There are many ways to capture the impact of the AI revolution, and the zooming valuations of AI companies is one of them.
As of March 10, 2026, MiniMax Group — a Shanghai-based AI startup that only went public in January — now commands a market capitalization of 345.31 billion HKD on the Hong Kong Stock Exchange, narrowly surpassing Baidu’s 318.31 billion HKD. It is a striking inversion: a two-year-old pure-play AI model company has overtaken one of China’s most established internet giants, a company that has been listed on public markets for over two decades.
A Rocket-Ship Debut
MiniMax’s rise has been swift. The company listed on the Hong Kong Stock Exchange on January 9, 2026, raising US$619 million in its IPO and trading under the symbolic stock code 0100.HK — a nod to binary code. Shares surged as much as 54% on its first day of trading, and retail investor subscriptions reached more than 1,830 times the available shares. Since then, the stock has continued to climb, posting gains of 162.25% from its listing price as of March 9, 2026.
Founded in December 2021 by former computer vision experts from SenseTime, MiniMax develops large language models across text, speech, music, images, and video. Its product suite — including MiniMax Agent, Hailuo AI video, and Talkie — now serves over 212 million users and 100,000 enterprise clients and developers across more than 200 countries. Notably, over 70% of its revenue comes from international markets, a remarkable feat for a Chinese AI startup.
Baidu’s Plateauing
The contrast with Baidu is stark. Once the dominant force in Chinese internet and search, Baidu’s stock is down 43.93% from its all-time high, having peaked at 214 HKD before falling steadily through 2022, 2023, and 2024. Despite Baidu’s own AI ambitions — most notably its Ernie Bot large language model — the market has not rewarded the company’s transformation efforts. Shares currently trade at around 120 HKD, giving it a market cap of 318.31 billion HKD, now below MiniMax’s.
This kind of generational transition in market value is precisely what investors mean when they talk about AI disruption. Baidu built its dominance on search and advertising in the pre-AI internet era. MiniMax, by contrast, was built from the ground up for the generative AI age.
Part of a Historic Wave
MiniMax was not the first Chinese AI company to go public in Hong Kong. Z.AI — formerly Zhipu AI — beat it by a day, listing on January 8, 2026, and raising approximately HK$4.35 billion ($560 million). Z.AI’s listing marked the first time a pure-play AI model company went public on any global stock exchange. MiniMax’s debut the very next day cemented what many observers are calling a historic moment for the AI industry: retail investors finally had a direct way to invest in companies building the frontier of AI, rather than accessing the technology only through large conglomerates like Alibaba, Microsoft, or Alphabet.
The back-to-back listings are likely the beginning of a broader trend. In the United States, OpenAI is reportedly preparing for what could be one of the largest IPOs in history, and Anthropic is also reportedly exploring public market options. The success of MiniMax and Z.AI will be closely watched by every AI company contemplating a listing in 2026 and beyond.
Strong Models, Fierce Competition
MiniMax’s valuation is not purely a story of investor euphoria. The company has been steadily advancing its technical capabilities. Its MiniMax M2 model placed fifth on the Artificial Analysis Intelligence Index, beating Google’s Gemini 2.5 Pro and becoming the highest-ranked Chinese model on that leaderboard at the time. More recently, its M2.5 model outperformed GPT-5.2 and Gemini 3 Pro on the SWE-Bench coding benchmark, with MiniMax claiming SWE-Bench scores rose from 56% (M1) to 80.2% (M2.5) in under a year.
That said, the picture is more nuanced on some benchmarks. On ARC-AGI-2, Chinese models including MiniMax’s M2.5 scored below 12%, still behind the scores achieved by top US frontier labs as of mid-2025. And competition from other Chinese AI companies is relentless — DeepSeek, Moonshot AI, and Z.AI are all pushing hard on both model capability and cost efficiency.
MiniMax has also not been immune to controversy. Anthropic has accused MiniMax and other Chinese AI companies of distilling capabilities from its Claude models, and has claimed that Chinese model benchmark performance is partly the result of optimization rather than genuine general intelligence improvements. MiniMax and its peers dispute these characterizations.
The Financials Behind the Valuation
MiniMax’s first earnings report as a public company showed full-year 2025 revenue of $79 million, up 158.9% year-over-year. Gross margins more than doubled to 25.4%, up from 12.2% in 2024. The adjusted net loss was $250.9 million — essentially flat year-over-year despite revenues more than doubling, which investors tend to view as a positive sign of improving unit economics. The company is well-capitalized, holding over $1 billion in cash as of December 31, 2025, further bolstered by IPO proceeds.
MiniMax is positioning its next phase as a transition from a large-model company to a platform company for the AI era, betting on AI coding agents, workplace productivity tools, and long-form multimodal content creation as its growth vectors.
What This Moment Means
The fact that a two-year-old AI startup now outvalues one of China’s most storied internet companies is more than a market curiosity. It reflects a broader realignment of where investors believe value will be created in the next decade of technology. The old guard — built on search, advertising, and the mobile internet — is being weighed against the new guard: companies whose entire existence is premised on the transformative potential of large language models.
Whether MiniMax can sustain and justify this valuation over the long term is an open question. The challenges are real: AI models are commoditizing, compute costs remain enormous, and geopolitical tensions add regulatory risk. But for now, the market has spoken — and it has placed a bet that the future belongs to AI-native companies, not the incumbents trying to adapt to the AI era.