Kimi Secures $2 Billion in Funding
On May 6, it was reported that Kimi, also known as “月之暗面,” is set to complete a new funding round of approximately $2 billion, pushing its post-money valuation above $20 billion. This round is led by Meituan’s Longzhu, with participation from China Mobile, CPE (CITIC Industry Fund), and others, with Longzhu alone contributing over $200 million.
From January to February this year, Kimi completed three rounds of financing totaling around $1.9 billion. Including this latest round, the company has raised over $3.9 billion in less than six months, surpassing MiniMax and Zhizhu to become the top AI model startup in China by total funding.
Financing Progress: Rapid Growth in Valuation
Kimi’s financing progress has been concentrated and rapid. In November last year, the company completed a round of approximately $500 million at a valuation of about $4.3 billion. Between January and February this year, it completed three rounds of financing totaling $1.9 billion, with valuations increasing from $10 billion to approximately $18 billion. Existing shareholders Alibaba, Tencent, and 5Y Capital also participated in the $10 billion valuation round. Following the completion of the new $2 billion financing, the post-money valuation will exceed $20 billion.
Compared to its competitors listed on the Hong Kong stock exchange, Kimi’s current valuation, approximately RMB 140 billion, is still lower than MiniMax’s RMB 210 billion and Zhizhu’s RMB 347 billion market capitalization. Some investors view this valuation gap as an opportunity for further investment.
Performance Driven: Kimi Claw Fuels Growth
The confidence behind this round of financing is driven by explosive revenue growth. In January, Kimi launched Kimi Claw, powered by the K2.5 model, which enables one-click deployment of cloud-based intelligent agents, making it one of the first companies in China to commercialize the OpenClaw trend.
According to data from global payment platform Stripe, Kimi’s revenue in the last 20 days of January exceeded its total revenue for the entire year of 2025. The number of paid subscription orders from individual users in January increased by over 8000% month-on-month, with a further growth of over 120% in February, placing Kimi among the top ten on Stripe’s global leaderboard. Additionally, data from Similarweb shows that Kimi’s overseas API open platform saw a daily visit increase of 10 to 20 times following the K2.5 release.
On April 20, Kimi released and open-sourced its latest model, K2.6, which enhances programming capabilities and agent cluster collaboration, supporting up to 300 sub-agents working in parallel. The company also initiated testing for new features in the Claw group.
Strategic Focus: From Diversified Attempts to Betting on Programming and Agents
The growth behind this round is a result of Kimi’s strategic contraction and refocus after a low point in early 2025. In January, Kimi released the K1.5 inference model, which was positioned against OpenAI’s o1, but market attention was largely captured by the simultaneously released and open-sourced DeepSeek-R1.
After this impact, Kimi established three core adjustments: prioritizing continuous achievement of SOTA (state-of-the-art), significantly reducing consumer marketing expenditures, and shifting from closed-source to open-source. The company is concentrating resources on programming capabilities and agent applications, aligning with founder Yang Zhilin’s long-standing advocacy for a “productivity scene first” approach.
On the technical front, Kimi has gained a certain level of influence in the open-source community. The Muon optimizer improvement, MuonClip, proposed on the K2 model, has been widely adopted in the industry. The Attention Residuals technology introduced in March received praise from Elon Musk on social media, who described it as “Impressive work from Kimi.”
Talent and Competition: Increased Incentives Amid Rising Poaching Pressure
The expansion of financing also serves to meet the needs of talent competition. According to LatePost, Kimi’s founder Yang Zhilin stated in an all-staff letter at the beginning of the year that the average incentive for 2026 will double compared to 2025, with plans to significantly increase the stock option buyback amount; the company is also launching a stock option incentive plan for interns. Given that the company’s valuation has quadrupled in a few months and it has yet to go public, the attractiveness of stock options has significantly increased.
However, due to Kimi’s technical performance in programming and agent fields, its personnel have also become targets for competitors, creating significant talent pressure.
From a business model perspective, domestic AI model startups generally rely on two monetization paths: charging based on API token usage and developing applications based on proprietary models to generate subscription income. Key variables in validating the sustainability of the business model include controlling inference costs, creating premium product experiences, and securing sufficient computing power in a timely manner. Ample funding reserves are a prerequisite for all of the above.
Comments
Discussion is powered by Giscus (GitHub Discussions). Add
repo,repoID,category, andcategoryIDunder[params.comments.giscus]inhugo.tomlusing the values from the Giscus setup tool.