Moonshot AI, a Chinese startup, is experiencing rapid valuation growth, with its pre-money valuation reaching $31.5 billion in June after a $2 billion financing round in May at a post-money valuation of $20 billion. This represents a more than 7-fold increase in valuation within half a year. The company's "Kimi K3" model, a 2.8 trillion parameter multimodal model with native visual capabilities, has garnered significant attention and is considered by experts to be a rival to advanced US offerings. Analysts expect K3 to open new pricing opportunities and project Moonshot AI's Annual Recurring Revenue (ARR) to exceed $1 billion by the end of the year.
The valuation logic for Moonshot AI is primarily based on replicating Anthropic's early revenue curve, driven by surging developer calls, growing API usage, increasing overseas paid usage, and model iterations that push pricing upward. For instance, if Moonshot AI achieves $1 billion ARR by the end of the year, its forward multiple against the $31.5 billion valuation would be approximately 30x, which is considered cheaper than Zhipu AI's current valuation multiple. This approach suggests investors are betting on a re-rating similar to Zhipu AI's path, which saw its market capitalization exceed $100 billion after listing from a primary round valuation of $13.7 billion.
In contrast, DeepSeek, another prominent Chinese AI company, has different drivers for its high valuation. DeepSeek's recent annualized revenue is estimated at $400 million to $500 million, primarily from API services. Its $74 billion valuation in a second financing round, after a $54.3 billion pre-money valuation in its first round, results in a Price-to-ARR ratio of around 148x. This steep valuation is supported by a "non-financial premium" including call volume, technical influence, geopolitical strategic scarcity, and founder endorsement. DeepSeek's first round of financing, at approximately $7.4 billion, was the largest single round in China's AI large model industry. Despite outperforming Anthropic and GPT models in global token volumes on platforms like OpenRouter, DeepSeek faces challenges such as a 95% deflation in inference pricing in the mid-tier market, raising questions about its revenue growth sustainability from call volume.
The broader context for both companies includes the rapid growth of the Chinese AI market, with Chinese open-source models contributing significantly to the doubling of OpenRouter's total call volume in Q2 2026. However, Jefferies warned on July 13 that the capability gap between Chinese and US models could widen again as US vendors acquire more new-generation computing power. This poses a risk to the valuation logic of current Chinese models, which implies a continuous narrowing of this capability gap. Both DeepSeek and Moonshot AI's aggressive primary market valuations will face scrutiny as they move toward potential IPOs, where the public market's continuous pricing mechanism will pressure their valuation models. While Moonshot AI's valuation is lower than Anthropic and OpenAI based on performance rankings, its aggressive financing pace suggests a focus on rapidly growing its market presence and revenue base.
