China to Invest in DeepSeek at $50 Billion Valuation

China’s DeepSeek is raising money from government-backed investors, aligning the artificial-intelligence startup with Beijing’s push for technology self-sufficiency.

Some prospective investors have valued DeepSeek at around $50 billion in recent talks, people familiar with the matter said. The valuation has surged in recent weeks after previous discussions envisioned a range between $10 billion and $30 billion.

Most Read from The Wall Street Journal

China’s National Artificial Intelligence Industry Investment Fund, a one-year-old government-backed fund with around $8.8 billion in capital, is in advanced talks to invest in the round in Chinese yuan, people familiar with the matter said.

Hangzhou-based DeepSeek aims to raise a few billion dollars in the new round, the people said. The company plans to use the new capital to advance its research and development as well as to expand computing infrastructure, they said. It also wants to anchor a market valuation to signal its worth and provide a benchmark for employees’ stock grants to help retain top talent, some of the people said.

Beijing has treated DeepSeek as a national AI champion since the startup rattled Silicon Valley and Wall Street early last year with a powerful model it said was built at a fraction of the cost of Western rivals. DeepSeek has since become a key component of China’s plan to have top-class homegrown companies in a range of AI fields.

Beijing views this plan as a way to hedge against U.S. export controls and take leadership in bringing AI to the world.

Last month, DeepSeek released its upgraded flagship model, called V4. While the new model was trained on Nvidia’s high-end chips, DeepSeek worked closely with Huawei and other domestic chip providers, signaling a pivot away from dependence on the U.S.

DeepSeek said V4 matched top-tier U.S. products released late last, year but in certain areas lagged leading U.S. models released this year such as Anthropic’s Claude Opus 4.6.

Investors expect the new DeepSeek model to drive a wave of AI adoption in China as industries look to automate more complex tasks from offices to the factory floor.

DeepSeek’s main models including V4 are open-source, meaning they are free for users to download and modify. The company earns a small amount of revenue from selling access to its models that run on its computing infrastructure.

While DeepSeek’s ability to rapidly increase its revenue is uncertain, some investors believe its connections to China’s industry will eventually help it build a lucrative business. Other China-based AI companies already sport high valuations. Hong Kong-listed Zhipu AI currently has a market capitalization of around $50 billion, while MiniMax stands at roughly $30 billion.

In 2024, China raised the equivalent of around $50 billion in the third installment of its national semiconductor fund, commonly known as the Big Fund, aiming to bolster the country’s chip-making capabilities. Months later, the Big Fund helped set up the AI fund.

In the initial months after it made a global splash, DeepSeek rebuffed external capital, including from government-linked funds, seeking to shield its decision-making from outside interference. It instead relied on the personal wealth of its founder, Liang Wenfeng, and profit from his hedge fund.

Now the company is aligning itself with Beijing’s call to bolster the China’s resilience against U.S. pressure.

China is tightening oversight of AI and other strategic technologies. It has recently told some companies developing AI not to accept U.S. investments without government approval. These curbs followed U.S. restrictions that limit American investments in China’s advanced AI companies and block exports of chip technology to the country.

Last month Beijing banned the acquisition by Meta Platforms of AI startup Manus, a Singapore-based company with Chinese roots.

Write to Raffaele Huang at raffaele.huang@wsj.com and Tracy Qu at tracy.qu@wsj.com

Most Read from The Wall Street Journal