Two of China’s most closely watched AI developers are edging towards public markets, with Moonshot AI, the company behind the Kimi chatbot, lining up a flotation as rival DeepSeek attracts far more investor cash than it originally sought.
Beijing-based Moonshot AI has completed its last private funding round, valuing it at about $50 billion (€44.6bn), up from $31.5 billion (€28.1bn) in a round this summer, according to Bloomberg, which cited people familiar with the matter.
It is now preparing an initial public offering in Hong Kong, targeting the first quarter of 2027, that could raise up to $5 billion (€4.46bn), though the timing could still change, they added.
The company has also started organising early-look meetings to test investor demand, possibly from this month. Bank of America is coordinating the deal, with China International Capital Corp, Deutsche Bank and Goldman Sachs acting as sponsors, according to Bloomberg.
Moonshot’s annual recurring revenue has climbed from $300 million (€268mn) in June to about $1 billion (€892mn) and is expected to reach $2 billion (€1.78bn) by December.
The company was founded in early 2023 by Yang Zhilin, a graduate of Tsinghua University, often dubbed “China’s MIT”, who previously worked at Meta AI and Google Brain.
Backed by Chinese tech giants Alibaba and WeChat operator Tencent, the company drew global attention in July with its Kimi K3 open model.
If the IPO goes ahead next year, it would become the latest Chinese AI developer to tap Hong Kong investors, after rivals Zhipu and MiniMax listed in January, with MiniMax’s shares more than doubling on their first day of trading.
DeepSeek’s bumper round
Hangzhou-based DeepSeek is also close to securing at least 80 billion yuan (€10.6bn) in a round with Tencent and battery maker CATL among the largest investors, according to Bloomberg.
Strong investor demand could push the total towards 100 billion yuan (€13.3bn), double the roughly 50 billion yuan (€6.66bn) the start-up initially sought.
DeepSeek, which relied mainly on money from founder Liang Wenfeng’s hedge fund High-Flyer until this year, had reportedly targeted a valuation of about 500 billion yuan (€66.5bn) for the round.
Its low-cost R1 model rattled global markets in January 2025, wiping almost $600 billion (€535bn) off chipmaker Nvidia’s market value in a single day. DeepSeek has since released its V4 models, adding a faster V4.1-Flash version in September.
It also recently teamed up with Huawei to develop programming tools for the Chinese tech group’s Ascend AI chips, as Beijing pushes to cut its reliance on Nvidia.
Both companies are being investigated by the Cyberspace Administration of China, the country’s internet regulator, according to reports last month, over whether they routed sensitive user data to Claude, the chatbot of US firm Anthropic. The US-based AI company has accused them of using its model to train their own.
It is currently unclear how the probe could affect their IPO plans.
DeepSeek has hired Chinese brokerage CITIC Securities to prepare for a potential listing on Shanghai’s STAR Market, a board for technology companies, though the timing, valuation and size of any offering have yet to be decided and confirmed publicly.