DeepSeek closes in on $12 billion funding round backed by Tencent and CATL ahead of 2027 IPO

DeepSeek, the Chinese artificial intelligence developer that startled global markets with its low-cost reasoning models, is closing in on a financing round that would rank among the largest ever assembled by a private AI company. The Hangzhou-based firm is close to securing at least 80 billion yuan, roughly $12 billion, according to people familiar with the matter, a total that would blow past the 50 billion yuan target it set when it began canvassing investors earlier this year.

The round, which is expected to close soon, has drawn commitments from some of China's most prominent industrial names. Battery giant Contemporary Amperex Technology, widely known as CATL, and Tencent Holdings, the operator of the WeChat messaging platform, have committed among the largest amounts, the people said. Their participation signals that appetite for Chinese AI assets now extends well beyond traditional venture capital and into the country's strategic manufacturing and internet champions.

The fundraising arrives ahead of a planned initial public offering on Shanghai's STAR Market in early 2027. DeepSeek has hired CITIC Securities to prepare for the listing, although the timing, valuation and size have yet to be finalized, according to Reuters. The company began seeking new capital in July at a targeted valuation of 500 billion yuan, or about $74 billion, just a month after raising roughly $7.4 billion in its first external funding round.

DeepSeek first came to global prominence with the release of its R1 model in early 2025. It followed that with DeepSeek-V4.1-Flash in September 2026, a model the company said was designed for greater capability, faster inference, higher throughput and scaling to larger models. The company has also been expanding its computing infrastructure at home, as Chinese AI firms seek greater access to domestic hardware and work to reduce reliance on US chip suppliers.

Key Facts

Bloomberg reported on October 6, 2026 that DeepSeek is close to securing at least 80 billion yuan, or $11.93 billion, in its latest round, a figure that exceeds the company's own capital-raising target by a wide margin. The Business Times reported on October 6 that DeepSeek initially sought about 50 billion yuan but that investor demand outstripped expectations following the successful release of its latest AI model. Based on term sheets already signed, the final tally could approach 100 billion yuan, or roughly $15 billion.

On valuation, DeepSeek had originally targeted about 500 billion yuan, roughly $74 billion to $75 billion, for this financing. Proactive Investors reported on October 6 that the round would value the company at a minimum of about $75 billion, and that proceeds could reach almost $15 billion if commitments approach the 100 billion yuan mark. That would compare with DeepSeek's first major external round earlier in 2026, which raised about $7.4 billion and valued the company at between $52 billion and $59 billion.

Tencent and CATL have committed among the largest amounts in the current financing, according to people familiar with the matter. Tencent had previously invested 10 billion yuan in the round that valued DeepSeek at roughly 350 billion yuan at the time. Founder Liang Wenfeng contributed again in that previous 50 billion yuan round, and his net worth more than doubled to around $36 billion when DeepSeek closed its first external round this summer.

On infrastructure, DeepSeek plans to deploy at least 160,000 of Huawei Technologies' top accelerators at a massive data centre it is building in Inner Mongolia, potentially one of the largest known clusters of Huawei AI chips. Last week DeepSeek announced a partnership with Huawei to develop programming tools optimized for Huawei's Ascend AI chips, an alternative to Nvidia's CUDA ecosystem.

The momentum is not only financial. TechStartups reported on October 6 that the news came less than two weeks after DeepSeek reached a $1 billion annualized revenue run rate, more than double its pace from a few months earlier. Bloomberg Intelligence said the performance gap between leading Chinese and US AI models has narrowed to about 3% on benchmark scores, down from roughly 9% in May.

Analysis

What this really means is that DeepSeek is no longer just a model developer competing on benchmark scores. It is becoming the anchor tenant of a domestic Chinese AI stack, with the balance sheet to match. When a battery supplier like CATL and a consumer internet platform like Tencent both write some of the largest cheques in the same round, the funding stops being a venture bet and starts looking like industrial policy executed through private markets.

The oversubscription is striking because DeepSeek did not need a new narrative to attract it. The company began seeking capital in July at a targeted valuation of 500 billion yuan, about $74 billion, just a month after raising roughly $7.4 billion. Investors were not asked to underwrite a promise. They were responding to a model release: the V4 Flash, also reported as DeepSeek-V4.1-Flash, which reset expectations for cost versus performance against leaders Anthropic and OpenAI. The fact that commitments ran past the original target, and that the final tally could approach 100 billion yuan, tells you that the marginal buyer now believes Chinese frontier models can compete on economics rather than only on national preference.

The presence of CATL matters as much as the size of the cheque. CATL is the world's largest electric vehicle battery maker, a company whose core business depends on manufacturing scale, supply chain control and the cost of compute-intensive engineering. Its decision to take a large position in an AI lab is a signal that China's industrial leaders view model capability as an input to their own products, not as a separate software category. Tencent, for its part, already had a seat at the table through the earlier 10 billion yuan commitment, and its willingness to increase exposure suggests it sees DeepSeek as strategically entwined with WeChat-scale services.

There is also a control question. Liang Wenfeng contributed again in the previous 50 billion yuan round, and his net worth more than doubled to around $36 billion when DeepSeek closed its first external round this summer. A founder who keeps writing personal cheques while outside capital floods in is deliberately preserving influence. That matters when the next step, according to the people familiar with the matter, is a restructuring ahead of an IPO. Rival Moonshot is also targeting an early 2027 IPO after closing funding at a $50 billion valuation, so the window for establishing a public-market narrative is narrowing at the same time as DeepSeek's private valuation climbs.

Why It Matters

The round matters first as a statement about hardware sovereignty. DeepSeek's plan to deploy at least 160,000 Huawei accelerators in Inner Mongolia, combined with its Huawei partnership on programming tools for Ascend chips, is a direct attempt to build a training and inference stack that does not depend on Nvidia's CUDA ecosystem. The funding gives DeepSeek the cash to buy domestic silicon at a scale few Chinese buyers can match. If the cluster works, it becomes a reference design for the rest of the industry.

Second, the numbers reframe the competitive gap. Bloomberg Intelligence's estimate that the performance gap between leading Chinese and US models has narrowed to about 3% on benchmark scores, down from roughly 9% in May, is the kind of claim that shifts procurement decisions. Combined with a $1 billion annualized revenue run rate reached less than two weeks before the funding news, it gives enterprise buyers a reason to evaluate DeepSeek on merit rather than geopolitics.

Third, the STAR Market listing would give public investors in China direct exposure to a frontier AI lab at a moment when such assets are scarce on domestic exchanges. CITIC Securities is already preparing the groundwork. If the IPO lands in early 2027, it would be a landmark for Shanghai as much as for DeepSeek.

Next Up

Once the round closes, the next step should be to restructure the company to prepare for an IPO, according to the people familiar with the matter. Reuters reported on October 6 that DeepSeek expects to complete the round in October, with investor commitments already exceeding the initial goal. The restructuring, the choice of listing venue and the final valuation will all be watched closely, though DeepSeek has said the timing and size of a STAR Market offering remain undecided.

For now, the immediate test is confirmation. DeepSeek did not respond to requests for comment during a Chinese public holiday, and CATL and Tencent either declined to comment or did not immediately respond, according to Reuters, which said it could not immediately verify the Bloomberg report. The silence leaves the 100 billion yuan upper figure as a term sheet projection rather than a closed number, and the gap between the two will define how the market prices the next phase.