A significant development in China's AI investment landscape emerged on September 2, 2026, as Moonshot AI (Kimi) reportedly submitted its A1 filing to the Hong Kong Stock Exchange in a confidential manner, officially commencing its IPO process. Simultaneously, the company is advancing its final pre-IPO funding round at a pre-money valuation of $50 billion, while Kimi's official response to the matter was a terse "no comment on market rumors."

From a valuation of approximately $4.3 billion at the end of 2025 to the current pre-money valuation of $50 billion, Moonshot AI has experienced a more than tenfold increase in less than one year. This AI large-model company, founded just over three years ago, is knocking on the doors of the capital markets at an astonishing pace. However, the critical question remains: behind these skyrocketing valuation figures, do they truly reflect solid fundamentals or merely represent a capital market frenzy?

The Valuation Surge: A Capital Narrative from $4.3 Billion to $50 Billion

Moonshot AI was established in April 2023, founded by Yang Zhilin, who holds a bachelor's degree from Tsinghua University and a PhD from Carnegie Mellon University, with prior experience at Google Brain and Meta AI. Since inception, the company has attracted substantial capital interest: it secured over $1 billion in funding in February 2024 at a $2.5 billion valuation, followed by an additional $300 million in August of the same year, raising its valuation to $3.3 billion.

The true valuation leap occurred in 2026. Between January and February, the company completed three consecutive rounds of "small-step, fast-paced" financing, elevating its valuation from $10 billion to $18 billion. By May, a $2 billion funding round led by Meituan Longzhu pushed the valuation past $20 billion. The E-round in June brought the pre-money valuation to $31.5 billion, and by July, the F-round raised over $3.5 billion, setting a post-money valuation of $35 billion. Now, the pre-IPO round is commanding a pre-money valuation of $50 billion.

In just three months, Moonshot AI's valuation catapulted from $18 billion to $50 billion, a trajectory so steep it is exceptionally rare in China's venture capital history. What core logic underpins this valuation phenomenon? The primary driver is the release of the Kimi K3 model. On July 17, 2026, Moonshot AI unveiled its next-generation flagship large model, Kimi K3, boasting a total parameter count of 2.8 trillion, making it the largest open-source model globally.

K3 employs a Mixture-of-Experts (MoE) architecture, with 896 routed experts where each token activates only 16, resulting in 104 billion activated parameters. It supports a 1 million token context window and native visual understanding capabilities. The model achieved a score of 1679 to top the Frontend Code Arena programming leaderboard, marking the first time an open-source model has surpassed leading closed-source models for the top spot. Within just 30 minutes of its release, K3 reached the top of Hugging Face's trending list, setting a record for the fastest growth since the platform's inception.

The commercial value of K3 has quickly translated into financial results. Moonshot AI reported revenue of RMB 1.699 billion for 2025 with a net profit of RMB 278 million. For the first half of 2026, projected revenue is approximately RMB 1.052 billion to RMB 1.128 billion, representing year-on-year growth of 35.62% to 45.41%. Annual Recurring Revenue (ARR) surged from $100 million in early March 2026 to over $200 million by May, and stabilized at $300 million by mid-June, tripling within three months. API revenue now accounts for over 70% of total revenue and continues to climb.

The capital market's valuation logic for Moonshot AI has surpassed traditional PS or PE frameworks. One investment firm partner candidly stated, "It's very difficult to judge the rationality of valuations for AI large-model companies using traditional PS or PE multiples. If you look at revenue and profit, the valuation certainly isn't supported. But if you evaluate AI projects using traditional industry metrics, you probably wouldn't invest in any of them." From his perspective, while the $35 billion valuation seemed high, "if you view it from the angle of betting on whether Kimi can grow into one of the few domestic AI companies capable of sustainable operations while maintaining international competitiveness as a platform company, this valuation reflects the capital market's confidence in Kimi."

