MiniMax Shares Plummet by 80%

Author: Shen Xiao

On July 13, shares of Hong Kong-listed AI giant MiniMax (00100.HK) suffered another steep decline, closing down 17.13% at HK$222.60 per share, with a total market capitalization of HK$69.82 billion. This marks a staggering drop of over 80% from its peak of HK$1,330 per share in March this year, with its market value having shrunk by approximately HK$340 billion from its zenith of over HK$410 billion.

Just three days earlier, on July 10, MiniMax unveiled a dual strategy: securing HK$16 billion in new financing while founder and CEO Yan Junjie announced in a company-wide letter that he would forgo his salary until the company achieves Artificial General Intelligence (AGI). Additionally, he pledged to transfer 5% of his personal stake in the company, allocating 4% for team incentives and 1% to an open-source community fund.

On one hand, the stock price has been nearly halving repeatedly; on the other, the founder has made a bold personal commitment. What has befallen this company, which was hailed as a "rising AI star" in the Hong Kong stock market just four months ago?

HK$340 Billion Vanishes in 4 Months: What’s Going On with MiniMax?

MiniMax was once the darling of investors, completing its journey from incorporation to Hong Kong Stock Exchange listing in just four years—one of the fastest AI IPOs globally at the time.

Founder Yan Junjie, who holds a Ph.D. from the Institute of Automation, Chinese Academy of Sciences, left his position as Vice President of the Research Institute and CTO of the Smart City Business Group at SenseTime at the end of 2021 to start his own venture.

On January 9, 2026, MiniMax made its debut on the Hong Kong Stock Exchange at an issue price of HK$165 per share. The stock soared on its first day, closing up nearly 110%, with its market value briefly exceeding HK$105 billion, eclipsing Zhipu, which had listed just the day before.

On March 18, MiniMax launched its new-generation model M2.7, triggering a brief intra-day surge of over 28% in its stock price, which peaked at HK$1,330 per share, giving the company a market value exceeding HK$410 billion and setting the benchmark for AI sector valuations in the Hong Kong stock market.

However, MiniMax has not revisited those heights since. The turning point came on June 1, when the company unveiled its flagship model M3 alongside a new token-based billing system, replacing the previous subscription model.

The new model’s resource consumption patterns caught many users off guard, as their usage quotas depleted far faster than anticipated, with little prior explanation from the company. Dissatisfaction quickly spread across developer forums and social media, with some users flocking to complaint platforms to demand refunds. That day, MiniMax’s stock price gapped down at the open and closed 15.71% lower.

A week later, MiniMax was forced to announce a permanent 50% price cut for M3, bringing it back in line with the previous model’s pricing.

This controversy also prompted a re-evaluation of MiniMax’s standing as a leader in model capabilities. The company had positioned itself as one of only four global first-tier full-modality AI companies, alongside OpenAI, Google, and ByteDance. Yet, according to independent evaluation agency Artificial Analysis’s intelligence index rankings, M3 placed only ninth globally and fell outside the top 40-50 in the more user experience-focused Chatbot Arena rankings.

Li Zeming, Chief Investment Officer at Red Ant Capital, explained that the core valuation metric for large model companies in the Hong Kong stock market is the price-to-sales ratio (market capitalization divided by revenue). "Whenever a large model company significantly cuts prices, the entire sector tends to plummet," he noted. MiniMax’s price reduction directly lowered market expectations for its revenue, dragging down its valuation.

In mid-June, JPMorgan downgraded MiniMax’s rating from "Overweight" to "Neutral" and slashed its target price from HK$1,100 per share to HK$400 per share, citing the lack of a new cutting-edge domestic model since M2 and MiniMax’s ongoing catch-up phase in pure model capabilities.

On July 6, Citigroup further lowered its target price to HK$533 per share, noting the tepid market reaction to the M3 release in early June and persistent uncertainties regarding user retention and monetization strategies, which continued to weigh on the stock price.

The erosion of the company’s model narrative marked the first major crack in this round of declines, dragging down valuations. However, the true catalyst for the market sentiment collapse was the subsequent lock-up expiration.

Relentless Stock Price Declines: Lock-Up Expiration Deals the Final Blow

On July 9, MiniMax faced its first major lock-up expiration since listing, with around 45% of its total share capital becoming freely tradable, significantly expanding the float.

The accumulated selling pressure was unleashed at the market open: the stock price gapped down sharply in the morning, plunging over 21% intra-day before closing 17.98% lower, with the market value evaporating by over HK$20 billion in a single day.

The following day, July 10, MiniMax announced its HK$16 billion financing round and Yan Junjie’s "no salary" pledge, but market sentiment failed to reverse; the stock price fell another 9.68%, closing at HK$268.60 per share.

Why did these seemingly substantial measures fail to halt the decline? Because the lock-up expiration was just one factor; the market was already re-pricing based on deeper, long-simmering concerns.

Beyond the cracks in its model capabilities narrative, MiniMax’s C-end commercialization efforts were also under pressure. In 2025, roughly 67% of the company’s revenue came from C-end AI products, primarily the overseas emotional companionship app Talkie, the domestic Xingye, and the video tool Hailuo AI.

The high valuations at the time of listing were largely betting on the story of "C-end globalization." However, industry observers noted that the monthly active users of Talkie and Xingye declined by about 60% quarter-over-quarter in the fourth quarter of 2025, while Hailuo AI had been surpassed by multiple models from Alibaba, ByteDance, and Kuaishou in Artificial Analysis’s video model rankings, losing its leading position.

