Alibaba and miHoYo refuse to cash out on MiniMax’s AI boom

Source Cryptopolitan

MiniMax is set to experience its first-ever share unlock on July 9 following its IPO. The company is an AI startup based in China that has more than doubled its value relative to its Hong Kong debut last January. Its two major strategic investors, namely Alibaba and miHoYo, claimed that they will not sell any shares, based on a report from Cai Lian She (Financial Associated Press) on June 22.

The pledge came at a time when China’s AI industry is growing rapidly and garnering funding that would be inconceivable just two years ago. MiniMax currently has an estimated user base of 300 million worldwide and more than one million enterprise developers. The decision by two of China’s biggest technology companies to hold their stakes in MiniMax instead of cashing out after a 400% surge highlights growing confidence in homegrown AI infrastructure and the long-term potential of China’s domestic AI ecosystem.

What the Minimax lock-up entails

The July 9 unlock relates to the restrictions placed on shares of pre-IPO shareholders whose lock-up periods end on the same day. The MiniMax founders took the initiative of locking up their shares for 12 months, double the required period, meaning that their shares and those held by employees will still be under restriction, Cai Lian She says.

According to Cai Lian She, Alibaba viewed its holding of the company’s shares as an opportunity to leverage the wider revolution in which artificial general intelligence (AGI) is expected to play. It said that it planned to enhance its cooperation with MiniMax in cloud computing and enterprise services. According to miHoYo, a Shanghai-headquartered video game developer behind Genshin Impact, AGI represents “the most important technology force of the next decade” and said its investment thesis aligns with MiniMax’s full-stack approach to multimodal AI.

China’s AI edge: speed, scale, cost

From the disclosures made by Minimax, the comments made by the founder quoted by Chinese media, and reports from Reuters on the growth strategy of the company, China’s strengths in relation to AI research and development lie in three main areas, which include speed, scale, and cost.

Speed. China’s AI firms are focusing on quickly launching products and commercializing them. According to comments made by the founder of MiniMax, the company is keen to maintain its pace of developing models and deploying products. The company has assured investors of quick expansion and diversification of its products following a strong 2025 revenue growth.

Scale. China’s AI developers enjoy the benefit of having large user bases domestically and a wide range of application ecosystems. MiniMax has expressed ambitions of becoming an AI platform firm globally and has diversified from foundation models to applications, including Talkie and Hailuo AI, as well as products for enterprises in order to test and monetize products at scale. Reuters has also observed that Chinese firms are competing fiercely with firms like DeepSeek and Zhipu in a fast-growing domestic market.

Cost. Reductions in the cost of inference and deployment have become a critical competitive theme among Chinese model builders. MiniMax has emphasized efficiency-oriented architectures and open models that promise lower computing costs, while Chinese AI firms in general have positioned themselves to provide high-performance systems at a lower cost than those from leading Western competitors.

Third-party analyses have highlighted MiniMax’s pricing advantage, with one comparison estimating M2.5 input costs at $0.15 per million tokens versus roughly $5 per million tokens for Anthropic’s Claude Opus 4.6. Several technology publications reported that MiniMax’s M1 training run cost only about $534,700, far below estimates for training GPT-4-class models. One article characterized this as “200 times lower” than GPT-4o’s estimated training cost.

Dual listing and the STAR Market strategy

MiniMax does not intend to stop at Hong Kong. According to Reuters, MiniMax submitted an A-share listing prospectus with the Shanghai securities authorities on May 29, with CITIC Securities as its sponsor, for listing in the STAR Market. By listing both in Hong Kong and on the mainland, MiniMax will be able to reach mainland Chinese investors, who traditionally assign a premium to tech stocks compared with those listed in Hong Kong.

MiniMax is racing Zhipu AI for the distinction of becoming the first major Chinese large language model company on the A-share market, according to Nikkei Asia. Both companies face the heavy capital expenditure demands of staying competitive in the AI race, Nikkei noted, which makes access to domestic capital markets a strategic priority rather than a vanity exercise.
Cryptopolitan has previously reported on how China’s regulatory environment, once seen as a constraint on its tech sector, has shifted toward actively supporting AI development. The STAR Market’s role as a destination for AI listings reflects that pivot.

