Reports emerge that DeepSeek, the Hangzhou artificial intelligence lab, has closed a second external funding round of roughly 100 billion yuan, nearly $16 billion, more than doubling its original target. The reported deal, described as one of the largest ever in China’s AI sector, would lift the company’s 2026 fundraising above $22 billion, in a deal that would reshape its place in China’s AI race.
What was reported, and what was not confirmed
The round was reported by Bloomberg on October 22, which said DeepSeek had raised about 100 billion yuan, or $15 billion, citing people familiar with the matter. Reuters followed the same day, citing four people with knowledge of the matter, and put the figure at 100 billion yuan, or $14.98 billion. DeepSeek did not comment to either outlet.
A later account, published October 30, put the figure higher, at 105.5 billion yuan, or roughly $15.9 billion. The difference between the October 22 and October 30 figures reflects different stages of the fundraising, different conversions and different sourcing, rather than two separate rounds. Readers should treat $15 billion to $16 billion as a range, not a confirmed total.
The reported pre-money valuation sits at 500 billion yuan, or roughly $74 billion, before the new capital. That figure comes from earlier Reuters reporting, and from the later account of the closed round. It is broadly flat against the valuation implied by the first external round in June, when it raised $7.4 billion at a valuation near $50 billion.
DeepSeek has not publicly confirmed the round, its size or its valuation, and no company statement was cited in the reports. That silence is part of the story: the figures rest on unnamed sources, not on a filing or an announcement. Any use of the numbers should carry that attribution, and treat the investor list as reported rather than confirmed.
One caution from the October 22 reporting still applies. A Reuters source said then that fundraising was set to complete in days, but details and the investor lineup were not ironclad and could still change. The October 30 account describes the round as closed, which would resolve that uncertainty, but without company confirmation the terms remain reported, not verified.
Separately, Tencent and Alibaba are reported to be in talks for a separate stake in DeepSeek, distinct from the funding round itself. That report, if it proceeds, would deepen ties between DeepSeek and China’s largest internet companies. It remains unconfirmed, and no terms were reported, so it should not be folded into the round’s total.
Who is reported to be investing
Battery maker Contemporary Amperex Technology, known as CATL, and WeChat operator Tencent Holdings, both existing DeepSeek shareholders, were among the biggest investors in the round, according to Reuters and Bloomberg. CATL alone is reported to have roughly doubled its earlier contribution, adding another 10 billion yuan on top of a previous 5 billion yuan ticket.
More than 15 investors took part, Reuters reported, with new names including home appliance maker Midea, food delivery company Meituan, automaker Geely and chipmaker GigaDevice. Other reported participants include NetEase, Boyu Capital, CPE, Legend Capital and existing shareholder IDG Capital. The breadth of the list is striking: it spans batteries, cars, phones, food delivery and home appliances.
State-backed money is reported to be a part of the round. The later account’s headline puts state capital at 27% of the total, led by China’s big chip fund, while Reuters noted that many investors were backed by the state and big corporations. That mix matters because it ties DeepSeek’s future to China’s industrial policy, not only to commercial returns.
The investor list also reads as a supply-chain map. CATL makes batteries, GigaDevice makes chips, Geely makes cars and Meituan runs delivery networks: all are companies that could use, distribute or build on DeepSeek’s models. As one analyst quoted by Reuters put it, DeepSeek could become a key part of these domestic supply chains, rather than a standalone software company.
DeepSeek is reported to have vetted prospective investors closely, turning away private investment funds raised from individual investors. That selectivity, if accurate, suggests a company choosing strategic partners over the largest possible cheque. It also means the cap table is a political and industrial statement, not just a financial one.
Why a lab that prized efficiency needs $16 billion
DeepSeek built its reputation on doing more with less, training capable models at a fraction of the cost claimed by Western rivals. A $16 billion round therefore looks like a contradiction. In practice, it reflects a different stage of competition: the cost of serving models at scale, building capacity and hiring talent has risen sharply, even for an efficient lab.
