大模型公司月之暗面在最新融资轮中筹集35亿美元，大幅超过原定的10至20亿美元融资目标 [1][2][3]。融资后公司估值达到350亿美元，投前估值为315亿美元 [2]。这笔融资由国家人工智能产业投资基金领投 [2]，参与方包括龙珠资本、中国移动、清华资本、中金公司和高盛集团 [3]。
公司创始人杨植麟现年33岁，毕业于卡内基梅隆大学，曾在Meta和谷歌工作 [3]。月之暗面近期业绩增长迅速，年化营收从4月的2亿美元增至6月的3亿美元 [3]，其中API收入占比超过70% [2]。7月16日公司发布了Kimi K3大模型，参数规模达2.8万亿，上下文窗口为100万token [1][2]，K3发布后日销售量增加了6倍 [2]。7月27日，月之暗面开源了K3完整模型权重与技术报告 [1]。
公司已正式启动VIE及红筹架构拆除工作，计划于7月份完成 [2]，为上市做准备。月之暗面计划以500亿美元的投前估值启动Pre-IPO融资 [1][2][3]，力争在今年内赴港上市 [2][3]。从其C轮融资时的43亿美元估值来看，估值增长已超过10倍 [2]。本轮融资规模在中国大模型公司单轮融资中排名第二，仅次于DeepSeek今年6月的约510亿元融资 [1]。
Moonshot AI, a Beijing-based large language model startup, has completed a funding round that raised $3.5 billion, significantly exceeding its initial target of $1 billion to $2 billion [1][2]. The financing round values the company at $35 billion post-money, up from a pre-money valuation of $31.5 billion [2]. The oversubscription represented a 75% increase above the company's original fundraising cap [1].
The company plans to launch a Pre-IPO funding round targeting a pre-money valuation of $50 billion and aims to complete the process and list on the Hong Kong Stock Exchange by the end of this year [1][2]. To prepare for the public market transition, Moonshot AI has formally begun dismantling its Variable Interest Entity (VIE) and red-chip structure, with completion targeted for July [2]. The funding accelerates the company's expansion at a time when its open-source Kimi K3 model, released on July 16 and featuring 28 trillion parameters with a 1 million token context window [1][2][3], has generated significant international attention and disrupted competitive dynamics in Silicon Valley [3].
The funding reflects strong investor confidence in the company's commercial momentum. As of June, Moonshot AI's annualized recurring revenue (ARR) reached over $300 million, with API revenue accounting for more than 70% of total income [2]. The latest valuation represents growth of over 10 times from the company's Series C round, when its valuation stood at $4.3 billion [2]. In the broader context of China's AI industry, this round ranks as the second-largest single-round financing for a domestic large language model company, surpassed only by DeepSeek's approximately $5.1 billion financing in June this year [1].
