DeepSeek's $12 Billion Funding Round and Shanghai IPO Plan: What Is Confirmed
Bloomberg says DeepSeek is close to raising at least 80 billion yuan, led by Tencent and CATL. Here is what is reported, what is not, and why the money matters.
In this story 11 sections
DeepSeek is reportedly close to raising at least 80 billion yuan, about $12 billion, in a round led by Tencent and CATL. Bloomberg reported the figure on October 6, and Reuters could not immediately verify it. DeepSeek has not announced the deal, and its valuation and IPO timing remain unconfirmed.
Emergent Wire covers AI models, capabilities, analysis and the business behind them. This explainer is for readers who want a clear separation between what has been reported, what is disputed and what is only rumor about China's most-watched AI lab. It covers the round, the investors, the IPO preparation and what the money could change.
What Has Been Reported About the Round?
The core report comes from Bloomberg, which cited people familiar with the matter. As syndicated by The Star, DeepSeek is close to securing at least 80 billion yuan ($11.93 billion). Battery maker Contemporary Amperex Technology (CATL) and Tencent Holdings are reported to have committed among the largest amounts.
The Next Web added detail. It said the round initially targeted about 50 billion yuan, and that demand rose after the launch of DeepSeek's latest model. The final total could approach 100 billion yuan.
Reuters said it could not immediately verify the Bloomberg report. That caveat matters. Until DeepSeek or its investors confirm terms, treat every figure in this story as reported, not final.
How Does This Compare With DeepSeek's Earlier Raise?
This is DeepSeek's second outside round this year, which is itself unusual for a lab that long ran on its founder's backing. The Next Web reported the first round raised about 50 billion yuan at a valuation of $52 billion to $59 billion. Founder Liang Wenfeng was the largest investor, contributing about 20 billion yuan of his own money.
Tencent put in about 10 billion yuan and CATL about 5 billion yuan in that earlier round. JD.com, NetEase and China's national AI investment fund also took part. The new round, if it closes at the reported size, would be well over the first in dollar terms.
The step up is fast. A second raise within months of the first suggests the company either needs more capital for compute and talent or sees a window it wants to use. DeepSeek has not said which, and no executive has commented in the reports we reviewed.
What Do We Know About the Valuation?
We do not know it, and reports conflict. A Dealroom summary of CNBC coverage put the round at $12 billion to $15 billion and the valuation at roughly $75 billion. A July Reuters headline said DeepSeek sought funding at a $74 billion valuation.
The Bloomberg report syndicated on October 6 did not include a valuation. The Dealroom page is a secondhand summary and does not name investors. Weigh those figures accordingly, and expect the numbers to move before any announcement.
| Item | Reported figure | Status |
|---|---|---|
| Round size | At least 80 billion yuan (about $12 billion) | Bloomberg, unverified by Reuters |
| Possible upper end | Near 100 billion yuan | Reported, not confirmed |
| Valuation | About $74 to $75 billion | Conflicting secondary reports |
| Lead investors | Tencent and CATL | Reported |
| IPO venue | Shanghai STAR Market | Preparation reported, no date set |
What Is the Shanghai IPO Plan?
DeepSeek has reportedly hired CITIC Securities to prepare a possible initial public offering on Shanghai's STAR Market. The Next Web, citing a Reuters report from the previous month, said the timing, valuation and scale of any listing have not been set. The STAR Market is Shanghai's board for technology companies.
An IPO would give domestic investors a way into one of China's best-known AI companies. The CNBC-based summary linked the interest to DeepSeek's low-cost models, which challenged assumptions about how much AI development costs. It also read the round as a sign of investor appetite for Chinese AI despite U.S. export restrictions on advanced chips.
Plans like this can change. Listing rules, market conditions and regulatory approvals all affect timing, and companies often delay or reshape deals. Treat any specific IPO date you see as speculation unless DeepSeek confirms it.
Who Are Tencent and CATL, and Why Would They Invest?