The timing also matters. The round closed four months after DeepSeek’s first external funding in June, and weeks after it released DeepSeek-V4.1-Flash, a model designed for greater capability, faster inference and higher throughput. Serving a fast, cheap model to millions of users still requires enormous infrastructure, and infrastructure is what this kind of money buys.
China’s AI race adds a second pressure. DeepSeek is no longer competing only with other Chinese labs: it is competing with American frontier labs that raised tens of billions each, and with a Chinese state that wants domestic champions. A round of this size, with state-backed funds inside it, signals that Beijing and corporate China will fund that competition directly.
There is also a defensive logic. Reports in October noted growing concern about massive AI spending, a sell-off in global technology stocks, and delays or withdrawals of data centre listings. Raising a very large round while investor appetite remains strong, even as sentiment cools elsewhere, locks in capital before conditions tighten further.
For Western rivals, the round narrows one advantage. OpenAI’s funding target was reportedly cut from $30 billion to $20 billion in the same week, as investors grew cautious about AI infrastructure spending. DeepSeek, by contrast, is reported to have more than doubled its target, suggesting that capital for Chinese AI remains available even as Western funding rounds face tougher questions.
What this changes, and what it does not
The most important change is structural, not technical. DeepSeek began 2026 having just started taking outside capital. It ends October with over $22 billion raised in a year, a cap table spanning China’s largest companies and state-backed funds, and a reported path toward a public listing. That is a transformation from research lab to national champion within a year.
The money does not, by itself, make DeepSeek’s models better. Model quality depends on data, talent, computing access and research choices, and the reports say nothing about how the new capital will be allocated between those uses. Investors are betting on the company’s future work, not certifying its current models, and the distinction should be kept in view.
The state-capital share also creates a tension that will follow DeepSeek abroad. A lab backed by Chinese state funds and major domestic companies may find it harder to win trust from foreign customers and regulators, particularly in the United States, where DeepSeek products have already faced restrictions in federal civilian IT. Capital that helps at home can complicate expansion overseas.
For Chinese peers, the round sets a new bar. DeepSeek and MiniMax won Beijing approval for global model deployment earlier, showing how Chinese labs are moving from domestic success to international ambition. A $16 billion round for the most prominent of those labs raises the funding expectations for every other Chinese lab seeking to compete at the same level.
For investors, the round tests a different thesis: that an open-weights lab, embedded in China’s supply chains and backed by its state, can build a durable business without the subscription revenues of its Western rivals. That thesis is unproved, and the reports give no revenue or profit figures to prove it. The bet is on position, not on demonstrated returns.
Why It Matters
The round matters because of its scale and its composition together. A $16 billion financing would be large anywhere; a $16 billion financing in which roughly a quarter of the money is reported to be state-backed, and the rest comes from China’s biggest industrial companies, is a statement about how China intends to compete in AI.
It matters for the global funding race too. While Western labs face harder questions about spending, valuations and returns, DeepSeek is reported to have doubled its target and closed one of China’s largest AI financings. That divergence may not last, but it changes the negotiating position of every lab raising money in the months ahead.
It matters for users and developers as well, though less directly. A well-funded DeepSeek is more likely to keep its models cheap and widely available, which has been its main competitive weapon. DeepSeek’s model tried to dodge shutdown in safety tests, a reminder that cheap and capable models also carry safety questions that funding alone does not answer.
Finally, it matters because so much remains unconfirmed. The figures, the investor list and the valuation rest on reporting, not on a company announcement. The honest conclusion is that DeepSeek has reportedly closed a very large round, with major Chinese and state-backed investors, and that the full terms will only be clear when the company or its investors confirm them.
Background
DeepSeek, based in Hangzhou, rose to prominence with open-weights models that challenged Western rivals on capability and cost. It took its first external funding in June 2026, raising $7.4 billion, and released DeepSeek-V4.1-Flash in September 2026. Reports in September said it aimed to begin a public listing process this year, with CITIC Securities preparing a potential STAR Market float.
Source: Reuters report and The Information account
0 Comments