven

MiniMax's rival Zhipu AI faced a similar lock-up expiry at roughly the same time. The SCMP reported that Zhipu and MiniMax both saw heavy pre-lock-up selloffs, with Zhipu's shares also down roughly 50% from their highs in the weeks before July 9 .

A key difference emerged in capital structure. While MiniMax's raise drew heavily from international institutional investors (sovereign funds, long-only funds, multi-strategy hedge funds) , Zhipu AI's capital structure featured significant state-capital participation, including Beijing state-backed funds . This contrast highlights the diverging strategies of China's top AI startups: MiniMax leaning into global private markets, and Zhipu AI anchoring itself closer to government-linked capital.

Summary

The market reaction was immediate and brutal. On July 9, shares opened at HK$359.8 and quickly came under heavy selling pressure. By the afternoon, the stock hit an intraday low of HK$290, down roughly 20%, before closing at HK$297.4, a 17.98% plunge . Trading volume surged to roughly 12 million shares, about six times the daily average . The company's market capitalization fell to roughly HK$92.8 billion ($11.8 billion) .

This selloff was not a complete surprise. Analysts and media had been warning for weeks. MiniMax stock had already fallen roughly 53% in the first two weeks of June as the lock-up expiry approached . By July 3, the stock was down 72% from its all-time high . Caixin reported on June 12 that shares in both MiniMax and rival Zhipu AI had fallen by about half over two weeks ahead of their respective lock-up expirations .

Crucially, the lock-up expiry did not involve MiniMax's founding team or employees, who had voluntarily set a 12-month lock-up period . The shares being freed belonged to pre-IPO cornerstone investors, including Alibaba, which held roughly 13% of the company as the largest single external shareholder . Alibaba and miHoYo (with about 5.24%) had expressed long-term support, but the market still braced for a selloff .

The Deal Structure: A Two-Tranche Rescue

On the evening of July 9, hours after the market closed, MiniMax launched a simultaneous two-tranche capital raise :

• Share placement: 35.6 million new Class A shares at HK$268 each, raising about HK$9.54 billion ($1.22 billion) . The placement price represented a 9.89% discount to the last closing price of HK$297.40 .
• Zero-coupon convertible bonds: HK$6.5 billion ($830 million) in zero-coupon, secured convertible bonds due July 2027, with an initial conversion price of HK$335 per share (a 25% premium to the placement price) .
• Total raise: HK$16.04 billion ($2.05 billion) .
• Bookrunners: Morgan Stanley and UBS acted as joint placement agents .
• Oversubscription: Despite the stock's collapse, the deal was reported to be over seven times oversubscribed, reflecting strong institutional demand from sovereign wealth funds, long-only institutions, top-tier Chinese funds, and multi-strategy hedge funds across Asia, Europe, and the U.S. .

The speed of the execution was remarkable: the company disclosed the lock-up expiry on July 8, watched the stock fall 18% on July 9, and launched the capital raise that same evening, with the deal announced before the Hong Kong market opened on July 10 .

Use of Proceeds: Betting on AGI Infrastructure

Per the Hong Kong exchange filing, proceeds were allocated as follows :

• ~80% for AI infrastructure and model R&D, including a planned 2.7 trillion-parameter model .
• ~10% for accelerating commercialization of the "Harness" AI agent product line .
• ~10% for working capital and general corporate purposes .

This allocation reflects MiniMax's aggressive bet on scale. The company has been one of China's most ambitious AI startups, and the 2.7 trillion-parameter model would be among the largest ever attempted .

Customer Growth: From 200K to 1 Million in Six Months

The capital raise came as MiniMax disclosed explosive customer growth. By the end of June 2026, the company's global enterprise and developer customer base exceeded 1 million, up roughly fivefold from ~200,000 at the end of 2025 . Consumer products reached about 300 million global users .

This growth was fueled by successful products like Talkie, MiniMax Code, and the open platform. By February 2026, annualized recurring revenue (ARR) had already crossed $150 million, with over 70% of revenue coming from international markets .

CEO Yan Junjie's Pledge: No Salary Until AGI

On July 10, along with the capital raise announcement, CEO Yan Junjie issued an internal memo that became a major media story :

• He will forgo his salary until the company achieves artificial general intelligence (AGI) .
• He will transfer 5% of his personal equity—4% for core team incentives and 1% to an open-source developer community fund .

Yan, 36, who also serves as board chairman and CTO, framed this as a long-term commitment: "Effective today, and until the day we achieve AGI, I will no longer accept any salary from the company" . The SCMP noted the move came as MiniMax's stock had plunged 80% from its peak .

Contrast with Zhipu AI: State Capital vs. Market-Driven

MiniMax's rival Zhipu AI faced a similar lock-up expiry at roughly the same time. The SCMP reported that Zhipu and MiniMax both saw heavy pre-lock-up selloffs, with Zhipu's shares also down roughly 50% from their highs in the weeks before July 9 .

A key difference emerged in capital structure. While MiniMax's raise drew heavily from international institutional investors (sovereign funds, long-only funds, multi-strategy hedge funds) , Zhipu AI's capital structure featured significant state-capital participation, including Beijing state-backed funds . This contrast highlights the diverging strategies of China's top AI startups: MiniMax leaning into global private markets, and Zhipu AI anchoring itself closer to government-linked capital.

Summary