#MoonshotAIPreIPOs开启 Behind Kimi’s $3.5 billion financing: The valuation logic of the foundation-model sector is being restructured
In August 2026, a financing announcement once again set China’s AI industry abuzz.
Moonshot AI announced the completion of its Series F financing, raising more than $3.5 billion and reaching a post-money valuation of $35 billion. Even more strikingly, the Pre-IPO round originally planned to launch in August has already started ahead of schedule, with the market-reported target pre-money valuation soaring to $50 billion.
What does this mean? A Chinese foundation-model company founded just over three years ago now has a valuation higher than that of the vast majority of A-share listed companies. From the $4.3 billion valuation in its Series C at the end of last year to its current push toward $50 billion, its valuation has multiplied more than tenfold in just over half a year.
Many people cannot understand it: How can an AI company that is still making massive investments and has yet to turn a profit be worth so much?
Today, let’s discuss what exactly happened behind Kimi’s latest financing round.
I. The phenomenon: Capital is voting with its money—and moving fast
First, let’s look at the fundamentals of this financing round.
According to media reports, Moonshot AI’s Series F financing exceeded $3.5 billion, with the National Artificial Intelligence Industry Investment Fund and Alibaba jointly leading the round. Because subscriptions exceeded the original target by more than three times, the company directly closed the subscription window ahead of schedule. This is not a normal financing pace. From the Series C at the end of 2025, to three consecutive rounds of financing in early 2026, to the $2 billion Series D led by Meituan Longzhu in May, and then the Series F in July, Moonshot AI completed multiple large-scale financing rounds in less than a year. Its cumulative financing has already far exceeded that of the vast majority of Chinese AI startups. Even more noteworthy is its shareholder structure. Alibaba, the National Social Security Fund, China Mobile, CITIC Industrial Fund, Meituan Longzhu… Put together, these names include internet giants, national-team funds, and industrial capital. This is not merely a venture-capital bet, but the entire Chinese technology industry betting on the future of foundation models. The same is true elsewhere. Zhipu and MiniMax have successively listed in Hong Kong, with their market capitalizations repeatedly reaching new highs; DeepSeek’s valuation also surpassed $50 billion after raising financing. China’s top-tier foundation-model companies are collectively entering the “capital finals.”
II. The turning point: The release of Kimi K3 gave the valuation a new narrative
Capital is willing to provide money on the premise that it sees potential. In July 2026, Moonshot AI released its next-generation flagship model, Kimi K3. This was not an ordinary version update. K3 has 2.8 trillion parameters and uses a mixture-of-experts (MoE) architecture, activating only 16 routed experts at a time, for 104 billion actively activated parameters. It supports an ultralong context of 1 million Tokens, offers native visual understanding, and is fully open-sourced. On Artificial Analysis’s globally authoritative Intelligence Index, K3’s overall score ranked just behind Anthropic’s Claude Fable 5 and OpenAI’s GPT-5.6, placing it among the top three worldwide. On the day of its release, K3 surged to first place on Hugging Face’s global trending chart. Forty-eight hours later, user request volume approached the cluster’s capacity limit, forcing Moonshot AI to announce the suspension of new consumer subscriptions and prioritize all computing power for existing paid users. The last time a product became so popular that its company had to “rate-limit” it was when ChatGPT had just launched. But K3’s significance goes beyond impressive technical specifications. More importantly, it gave the market a fresh understanding of Moonshot AI’s technological potential. After DeepSeek emerged with high performance at low cost, many people were asking: What differentiated advantages does Kimi still have? K3’s answer is that Kimi remains competitive on three key battlegrounds: long context, multimodality, and the open-source ecosystem.
III. The confidence: Commercialization has turned from a “story” into “numbers”
Even excellent technology is merely a laboratory toy if it cannot be monetized. The real anchor behind Moonshot AI’s valuation surge this round is its commercialization data. In March 2026, the company’s annual recurring revenue (ARR) surpassed $100 million. In May, it surpassed $200 million. By mid-June, it surpassed $300 million. It tripled in three months. More importantly, the revenue structure. API revenue now accounts for more than 70% of total revenue, with enterprise customers becoming the primary source of income. This means Kimi is no longer merely a consumer-facing application for users to chat and ask questions, but is becoming enterprise-level AI infrastructure. Enterprises’ willingness to pay and retention are far higher than those of individual users. Growth in API calls brings predictable, sustainable cash flow.
According to media reports, Kimi’s overseas paying users grew 400%, API revenue grew 400%, and the product has entered more than 200 countries and regions. One customer even offered a tens-of-millions-of-dollars spending commitment and prepaid guarantee solely to secure priority API access.
