Watch Out OpenAI, Anthropic: DeepSeek, Moonshot Raise Fresh Capital As Chinese Open-Weight AI Models Gain US Users

China’s AI upstarts are raising bigger pools of capital and gaining traction with lower-cost open-weight models.

• Moonshot AI has reportedly reached a $50 billion valuation and is preparing for a potential Hong Kong IPO.
• DeepSeek is reportedly nearing a $12 billion funding round.
• Tencent and CATL are among the backers of DeepSeek's round, according to the report.

China’s AI challengers are rapidly building up their war chests and expanding their global reach, adding pressure on U.S. AI leaders and IPO hopefuls OpenAI and Anthropic.

Moonshot AI has closed a large funding round at a valuation of roughly $50 billion, while DeepSeek is nearing a $12 billion capital raise, Bloomberg reported Tuesday.

Moonshot, DeepSeek Step On The Gas

Moonshot AI has closed its latest private fundraising round and is preparing for a Hong Kong initial public offering in the first quarter of 2027, according to the report. The company has started laying the groundwork for meetings with potential investors ahead of the listing, which could raise as much as $5 billion.

Moonshot AI’s reported $50 billion valuation marks a sharp jump from the $31.5 billion valuation it secured earlier this year. Its Kimi AI platform has become one of China’s most prominent AI products, while its latest Kimi K3 open-weight model has challenged leading U.S. systems on coding and agentic tasks at significantly lower costs.

DeepSeek is close to securing at least 80 billion yuan (about $12 billion) in its latest funding round, Bloomberg reported separately. The company initially targeted roughly 50 billion yuan, but strong investor demand has pushed the round higher. Tencent and battery giant CATL are among the major backers, the report said.

DeepSeek’s latest V4 Flash model has further strengthened its reputation for delivering frontier-level performance at a fraction of the cost of U.S. rivals.

The company first became a global phenomenon in early 2025 after its R1 reasoning model demonstrated that highly capable AI could be developed with far lower computing costs than Silicon Valley had assumed.

OpenAI is also eyeing fresh funding of as much as $30 billion. 

Cheap Chinese AI Gains U.S. Ground

Chinese AI labs are increasingly embracing open-weight models, allowing developers and companies to download, customize and deploy models instead of paying for access to closed systems.

That strategy is gaining traction in the U.S. as companies look to rein in soaring AI costs. Other Chinese models have become some of the most popular offerings on AI model-routing platforms such as OpenRouter, while DeepSeek has sharply increased its share of token usage on Vercel, according to a June report from Rest of World.

Airbnb has used Alibaba’s Qwen for customer-service applications, while Coinbase CEO Brian Armstrong said the company has nearly halved its AI spending by steering employees toward lower-cost models, including Moonshot AI’s Kimi and Z.ai’s GLM, according to The Information.

America’s AI Complex Joins Open-Weight Push

U.S. technology companies also offer open-weight models. Meta has its Llama family, Google has released its Gemma models and Microsoft offers Phi. However, OpenAI, Anthropic and U.S. officials have accused DeepSeek and Moonshot of using outputs from American models to speed up their development.

As of Tuesday, retail sentiment was ‘neutral’ for both OpenAI and Anthropic. Their latest private market valuations are $953.8 billion and $1.32 trillion, respectively, according to Nasdaq Private Market data.

For updates and corrections, email newsroom[at]stocktwits[dot]com.<