HANGZHOU, China: Chinese AI lab DeepSeek is preparing to raise about 50 billion yuan, or roughly $7.4 billion, in its first-ever outside funding round, at a valuation of between $52 billion and $59 billion, the South China Morning Post reported on June 3. Tencent and battery maker CATL are lined up as the largest external investors, and founder Liang Wenfeng is putting in about 20 billion yuan of his own money, according to financial outlet 36Kr. The round is expected to close within the next few weeks.
The valuation is a sharp jump from the roughly $20 billion figure that investors were circling in mid-April, when DeepSeek’s first round was still forming. Grey Journal covered that earlier stage in an April news piece on the $20 billion DeepSeek raise, noting that Tencent had at one point proposed acquiring a 20% stake before the company pushed back on ceding that much control. In six weeks, the implied price has nearly tripled.
How DeepSeek Moved From $20 Billion to $59 Billion in Six Weeks
DeepSeek built its reputation on capital efficiency. The Hangzhou-based lab, founded by quant hedge-fund manager Liang Wenfeng, became globally visible in early 2026 when its V3 and R1 open-weight models matched frontier US systems while reportedly costing a fraction as much to train. Until this round, the company operated almost entirely on Liang’s personal capital and revenue from its sister hedge fund, High-Flyer.
The current round’s structure reflects that history. According to 36Kr’s breakdown of the financing, Liang is putting in about 20 billion yuan personally, roughly 40% of the total, which means he and his existing entities still control the company after the raise. Tencent is weighing about 10 billion yuan, CATL about 5 billion yuan, and DeepSeek is in late-stage talks with China’s national AI fund, gaming developer NetEase, e-commerce platform JD.com, and Hong Kong’s IDG Capital, among others. The plan calls for fewer than 10 outside investors.
The valuation jump tracks DeepSeek’s product cadence and the strategic logic of the new backers. DeepSeek V4, the company’s trillion-parameter flagship, launched on April 24, the same day the funding talks first became public. CATL, which dominates global EV batteries, has been pushing into power solutions for AI data centers, where energy access is now as scarce as compute. Tencent runs one of China’s largest cloud businesses and a games platform that increasingly relies on AI features. Both are strategic backers buying into the supply chain, not financial investors hunting a markup.
What Does DeepSeek’s $59 Billion Valuation Mean for Founders?
It signals that capital efficiency alone is no longer enough to stay in the AI frontier race. Even the world’s most cost-efficient lab has concluded it needs a multibillion-dollar war chest to keep pace, and DeepSeek’s founder is willing to dilute himself by 60% to get one. The bar for “lean” has reset.
For startups, the most useful piece of the deal is what Liang did with his own check. By writing the single largest line on the cap table himself, he kept founder control through a raise that would otherwise hand strategic backers near-veto power. That option is available to almost no one outside founders with prior fund or trading wealth, but the principle generalizes: when outside money is unavoidable, structure it so the strategic logic, not the share count, drives the decisions.
The deal also clarifies where the market now prices a credible AI lab. DeepSeek at $59 billion places it in the same tier as Anthropic and OpenAI on a private-market basis, even though its revenue is a fraction of theirs. CNBC’s reporting on the maiden fundraising attributes the raise to soaring R&D costs, intensifying talent competition, and the need to price employee equity, not a cash crunch. Founders pitching AI-adjacent companies are now negotiating against a market that has accepted those costs as table stakes.
What to Watch as the Round Closes
The round is expected to close within the next couple of weeks, though terms could still shift. Three items are worth tracking. First, whether Tencent’s commitment lands at 10 billion yuan or higher, because anything that pushes its stake past 5% reopens the control questions that surfaced in April. Second, whether DeepSeek discloses the size of its employee equity pool, which is one of the stated reasons for the raise and a tell on how aggressively it plans to hire. Third, how US export-control regulators read the deal, given that CATL is on multiple Western watchlists and that DeepSeek’s commercial AI products are increasingly used by Western developers through API resellers.
The competitive read matters too. A closed DeepSeek round at this size puts price pressure on every other major AI raise in the second half of 2026, and it gives Chinese cloud and infrastructure firms a domestic frontier-model partner that does not depend on US chips or US capital. For founders building on top of frontier models, it is one more reason to design for a multi-provider future rather than betting the product on any single lab.
DeepSeek has not commented publicly on the terms. The company said in its V4 launch materials that it plans to release new models on a quarterly cadence through the end of the year.