What to watch?

MiniMax introduced its next-generation flagship model, M3, with a one-million-token context window and native multimodal abilities in early June 2026, as per Cai Lian She. The firm continues to be loss-making, with a net loss of $1.87 billion in 2025, but most of it was due to changes in the value of financial assets and not due to operating cash burn, as per Reuters.

The July 9 unlock will prove the investors’ belief or disbelief regarding Alibaba and miHoYo’s confidence. Annual recurring revenue of more than $300 million on a valuation of $33 billion means the price/revenue ratio is more than 100x, as per CryptoBriefing. It will be interesting to see how fast/slow the STAR Market application can be processed by Chinese authorities.

If you're reading this, you’re already ahead. Stay there with our newsletter.

Disclaimer: For information purposes only. Past performance is not indicative of future results.
placeholder
MicroStrategy Chair Michael Saylor Breaks 13-Week Bitcoin Buying RitualStrategy (MicroStrategy) may have skipped its weekly Bitcoin (BTC) purchase for the first time since late December, potentially ending a 13-week accumulation streak.Executive Chair Michael Saylor did
Author  Beincrypto
Mar 30, Mon
Strategy (MicroStrategy) may have skipped its weekly Bitcoin (BTC) purchase for the first time since late December, potentially ending a 13-week accumulation streak.Executive Chair Michael Saylor did
placeholder
MicroStrategy Posts $12.5 Billion Q1 2026 Loss on Bitcoin SlideMicroStrategy Inc posted a $12.54 billion net loss for the first quarter of 2026, the largest in the firm’s history. The deficit reflects a $14.46 billion unrealized markdown on its Bitcoin (BTC) hold
Author  Beincrypto
May 06, Wed
MicroStrategy Inc posted a $12.54 billion net loss for the first quarter of 2026, the largest in the firm’s history. The deficit reflects a $14.46 billion unrealized markdown on its Bitcoin (BTC) hold
placeholder
TON price soars 13% as Telegram revives original Gram token brandThe price of Toncoin surged more than 13% within 24 hours after Telegram announced a major branding shift that brings back the token’s original name, “Gram.” After long shelving its crypto network plans amid regulatory scrutiny, Telegram has taken control of The Open Network. The rebranding news sent the token to a high of $2.26...
Author  Cryptopolitan
Jun 02, Tue
The price of Toncoin surged more than 13% within 24 hours after Telegram announced a major branding shift that brings back the token’s original name, “Gram.” After long shelving its crypto network plans amid regulatory scrutiny, Telegram has taken control of The Open Network. The rebranding news sent the token to a high of $2.26...
placeholder
Trump Calls Stock Buybacks Fake: MicroStrategy Bitcoin Model Shows Another Way to Boost ValuationsPresident Donald Trump has again branded stock buybacks a fake way to lift share prices, yet the MicroStrategy Bitcoin model points to a different route to higher valuations, one built on issuing shar
Author  Beincrypto
11 hours ago
President Donald Trump has again branded stock buybacks a fake way to lift share prices, yet the MicroStrategy Bitcoin model points to a different route to higher valuations, one built on issuing shar
placeholder
Why are prediction market traders suddenly bearish on Nvidia's stock?Nvidia (NASDAQ: NVDA) stock is still green for 2026, but the trade no longer looks clean from the company that outperformed every other company and country in 2024 and 2025. NND is up about 12% this year, yet they have slipped roughly 3% over the past month. The gap with the rest of the chip...
Author  Cryptopolitan
11 hours ago
Nvidia (NASDAQ: NVDA) stock is still green for 2026, but the trade no longer looks clean from the company that outperformed every other company and country in 2024 and 2025. NND is up about 12% this year, yet they have slipped roughly 3% over the past month. The gap with the rest of the chip...
goTop
quote