Announced 22 Oct 2026 · Sources checked
What has been reported
DeepSeek, the Hangzhou AI lab behind some of the most widely used open-weight models, has raised about 100 billion yuan, roughly $15 billion, in its latest funding round, according to Bloomberg, which first reported the deal on 22 October 2026 citing people familiar with the matter. Bloomberg described it as one of the largest funding rounds in China's history.
Reuters confirmed the size with two people of its own, in a report republished by DealStreetAsia. It said DeepSeek is "wrapping up" the round, which is set to be completed in the coming days, and that financial details "are not ironclad and could still change." DeepSeek, CATL and Tencent did not respond to requests for comment from either outlet, and the company has not announced the round.
So the headline number is a well-sourced report, not a closing announcement. Two major news agencies landing on the same figure from separate sources on the same day is about as firm as private Chinese fundraising news gets before a company files paperwork.
Who is investing
Battery maker Contemporary Amperex Technology (CATL) and WeChat operator Tencent were among the biggest investors, according to both Bloomberg and Reuters. Neither report gives exact amounts for individual backers.
Chinese financial media went further. A Gelonghui report carried by Sina Finance on 22 October, citing company-registry data from Tianyancha and other media, describes the round as a Series B. It lists new investors including Stony Creek Capital, SMIC's investment arm China Fortune-Tech Capital, CPE Yuanfeng, Legend Capital and a Geely automotive fund, with existing shareholders CATL, Tencent, IDG Capital, Monolith and Shixiang following on. It says CATL plans to put in about 10 billion yuan, double its roughly 5 billion yuan in the first round, and that new and existing shareholders are each contributing about half.
The mix is telling. A chipmaker's fund, a carmaker and a battery giant alongside the usual venture firms suggests DeepSeek is lining up partners for computing infrastructure and real-world deployments, not only cash. Earlier this month CNBC reported that DeepSeek had been vetting investors closely, turning away private funds raised from individuals and limiting the pool largely to government and corporate money.
How the round grew
The round got bigger as it went. DeepSeek started with a 50 billion yuan target. On 6 October, CNBC reported that the company was considering doubling it to as much as 100 billion yuan, with state-backed funds and corporate investors "jostling for stakes", while Bloomberg said DeepSeek was close to securing at least 80 billion yuan. The final figure, by both Bloomberg and Reuters accounts, landed at about 100 billion.
Valuation is less clear. CNBC's sources said DeepSeek was seeking about 500 billion yuan, or roughly $75 billion. Gelonghui reports a pre-money valuation of about 500 billion yuan, unchanged despite the larger raise. Bloomberg and Reuters did not give a valuation, so treat that number as reported, not confirmed.
The round also took longer than planned. CNBC said it was initially expected to close by the end of August but dragged on as some prospective investors wavered over the price, and that talks were briefly suspended after founder Liang Wenfeng reportedly told investors privately that Chinese AI companies still lagged far behind US rivals and relied heavily on Nvidia chips.
Why DeepSeek is raising now
Until this year DeepSeek did not take outside money; it was funded by Liang's quantitative hedge fund, High-Flyer. Its first external round closed in June at about 50 billion yuan, according to CNBC. Gelonghui's account of that round says Liang personally contributed about 20 billion yuan, Tencent about 10 billion and CATL about 5 billion, with China's national AI industry investment fund, NetEase and JD also taking part. If both rounds close as reported, DeepSeek will have raised around 150 billion yuan in about six months.
The reason is the cost of staying at the frontier. Bloomberg says DeepSeek has started investing more aggressively in AI data centres and in more open-source models for use in China and abroad. Its recent releases show the pace: the V4.1 Flash 1017 open-weights update last week, the DeepInfer inference engine it open-sourced on 20 October, and search and fetch APIs for agents. Each of those needs more computing power to train and serve.
Bloomberg also notes the gap that remains. Even at $15 billion, DeepSeek's war chest is far smaller than those of Anthropic and OpenAI, each of which has raised more than $100 billion, according to the agency.
