Wenfeng Liang Halts DeepSeek's Second Round of Funding Due to Dissatisfaction with Leaked Closed-Door Meeting Details

Original Title: "Liang Wenfeng Stops DeepSeek's Second Round of Financing Due to Dissatisfaction with Leakage of Closed-Door Meeting Content"
Original Author: Dynamic Insight Beating

On July 25, Bloomberg reported, citing informed sources, that DeepSeek has paused its ongoing second round of financing and verbally informed some potential investors that the company will not sign the investment agreement as previously expected.

The pause is partly due to founder Liang Wenfeng's frustration over the leakage of content from a closed-door meeting. The meeting took place during the first round of financing, where Liang Wenfeng spoke to potential investors for nearly four hours, discussing commercialization, roadmap models, computing power reserves, domestic chips, and future revenue. Two months later, a tens of thousands of words transcript of the recording began circulating on the internet.

According to Bloomberg's disclosure, this round of financing is only temporarily suspended and has not been completely canceled. Negotiations may still resume in the future. DeepSeek's domestic IPO preparation is also not halted, with the possibility of submitting a listing application as early as this year.

In June this year, DeepSeek completed its first external financing since its establishment, with the amount disclosed by several media outlets being around 50 billion yuan, and a post-investment valuation of about 3.5 trillion to 4 trillion yuan.

Just after the first round was concluded, the company began reaching out to investors in mid-July to plan for the second round. Reuters reported on July 15, citing another group of informed sources, that the company aimed to raise funds at a valuation of around 500 billion yuan, with a potential scale of up to 50 billion yuan.

It was reported that some institutions were expecting to enter the signing phase soon.

This round of financing is considered directly related to computing power expansion and the IPO plan. Bloomberg previously reported that DeepSeek has started engaging financial and banking advisors, with an internal goal of submitting an application as early as this year to aim for a listing completion in 2027. If the second round lands, it could not only supplement funds before the IPO but also establish a higher valuation benchmark.

Leaked Documents

The meeting took place on May 20, lasting approximately 3 hours and 44 minutes, with Liang Wenfeng introducing the business philosophy, open-source logic, and AGI roadmap in the first half, followed by an investor Q&A in the second half.

From July 22 to 23, this material began to spread in the domestic science and technology and investment circles, and was subsequently widely reposted by numerous media, self-media, and overseas communities. According to the media's verification with institutions involved in DeepSeek's investment, it was confirmed that the closed-door meeting did occur in May, and the overall content is deemed authentic and credible.

Before Liang Wenfeng finished his keynote speech and the conference moved to the Q&A session, the host specifically reminded the attendees: "Liang Wenfeng has shared a lot of sensitive information, so please do not leak any numbers or details, including card volumes. Also, absolutely do not screen record for sharing outside."

Two months later, the content that was explicitly mentioned not to be shared leaked almost in its entirety on the internet.

What Was Discussed at the Conference

The rapid spread of the transcript was due to Liang Wenfeng rarely publicly discussing DeepSeek at length and a significant amount of previously undisclosed company information being revealed during the conference.

Liang Wenfeng still places AGI at the forefront of the company's goals. He believes that both B2C users and B2B business, as well as commercial revenue, are merely transitional products in the process towards AGI. The company is currently not planning to arrange product and research directions around short-term profits.

He summarized AI development as a series of ascending steps: starting with language models, then moving to mind chains, and next to Agents, with the most critical issue being continual learning. DeepSeek hopes that the next-generation models will not only be used by external users but will also assist their own researchers in developing the next generation of models. At the current stage, the company places more emphasis on General Agents and Coding Agents, while the priority for vertical Agents in finance, healthcare, and other sectors is relatively low.

Computing power and chips are the most sensitive topics. The transcript stated that the primary gap between Chinese and American AI companies is not talent but resources. Liang Wenfeng estimated that DeepSeek may have only used one-twentieth of the computing power of its American counterparts in the past, leading to an overall model time lag of about 12 to 18 months. The future goal is to reduce this to 3 to 6 months and achieve breakthroughs in certain directions. The document also mentioned DeepSeek's equivalent computing power, the scale of chips required for future models, the number of chips provided by Huawei, and the performance conversion between Huawei cards and NVIDIA cards. These numbers were repeatedly cited in the dissemination, but there were no public procurement contracts or formal confirmation from DeepSeek or Huawei.