Two core variables support this valuation: first, the ability to sustain a global top-tier technological position, and second, the capacity to convert technological advantages into sustained high revenue growth. Based on K3's performance, the market has clearly given a provisional affirmative answer to both questions. Additionally, Chapter 18C listing rules introduced by the Hong Kong Stock Exchange in March 2023 have significantly lowered listing barriers for specialty technology companies including AI firms. Zhipu AI and MiniMax previously successfully listed on the Hong Kong exchange in January 2026 with strong stock performance, having surged approximately fourfold and sevenfold respectively since their debuts. This demonstration effect has undoubtedly paved the way for Moonshot AI's IPO journey.

Concerns Beneath the Glory: User Attrition, Profitability Pressures, and Valuation Bubble Risks

However, the $50 billion valuation is not without controversy. Behind the glossy capital narrative, Moonshot AI faces multiple real-world challenges. The most prominent issue is a notable decline in user activity. According to QuestMobile data, Kimi App's Monthly Active Users (MAU) fell from 21.653 million in Q1 2025 to 9.027 million by year-end, a decline of 58.3% from its peak. Industry ranking slipped from TOP2 to fifth place. During the same period, DeepSeek's daily active users across all platforms exceeded 110 million, while Kimi's were only 6.88 million, less than 6% of DeepSeek's figure.

Despite achieving breakthroughs in commercialization, with paid subscriber numbers growing 170% month-over-month, the shrinking user base remains a signal that cannot be ignored. If user attrition cannot be effectively curbed, the ceiling on commercialization will be readily apparent. Profitability pressures are equally significant. Large-model companies face a structural dilemma: the marginal cost of new revenue is high, and more powerful models generate more usage but also rapidly consume more computing power. Following the K3 release, Moonshot AI temporarily suspended new C-end user subscriptions due to user demand exceeding expectations, prioritizing computing resources for existing paid users.

This demonstrates that even with rapidly growing ARR, computing capacity remains a bottleneck restricting commercial expansion. To support the next round of model iteration and commercial growth, the company must continuously raise funds and expand computing capacity, creating a veritable bottomless pit for capital. The competitive landscape is also undergoing dramatic shifts. Zhipu and MiniMax have already listed, and DeepSeek is widely expected to go public in the first half of 2027. Leading large-model companies are engaged in intense arms races. Looking at the latest model iterations, Zhipu focuses on Coding, Agent, and long-horizon tasks, MiniMax concentrates on multimodal global expansion, and DeepSeek continues to strengthen its Coding and complex task capabilities.

While Moonshot AI excels in long-text processing and knowledge work, it faces formidable competitors across every specialized segment. Notably, K3's pricing at $2.3 per million tokens sets a new high for domestic models, exceeding Zhipu GLM5.2's $0.9 and DeepSeek V4's pricing. Whether this premium pricing strategy can sustain market acceptance remains to be tested. In terms of valuation itself, bubble risks cannot be ignored. At the end of 2025, Moonshot AI was valued at just $4.3 billion; now its pre-IPO valuation stands at $50 billion, an approximate twelvefold increase in over six months. Even accounting for K3's technological breakthroughs and accelerated commercialization, this valuation growth rate far exceeds revenue growth, with first-half 2026 projected revenue of only RMB 1.052 billion to RMB 1.128 billion.

At the $50 billion valuation, the price-to-sales ratio reaches hundreds of times. While supporters frame this as a bet on the future of a "platform-type AI company," history has repeatedly shown that excessive expectations often accompany significant correction risks. Another aspect worth noting is the confidential filing itself. The A1 document is one of the formal application forms companies submit when seeking listing on the Hong Kong Stock Exchange. Filing in a confidential manner means the company chooses not to disclose financial details during the preparation phase. While not uncommon in Hong Kong listings, this approach reflects the company's prudent consideration of market sentiment, as premature disclosure could trigger unnecessary volatility, particularly at a sensitive time when valuations are already elevated.

Moonshot AI is writing one of the most aggressive capital stories in China's AI entrepreneurship history with its $50 billion valuation. From technological breakthroughs to commercial implementation, from financing pace to IPO progress, this company has pressed the accelerator across virtually every dimension. Yet capital markets will ultimately return to rationality. When the prospectus is made public and financial data faces public scrutiny, the market will deliver its final verdict on whether the $50 billion valuation represents value discovery or bubble inflation. For investors, Kimi's IPO represents both a bet on AI's future and a renewed examination of valuation fundamentals.