DeepSeek Eyes $71B Valuation in Second Funding Round as IPO Preparations Begin

DeepSeek, the Hangzhou-based AI lab that stunned Silicon Valley at the start of 2026 with models that outperformed Western frontrunners at a fraction of their training cost, is moving with remarkable speed toward becoming China’s next major technology listing. According to people familiar with the matter, the company is in early discussions with prospective investors for a follow-on funding round at a pre-money valuation of at least 480 billion yuan — approximately $71 billion — and has begun preliminary conversations with investment banks and accounting firms about a public offering on a Chinese stock exchange as early as late 2026.

A Valuation That Has Multiplied Seven Times in Under a Year

The trajectory of DeepSeek’s valuation over 2026 reads less like a corporate fundraising history and more like a compressed chart of the AI investment supercycle itself. When reports first surfaced in April of outside investors taking a stake, conversations pegged the company’s value at roughly $10 billion. Within weeks, that figure had climbed above $20 billion. By early May, China’s National Artificial Intelligence Industry Investment Fund — a state-backed vehicle that has become the defining actor in China’s domestic AI funding market — entered talks to lead a round at approximately $45 billion.

In early June, DeepSeek completed what it described as its first-ever external funding round: approximately $7.4 billion raised at a post-money valuation of $52 billion. The round drew participation from Tencent, CATL — the world’s largest battery manufacturer, now a significant AI investor — and several other strategic backers. It was the largest AI funding round in Chinese history, and the company had never taken outside capital before.

Now, just weeks later, a second round is taking shape at a pre-money figure that implies DeepSeek’s valuation has grown 37% since the ink dried on the first. For context: $71 billion would make DeepSeek the third most valuable private AI company globally, behind only OpenAI and Anthropic.

The Founder Who Became China’s Wealthiest AI Billionaire

DeepSeek’s founder, Liang Wenfeng, holds roughly 78% of the company’s equity through a specialized transaction structure. His net worth, according to Bloomberg’s estimates, has climbed to approximately $36 billion — placing him among the wealthiest founders in global AI, and the wealthiest among AI-focused founders in China.

Liang built DeepSeek from within High-Flyer, his quantitative hedge fund. For years, the company operated entirely without outside capital, funding its AI research through the profits of High-Flyer’s trading operations. That self-sufficiency gave DeepSeek unusual freedom from investor pressure — a freedom that apparently enabled the patient, research-intensive approach that produced DeepSeek-R1, the reasoning model whose January 2026 release triggered a sharp selloff in Nvidia’s stock and a reassessment of AI infrastructure economics across Silicon Valley.

What DeepSeek Plans to Do With the Money

The funding rounds and prospective IPO come amid an increasingly urgent need to secure infrastructure. DeepSeek’s efficiency advantages — its models achieve competitive performance with significantly fewer computational resources than rivals — have not made it immune to the fundamental economics of frontier AI research, which requires continuous scaling of training infrastructure.

The company is also reportedly developing its own AI inference chip, a move that would reduce its dependence on both Nvidia’s restricted export products and Huawei’s Ascend chips. Access to advanced chips remains one of the most significant constraints on Chinese AI development following U.S. export controls; a proprietary chip program would give DeepSeek more control over its cost structure and scaling trajectory.

IPO Timeline and Listing Venue

According to sources cited by Bloomberg, DeepSeek could file listing documents as soon as late 2026, with a market debut targeted for 2027. The company has begun discussions with accounting firms and investment banks — steps that mark a clear transition from exploratory conversations to active preparation.

The likely venue is one of the Chinese mainland exchanges, either the Shanghai STAR Market or Shenzhen’s ChiNext — both of which have hosted high-profile technology listings and have mechanisms for companies with non-standard governance structures. A Hong Kong listing is also possible; the city’s exchange has actively courted Chinese technology companies following regulatory changes that streamlined the approval process for dual-class share structures.

The decision on venue carries strategic dimensions beyond pure finance. A mainland listing would provide DeepSeek with access to China’s deep retail investor market and align it more closely with state priorities. A Hong Kong listing would offer greater international visibility and potentially access to global institutional capital.

The Broader Stakes

DeepSeek’s emergence has reshaped several assumptions that had governed AI investment logic since the GPT-4 era. Chief among them: the belief that frontier AI performance required the kind of capital and compute scale that only a handful of hyperscalers and well-funded Western startups could assemble.

DeepSeek’s V3 and R1 models demonstrated that a Chinese company operating under significant chip export restrictions could produce models competitive with GPT-4o-class systems. The V4 model released in April 2026 extended that performance advantage further, and DeepSeek’s API has seen rapid adoption globally — including, paradoxically, significant usage from developers in the United States and Europe who access it through API aggregators.

A $71 billion second-round valuation and a public market debut would convert DeepSeek from a research-famous private company into a publicly accountable institution — one with obligations to shareholders, disclosure requirements, and a market capitalization that will sit alongside those of Baidu, Alibaba, and ByteDance as a benchmark for Chinese AI value creation.

For investors who missed Nvidia’s AI infrastructure run and OpenAI’s private markets, DeepSeek’s IPO would represent one of the few remaining ways to take a position on the Chinese AI stack at a stage where the technology’s value has already been validated but the full commercial arc remains ahead.

Whether that opportunity materializes on the timeline suggested — late 2026 filing, 2027 debut — depends heavily on regulatory approvals, market conditions, and DeepSeek’s own assessment of when its story is most legibly compelling to public market investors. Given the speed at which everything else in DeepSeek’s 2026 has moved, betting against the accelerated timeline seems unwise.