Even more challenging was the mounting pressure from sustained losses. According to the company’s 2025 annual report, MiniMax’s total revenue for the year was approximately $79.038 million, up 158.9% year-over-year, with over 70% coming from overseas markets. However, its adjusted net loss was approximately $250 million. Revenue growth and capital consumption accelerated almost in tandem, which is why MiniMax urgently presented the HK$16 billion financing plan on the same day as its stock price plummeted.

These pressures are not unique to MiniMax. A glance at Zhipu, which listed around the same time, reveals that despite similar lock-up expirations and financing efforts, the capital markets’ responses were diametrically opposite.

Why Did Markets React Differently to Similar Financings?

On July 8, Zhipu completed the lock-up expiration for its cornerstone investors, with its stock price rising over 13% that day. On July 9, Zhipu announced a placement financing of approximately HK$31.4 billion, offering up to 19.78 million new H-shares at HK$1,588 per share, a discount of about 12.99% from the previous trading day’s closing price. This scale set a new record for single placement fundraising by Hong Kong-listed technology companies in 2026 and represented the largest equity financing by a domestic large model company after listing.

On the day of the announcement, Zhipu’s stock price rose 11.34%, with its market value rebounding to HK$905.953 billion. One day later, on July 10, MiniMax announced the completion of its HK$16 billion financing, but its stock price still fell 9.68% that day. Despite both being large financing announcements, the market reactions were completely different.

Behind this divergence lies a fundamental difference in business models: Zhipu focuses on localized deployment, enterprise-grade APIs, and developer platforms, with revenue primarily coming from domestic government and enterprise clients, having served over 12,000 institutions as of May 2026. MiniMax, on the other hand, primarily targets C-end AI native applications, relying on advertising and subscription revenue from emotional companionship products, with a relatively "lighter" business model but one that struggles to maintain gross margins amid soaring computing costs.

The two companies’ revenue scales were roughly comparable in 2025: Zhipu at RMB 724 million and MiniMax at approximately RMB 569 million, but their gross margins showed a significant gap: Zhipu at 41% and MiniMax at only 25.4%. This 16-percentage-point difference reflects the true cost structures of the two business models.

The capital markets’ "vote with their feet" was equally telling: as of May 29, 2026, Zhipu’s stock price touched a historical high of HK$1,993 per share, with its market value nearing HK$900 billion, surpassing Xiaomi. Meanwhile, MiniMax’s market value had fallen to HK$209.4 billion, less than one-third of Zhipu’s. By June 22, Zhipu’s total market value once exceeded HK$1 trillion, with cumulative year-to-date gains exceeding 1900%.

However, the core capital allocation directions of both companies are actually highly aligned, both pointing towards AI infrastructure and model research and development. Eighty percent of the net proceeds from MiniMax’s HK$16 billion fundraising will be used to continuously strengthen AI infrastructure and model R&D. Zhipu’s HK$31.4 billion fundraising is also planned to be fully invested in core R&D, computing infrastructure, and commercial expansion by the end of 2027, while it is also advancing its listing on the A-share Science and Technology Innovation Board, establishing a "A+H" dual-listing structure.

The divergence between MiniMax and Zhipu is, to some extent, a microcosm of the financing rhythm of the entire domestic large model industry in 2026. DeepSeek has completed its first round of financing exceeding RMB 50 billion, while companies like Yuezhiyimian and Jieyuexingchen are also intensively advancing new rounds of financing or listing plans.

Going public is no longer the endpoint for financing but has become the starting point for a new round of capital operations, a phenomenon extremely rare in traditional industries but increasingly becoming the norm for large model companies: training, inference, computing power lock-in, and globalization investments all require continuous capital support, with no breaks in the technological race.

In his company-wide letter on July 10, Yan Junjie wrote that while there may be market fluctuations and external noise, the direction of progress remains unchanged. The team on the front lines of the industry understands the true pace of technological evolution better than anyone and is also clearer about the long-term value the company is creating and accumulating.

The founder’s decision to forgo salary and contribute personal shares often signifies a public endorsement of the company’s long-term beliefs with personal stakes. However, whether the market will re-price accordingly does not lie in this letter but will be answered by MiniMax’s subsequent model capability iterations, commercialization progress, and profitability path.

References:

"Financing HK$16 Billion, Founder Announces 'Zero Salary': Can MiniMax Rewrite the Endgame of the Large Model Capital Race?" by Entrepreneur China;

"MiniMax Raises HK$16 Billion Overnight, but the Large Model Race is Far from Over" by China Money Network;

"In-Depth | I Witnessed MiniMax's 'Darkest Hour'" by Guixingren Pro;

"From HK$410 Billion to HK$100 Billion: What’s Happening with MiniMax?" by Dingjiao One;

"MiniMax Shakes on First Day of Lock-Up Expiration" by Yicai;

"Just Four Years from Seed Round of RMB 1.2 Billion to HK$80 Billion Market Value, MiniMax Just IPO'd" by Entrepreneur China;

"Zhipu's Market Value Surpasses HK$1 Trillion" by Cailian Press;

"Who Exactly is China's First Large Model Stock 'Zhipu'?" by 36Kr.

THE END

Copyright and Disclaimer

1. Content Copyright: Except for quoted public data, policies, and case studies, all content herein is original. Professional data is sourced from authorized databases and government websites, while case studies are compiled from real events.

2. Image Licensing: Some images herein are proprietary or officially licensed, with others generated by AI; for network images without clear copyright, ownership remains with the original authors, and removal will be executed upon notification.

3. Republication Guidelines: Unauthorized republication is prohibited; full attribution and author credit must be retained for permitted republications.

4. Liability Disclaimer: This article represents the author’s observations on business figures and industry commentary based on public information, intended for reference only and not constituting professional advice. Risks arising from use are borne by the user. Xinzhidongneng reserves the final interpretation rights for this article.