Chinese AI startup MiniMax made a thunderous entrance into public markets, with its Hong Kong IPO turning into a full-blown spectacle.
Shares of the Shanghai-based company surged as much as 100% in their trading debut, rocketing from an offer price of HK$165 to highs near HK$330. At its peak, the rally briefly valued MiniMax at around $13 billion—an eye-catching figure for a company that’s barely three years old.
The demand was overwhelming. Retail investors oversubscribed the offering by more than 1,100 times, a level of enthusiasm that underscores just how hungry markets are for pure-play AI stories—especially ones coming out of China.
China’s AI startups are beating the US to the market
MiniMax’s debut didn’t happen in isolation. Just a day earlier, Beijing-based Zhipu jumped about 16% in its own Hong Kong listing. Together, the two companies are emerging as the public faces of China’s so-called “AI tigers”—startups racing to build large language models that can rival Western leaders.
What’s notable is the timing. While US heavyweights like OpenAI and Anthropic remain private and are still mapping out eventual IPOs, Chinese competitors are already tapping public capital. That head start is starting to matter.
Hong Kong, in particular, is becoming a launchpad: global enough to attract international investors, yet accessible for mainland tech firms navigating tighter domestic conditions.
More than just text: MiniMax’s multimodal bet
Founded in early 2022 by former SenseTime executive Yan Junjie, MiniMax has leaned hard into multimodal AI. Its models don’t just generate text—they work across audio, images, video, and even music.
That strategy appears to be paying off in reach. Monthly active users jumped from just over 3 million in 2023 to nearly 28 million by late 2025, according to IPO disclosures. Even more telling: less than one-third of its revenue comes from China. The rest is spread across the US, Singapore, and other international markets, signaling ambitions well beyond its home turf.
The money problem behind the hype
Still, the numbers reveal a tougher reality beneath the celebration.
MiniMax is losing money—lots of it. Heavy spending on cloud computing and research pushed losses past $200 million through September 2025. Annual costs to train and run its models hover around $400 million, a reminder of just how expensive the AI arms race has become.
Revenue is growing fast, but margins remain thin. Like many AI startups, MiniMax is betting that scale, efficiency gains, and time will eventually close the gap.
Why investors are piling in anyway
For now, investors seem willing to overlook the red ink. The logic is simple: whoever reaches scale first may define the next generation of AI platforms. MiniMax’s explosive debut suggests markets believe Chinese AI firms don’t just belong in that race—they may already be ahead on the funding track.
In a market starved for fresh AI plays, MiniMax didn’t just go public—it set the tone.