Tencent runs WeChat and a large cloud business, so a stake in a leading model lab helps it compete with domestic rivals. Reuters reported in June that closer ties to DeepSeek could help Tencent keep up. CATL makes electric vehicle batteries and is developing solid-state batteries.
CATL's interest is less obvious. A battery maker may see AI as a tool for materials research, manufacturing and energy management. It may also see a financial opportunity in a company that has become a symbol of Chinese AI.
These motives are our inference, not statements from the companies. Neither has publicly explained its investment in the reports we reviewed. Strategic investors often want access to technology as much as a return, and that can shape how a lab operates.
What Could the Money Change for DeepSeek and Its Models?
More capital usually buys compute, talent and time. Training frontier models costs enormous sums, and chips are the largest expense. Funding at this scale would let DeepSeek buy or rent more hardware and hire researchers away from rivals.
It could also affect how the company releases models. DeepSeek built its reputation on open releases and low prices. Outside investors and a public listing bring pressure for revenue, and that can change the balance between open weights and paid services.
We have seen the pricing side of its work recently. Our piece on DeepSeek V4.1 Flash and why agent runs get cheaper explains the cost angle. Our analysis of the AI compute buildout as a power problem shows why money alone does not guarantee capacity.
How Did DeepSeek Get Here?
DeepSeek grew out of High-Flyer, the quantitative hedge fund that Liang Wenfeng founded. It drew wide attention in early 2025 with its open-source R1 reasoning model and launched its V4 family of models this year. That track record explains investor interest more than any single financial metric.
The company has been unusually quiet about its business. It has not published audited financials in the reports we reviewed, and it rarely gives interviews. Investors are betting on technical reputation, low serving costs and the founder's record.
That reputation is also a risk. A lab known for efficiency has to keep delivering new models while its costs rise. If capital pushes it toward heavier spending, the efficiency story that made it famous could weaken.
What Risks Do Investors and Users Face?
Several risks stand out. Regulatory approval for a STAR Market listing is not guaranteed. Chip supply remains constrained by export controls. Competition from both Chinese and U.S. labs is intense, and price cuts across the industry squeeze margins.
For users, the main risk is change. A company preparing for an IPO may adjust pricing, terms of service or licensing. Open-weight releases that were free yesterday may come with new conditions tomorrow.
None of this means DeepSeek is a bad bet. It means the picture is incomplete, and the honest summary is that the company is well funded, closely watched and still unconfirmed on the details that matter most.
How Does This Fit the Wider AI Funding Picture?
A $12 billion round is large but not unique. Our coverage of OpenAI's $30 billion raise and Anthropic's IPO valuations shows how big the numbers are across the industry. Frontier labs now raise money in the billions each year.
What is different about DeepSeek is the source of the money. The round is reportedly backed by Chinese strategic and state-linked investors, and it is aimed at a domestic listing. That separates it from the dollar-denominated rounds at U.S. labs and affects how investors, customers and governments see it.
Chip export controls add another layer. DeepSeek can raise capital, but access to advanced hardware depends on policy and supply chains. How it spends the money will show how it handles that constraint.
What Should Readers Watch Next?
Watch for four things. First, an official announcement from DeepSeek or an investor confirming the round. Second, a disclosed valuation. Third, a formal IPO filing or exchange application. Fourth, the next model release and any change in licensing.
At Emergent Wire, our editors treat unconfirmed funding numbers as a range, not a point. The pre-IPO period is when leaks are common and terms shift. We will update our reading when a primary source appears.
If you use DeepSeek models in a product, plan for a stable API and pricing, but have a fallback. Ownership changes and IPO preparation can alter terms. A second provider on standby is cheap insurance.
Conclusion
DeepSeek's reported $12 billion round, led by Tencent and CATL, would make it one of the best-funded AI labs outside the United States. Valuation, final size and IPO timing are unconfirmed, and the numbers differ by source. The sensible reading is a range with clear caveats.
Next step: if you depend on DeepSeek models, bookmark the company's official channels, track any announcement of terms, and confirm you have a backup model ready.