In August 2026, a financing announcement once again set China’s AI industry abuzz.
Moonshot AI announced the completion of its Series F financing, raising more than $3.5 billion and reaching a post-money valuation of $35 billion. Even more strikingly, the Pre-IPO round originally planned to launch in August has already started ahead of schedule, with the market-reported target pre-money valuation soaring to $50 billion.
What does this mean? A Chinese foundation-model company founded just over three years ago now has a valuation higher than that of the vast majority of A-share listed companies. From the $4.3 billion valuation in its Series C at the end of last year to its current push toward $50 billion, its valuation has multiplied more than tenfold in just over half a year.
Many people cannot understand it: How can an AI company that is still making massive investments and has yet to turn a profit be worth so much?
Today, let’s discuss what exactly happened behind Kimi’s latest financing round.
I. The phenomenon: Capital is voting with its money—and moving fast
First, let’s look at the fundamentals of this financing round.
According to media reports, Moonshot AI’s Series F financing exceeded $3.5 billion, with the National Artificial Intelligence Industry Investment Fund and Alibaba jointly leading the round. Because subscriptions exceeded the original target by more than three times, the company directly closed the subscription window ahead of schedule. This is not a normal financing pace. From the Series C at the end of 2025, to three consecutive rounds of financing in early 2026, to the $2 billion Series D led by Meituan Longzhu in May, and then the Series F in July, Moonshot AI completed multiple large-scale financing rounds in less than a year. Its cumulative financing has already far exceeded that of the vast majority of Chinese AI startups. Even more noteworthy is its shareholder structure. Alibaba, the National Social Security Fund, China Mobile, CITIC Industrial Fund, Meituan Longzhu… Put together, these names include internet giants, national-team funds, and industrial capital. This is not merely a venture-capital bet, but the entire Chinese technology industry betting on the future of foundation models. The same is true elsewhere. Zhipu and MiniMax have successively listed in Hong Kong, with their market capitalizations repeatedly reaching new highs; DeepSeek’s valuation also surpassed $50 billion after raising financing. China’s top-tier foundation-model companies are collectively entering the “capital finals.”
II. The turning point: The release of Kimi K3 gave the valuation a new narrative
Capital is willing to provide money on the premise that it sees potential. In July 2026, Moonshot AI released its next-generation flagship model, Kimi K3. This was not an ordinary version update. K3 has 2.8 trillion parameters and uses a mixture-of-experts (MoE) architecture, activating only 16 routed experts at a time, for 104 billion actively activated parameters. It supports an ultralong context of 1 million Tokens, offers native visual understanding, and is fully open-sourced. On Artificial Analysis’s globally authoritative Intelligence Index, K3’s overall score ranked just behind Anthropic’s Claude Fable 5 and OpenAI’s GPT-5.6, placing it among the top three worldwide. On the day of its release, K3 surged to first place on Hugging Face’s global trending chart. Forty-eight hours later, user request volume approached the cluster’s capacity limit, forcing Moonshot AI to announce the suspension of new consumer subscriptions and prioritize all computing power for existing paid users. The last time a product became so popular that its company had to “rate-limit” it was when ChatGPT had just launched. But K3’s significance goes beyond impressive technical specifications. More importantly, it gave the market a fresh understanding of Moonshot AI’s technological potential. After DeepSeek emerged with high performance at low cost, many people were asking: What differentiated advantages does Kimi still have? K3’s answer is that Kimi remains competitive on three key battlegrounds: long context, multimodality, and the open-source ecosystem.
III. The confidence: Commercialization has turned from a “story” into “numbers”
Even excellent technology is merely a laboratory toy if it cannot be monetized. The real anchor behind Moonshot AI’s valuation surge this round is its commercialization data. In March 2026, the company’s annual recurring revenue (ARR) surpassed $100 million. In May, it surpassed $200 million. By mid-June, it surpassed $300 million. It tripled in three months. More importantly, the revenue structure. API revenue now accounts for more than 70% of total revenue, with enterprise customers becoming the primary source of income. This means Kimi is no longer merely a consumer-facing application for users to chat and ask questions, but is becoming enterprise-level AI infrastructure. Enterprises’ willingness to pay and retention are far higher than those of individual users. Growth in API calls brings predictable, sustainable cash flow.
According to media reports, Kimi’s overseas paying users grew 400%, API revenue grew 400%, and the product has entered more than 200 countries and regions. One customer even offered a tens-of-millions-of-dollars spending commitment and prepaid guarantee solely to secure priority API access.