The road to a Shanghai listing
Both reports frame the round as a step toward a stock-market listing. Reuters reported in September that DeepSeek had hired CITIC Securities to prepare for a possible initial public offering on Shanghai's STAR Market, the exchange's technology board. Bloomberg says a DeepSeek IPO would be one of China's most anticipated listings in years, and that its domestic rival Moonshot AI is preparing for one too.
Gelonghui, citing Hong Kong media, adds that DeepSeek could file its listing application as early as the end of this year with a 2027 debut in mind, and that the sponsor bank is reviewing investor backgrounds so that the shareholder count stays under China's 200-holder limit for private companies. It also reports that Yan Wentao, a former partner at Hillhouse's venture arm, joined DeepSeek as chief financial officer in late September. None of those details has been confirmed by the company.
The listing plans arrive in a busy season for AI IPOs. Anthropic's draft filing this week showed how much the leading US labs are spending, and investors will compare DeepSeek's finances with those numbers if it does file.
What it means for the AI race
DeepSeek matters beyond its own balance sheet because its cheap, openly licensed models set price expectations for the whole market. Bloomberg says its September V4.1 Flash release knocked the shares of local rivals MiniMax and Z.ai, and that a widely shared Artificial Analysis chart described a "DeepSeek kill zone" in which competitors struggle unless they beat its performance or its prices.
More money means DeepSeek can keep that pressure on: more training runs, more inference capacity and the ability to keep weights open without needing every model to pay for itself. It also deepens the political split around the company. The US Department of Homeland Security ordered federal civilian agencies to remove DeepSeek this month, a reminder that the company's reach outside China is limited by politics as well as money.
For developers outside China, the practical effect is likely to be more DeepSeek releases and continued low API prices, alongside rising compliance questions for organisations covered by government bans.
What is not confirmed
Several things remain unsettled. The amount is reported by sources who say terms could still change, and the round has not formally closed. The valuation comes from CNBC's sources and Chinese media, not from Bloomberg or Reuters. Investor names beyond CATL and Tencent come from registry-based Chinese reports, and the per-investor amounts have not been disclosed by the companies.
DeepSeek itself has said nothing publicly, which is consistent with its low-profile style but means there is no official statement on how the money will be spent. The IPO timetable is also reported, not announced, and depends on regulators.
What is clear is the direction. A lab that became famous for doing more with less now has one of the largest private war chests in China, and the investors include companies that build chips, batteries and cars. Watch for a formal close, any filing with the STAR Market, and whether the new capital shows up first as bigger models or as more data-centre capacity.
Common questions
How much did DeepSeek raise?
About 100 billion yuan, roughly $15 billion, according to Bloomberg and Reuters on 22 October 2026. Reuters' sources said the round was being wrapped up and details could still change.
Who invested in DeepSeek's new round?
CATL and Tencent were among the biggest investors, per Bloomberg and Reuters. Chinese media also list Legend Capital, CPE Yuanfeng, a Geely fund, SMIC's investment arm, IDG Capital and others, which the companies have not confirmed.
Is DeepSeek going public?
Reuters reported in September that DeepSeek hired CITIC Securities to prepare a possible STAR Market IPO in Shanghai. Chinese media say a filing could come by the end of 2026, but DeepSeek has not announced a listing.
Sources & further reading
- DeepSeek Raises $15 Billion in Challenge to Leading US Labs (Spanish edition: DeepSeek recauda US$15.000 millones para competir con los líderes de IA de EE.UU.) ↗
- CATL, Tencent back DeepSeek's latest $15b financing (Reuters) ↗
- DeepSeek considers doubling latest funding round to up to $15 billion, sources say ↗
- DeepSeek完成约千亿元融资：金额半月翻倍，IPO铺路进行时 ↗
How this story was made
Written by Kristian Kostov with AI assistance and checked against the linked sources. Company performance claims are attributed to the company. Analysis reflects AiLookout’s interpretation; we have not independently tested the products discussed. The cover places the companies’ own logos, taken from public-domain or openly licensed files credited in the caption, on a plain background. It is not a product photograph, a screenshot, or an endorsement.
Our editorial standards