Regarding the team, he mentioned that the company can exercise restraint in many aspects of interests, but there can be no compromise on the stability of the core team. The recent funding round provided a true valuation of employee stock options, largely mitigating the talent outflow risk.

From the leaked content, it is evident that what was revealed was not a DeepSeek scandal, but rather a quite positive spread of information. Moreover, the emphasis on open-source, prioritizing AGI research, and focusing on model efficiency were all judgments that have been present in his previous public interviews.

The issue lies in the fact that this was a closed-door meeting with investors that had a clear confidentiality requirement. In addition to ideas and judgments, the meeting also involved discussions on computing power reserves, chip supplies, model scales, revenue forecasts, and the use of funds. This information not only concerns DeepSeek itself but may also involve suppliers, investment institutions, and ongoing business negotiations.

For a company that is about to undergo a second round of funding and is preparing to enter the public market, the ability to maintain confidentiality is part of the investors' due diligence. The full content of a closed-door meeting will be widely disseminated two months later, forcing the company to reassess the attendee list, communication methods, and the boundaries of information disclosure.

Why the Confidence to Pause Funding

One key reason the second round could be paused just before signing is that DeepSeek is currently not under cash pressure.

Prior to this year, DeepSeek had not engaged in external equity financing for a long time. The company emerged in 2023 from Phantasm Quantitative's AI research system, with early R&D investment mainly coming from Phantasm's operating income and R&D budget. Funding rumors started circulating in April this year, and within a few weeks, the valuation and scale rapidly escalated. By June, the first round was completed, becoming one of the largest single rounds of financing in the Chinese AI industry.

According to Reuters, Liang Wenfeng personally invested around 20 billion yuan in this round, making him the largest individual investor. Tencent invested about 10 billion yuan, CATL about 5 billion yuan, with NetEase, JD.com, IDG, and the National Artificial Intelligence Industry Investment Fund, among other institutions, also participating.

The post-financing equity structure shows that Ningbo Cheng'en Enterprise Management Consulting Partnership holds about 60.19% of the operating entity, Liang Wenfeng directly holds about 31.01%, Hangzhou Chengli Platform, where external investors are located, holds about 8.52%, and the National Artificial Intelligence Industry Investment Fund directly holds about 0.28%. Domestic media, after penetrating through multiple layers of partnerships, estimated that Liang Wenfeng's direct and indirect controlling interests are around 84.29%, a figure that is more in line with the control rights method, and should not be understood as him directly holding 84.29% of the shares.

According to The Information and other media reports, most commercial investors enter through limited partnership platforms, face a five-year lock-up period, and do not have voting rights in the operating company; the National Artificial Intelligence Industry Investment Fund holds direct shares and retains voting rights.

In other words, even after external capital comes in, the main decision-making power still lies with Liang Wenfeng. Pausing the second round after just completing a funding round of about 50 billion yuan will not immediately impact the company's operations.

Phantasm Quantitative is often seen as another source of his confidence. The average return rate of the fund under this organization is about 56.6% in 2025, ranking second among large domestic quantitative institutions with over 10 billion yuan in assets under management, and currently holding assets under management exceeding 70 billion yuan. What truly demonstrates Liang Wenfeng's financial strength is the 20 billion yuan he put forth in the first round of funding.

Still Not Enough Money

DeepSeek's current flagship model is the V4 preview version released in April this year, originally scheduled to launch the official version in mid-July, divided into Pro and Flash versions. V4-Pro has a total of 1.6 trillion parameters, with 490 billion parameters activated per token; V4-Flash has approximately 284 billion total parameters, with 13 billion parameters activated. Both versions support a context of 1 million tokens and have opened up model weights.

As competition moves from chat and reasoning to Agents, continual learning, and larger-scale model training, the company needs to invest in more chips, data centers, high-quality data, and engineering talent. Although the scale of the first round of about 50 billion RMB is considerable, it still lags significantly behind the capital investment of top AI companies in the United States. This is also the reason why the second round quickly started soon after the first round closed.

As of the time of publication, DeepSeek has not publicly responded to the suspension of the second round of financing, nor has it provided any explanation regarding the authenticity of the leaked transcript, the dissemination process, and the sensitive figures.

Original Article Link

Welcome to join the official BlockBeats community:

Telegram Subscription Group: https://t.me/theblockbeats

Telegram Discussion Group: https://t.me/BlockBeats_App

Official Twitter Account: https://twitter.com/BlockBeatsAsia