DeepSeek nears $7.4 billion funding round at $74 billion valuation ahead of 2027 IPO

DeepSeek is nearing a massive new funding round that could bring in about 50 billion yuan, or $7.4 billion, valuing the Chinese AI startup at roughly $74 billion before the investment, according to reports from the Wall Street Journal and South China Morning Post (SCMP).

The financing is expected to close by the end of August, as the Hangzhou-based company prepares for a potential initial public offering on Shanghai’s Star Market. DeepSeek could file for an IPO as early as the end of 2026, with a public debut targeted for 2027, according to people familiar with the plans.

“DeepSeek is nearing the completion of a new funding round valuing the company at about 500 billion yuan (US$74 billion) before investment, as the Chinese artificial intelligence start-up moves closer to a potential listing on Shanghai’s Star Market,” SCMP reported, citing people familiar with the ongoing talks.

One source told SCMP the AI startup was “seeking to raise about 50 billion yuan in the round, which was expected to close before the end of August.”

The Wall Street Journal separately reported that DeepSeek was close to securing the financing, saying, “DeepSeek is set to soon secure new funding that would value the Chinese artificial-intelligence startup at $74 billion, building its war chest as it targets an initial public offering next year.”

The numbers put DeepSeek in rare territory. A full $7.4 billion raise at a $74 billion pre-money valuation would imply a post-money value of roughly $81 billion, cementing its position among the most valuable private AI companies outside the United States.

Existing backers Monolith, Shixiang Capital and Chinese battery giant Contemporary Amperex Technology Limited, better known as CATL, are participating in the round, according to reports. New investors in discussions include CPE, Legend Capital and semiconductor-focused private equity firm Stony Creek Capital. Funds backed by Chinese chipmaker GigaDevice and state investment vehicles from Hefei are reportedly involved as well.

DeepSeek and the investors have not publicly confirmed the financing.

DeepSeek’s cheap AI models are meeting an expensive reality

DeepSeek rose to global prominence after releasing AI models that challenged assumptions about how much money and computing infrastructure it took to compete with leading American AI labs.

Its rise was unusual for another reason. Founder Liang Wenfeng had largely financed the AI lab through High-Flyer Quant, the hedge fund he founded before DeepSeek. High-Flyer had spent years using AI and deep learning to trade stocks, giving Liang access to capital and computing resources without relying on the traditional Silicon Valley venture capital model.

That arrangement is becoming harder to sustain as DeepSeek grows.

The company is reportedly planning to add roughly 1 gigawatt of computing capacity as it builds larger models and competes for scarce AI talent. Those requirements can turn even a company known for efficient models into a capital-hungry infrastructure operation.

DeepSeek has opened itself to outside investors as those costs climb. Reports indicate that an earlier financing brought in investors including Monolith Management, Tencent, JD.com, NetEase and CATL, CNBC reported. The latest round would push DeepSeek further from its origins as an AI project financed primarily by Liang’s hedge fund, turning it into a heavily capitalized standalone AI company backed by outside investors, including Liang’s hedge fund, and increasingly positioned for the public markets.

“DeepSeek’s founding team, including Liang, are still traders at heart and inclined to chase maximum upside,” Ke Zong, a portfolio manager at a Shanghai-based hedge fund, said.

DeepSeek has raised API access prices amid a broader reset across China’s AI market. Alibaba, Tencent, Baidu and Zhipu AI have made similar moves this year as companies shift from aggressive discounts to finding sustainable economics for increasingly expensive AI services.

That pricing shift points to a larger contradiction running through the AI industry. Model inference can get cheaper, yet building the infrastructure, acquiring chips, training new systems, and keeping elite researchers can still consume enormous amounts of capital.

From High-Flyer experiment to potential public company

The fundraising comes as DeepSeek’s relationship with China’s broader technology sector is changing.

High-Flyer affiliates have secured pre-IPO allocations in some of China’s most sought-after technology listings, including memory chipmaker CXMT and humanoid robotics company Unitree Robotics. DeepSeek became a strategic investor in Unitree, taking a 2.31% stake in its offering and agreeing to a 36-month lock-up period.

Those investments connect Liang’s growing business interests to sectors Beijing considers strategically significant, including artificial intelligence, semiconductors and robotics.

“There’s a genuine distinction between the quant funds seeking returns and DeepSeek selectively using its corporate balance sheet to build strategic relationships around the future AI stack,” Hutong Research tech analyst Sigrid Wang said.

An IPO would mark the biggest transformation yet.

DeepSeek entered the global spotlight as an unconventional AI lab backed by a quantitative trading firm. Less than two years later, it is raising billions from outside investors, adding huge amounts of computing capacity and preparing for a possible public listing.

That progression says something about frontier AI economics. DeepSeek helped prove that competitive AI models could be built more efficiently than many in the industry expected. Its latest fundraising effort shows that competing at global scale can still carry an enormous price tag.