This is an emergency translation of an account of the recently closed DeepSeek fundraising. We think this offers a unique glimpse of what type of company DeepSeek is and what type of person Liang Wenfeng is.
The original article was written in Chinese by Elsewhere别处发生, a writer and podcaster who typically reports on China’s venture capital and startups. It’s a partial translation that leaves out some details unrelated to DeepSeek.
The translation below is not authorized by the original author. Emphasis is our own.
The “four-hour meeting”
First, let’s talk about that investor meeting—the one whispered about as the “four-hour meeting.”
In mid-May, DeepSeek hosted this virtual meeting via Tencent Meeting. By then, the list of investors was mostly settled. Each firm was allowed to send two attendees to the meeting. The session began with a presentation by Liang Wenfeng, followed by a Q&A where he addressed the investors’ inquiries.
Legend has it that one CVC (Corporate Venture Capital) investor spent a full ten minutes just asking questions—starting with a lengthy self-introduction followed by three drawn-out questions. The atmosphere in the virtual room turned awkward for a moment. It’s easy to imagine the scene, with some participants whispering to one another, wondering who this person was.
Liang, however, seemed unfazed and continued to answer every question earnestly. At one point, after finishing the second question, he forgot what the third one was and had the investor repeat it.
For many of those present, even those who had already committed capital, this was their actual first interaction with Liang Wenfeng, as most had never met him in person. By the end of the call, many were left with a lasting impression, as if his words were still echoing in their minds.
Liang Wenfeng’s aphorisms and spirit began to circulate throughout the market.
Restraint and AGI
Liang Wenfeng is not a naturally gifted orator, and he speaks at a measured pace. Yet, several quotes from that meeting are frequently repeated (paraphrased, as the exact wording may vary slightly):
“We have no fame and no influence; we are just a group of very ordinary people.”
“The narrative I like is a group of ordinary people achieving something extraordinary. For DeepSeek, our core focus is ensuring team stability, which is more important than money or resources. It is the greatest risk and the greatest challenge.”
“Assuming AGI will account for 20% of GDP, those who want 5% will lose to those who want 1%; those who want 1% will lose to those who want 0.1%.”
“What we want to do is anything related to improving intelligence; we will not do anything else.”
“AGI is a big enough endeavor. Everything else is just a process.”
Throughout those four hours, Liang consistently emphasized the same themes: the singular pursuit of AGI, the philosophy that “less is more,” and the importance of restraint.
From 5 Billion to 1.5 Billion
Official fundraising began around April of this year. It is said there was a “whitelisting” of funds—a list consisting of institutions with significant capital capacity and strong brand reputations.
Initially, DeepSeek required a minimum investment of 5 billion RMB, prohibited the syndication of shares, and insisted on a “pure” RMB fund structure. Given this, one prominent investor who was widely rumored to be a guaranteed participant ultimately did not join. Industry rumors suggest this was because they had tried to sell portions of their stake to Middle Eastern LPs.
However, in China, there are very few market-oriented funds capable of committing 3 billion RMB in a single shot. Consequently, many VCs negotiated to see if they could participate through multiple fund entities.
Ultimately, the minimum investment threshold was lowered to 1.5 billion RMB, and the fundraising structure became more flexible.
There was also a claim that participating institutions were preferred if they had not invested in other LLM companies. This was, in all likelihood, an impossibility for a simple reason: finding a top-tier fund that hasn’t invested in any LLM companies is a contradiction in terms.
During the investor meeting, Liang Wenfeng also expressed that DeepSeek is willing to help other companies, including other large-model developers. The caveat, of course, remains: as long as they don’t try to poach his talent.
[Some parts about participating venture capital funds are redacted. You read about them in the original article.]
The most important demand: don’t poach DeepSeek’s people.
I won’t go into DeepSeek’s investment terms here. I’ll add only one point: whether dealing with big tech companies or VC funds, Liang Wenfeng’s most important demand was this: do not poach DeepSeek’s people, or encourage them to leave and start their own companies.
“China’s largest listed company“
If the issue is reduced to the level of a single investment, what kind of company is DeepSeek, in terms of odds of success?
One investor who participated in the deal once worried, perhaps unnecessarily, whether this might be an investment with too much consensus behind it. According to the experience of the investment industry, consensus investments usually do not end well.
But with DeepSeek, it is hard to say that. Or rather, we cannot understand it through a simple investment lens, whether in terms of its significance for China’s AGI industry or on a more spiritual level.
As one investor put it: “They are doing all of this with enormous goodwill. He seems to be telling us: the world was always supposed to be this way.”
So some hesitated, while others believed there was “no reason at all not to invest.” If, as some expect, DeepSeek eventually goes public, could it become the largest listed company in China’s A-share market?
Low profile and ambition
In following DeepSeek’s fundraising progress, one feeling stood out: for most people, DeepSeek was probably an utterly unexpected answer. Even the investors who most love to win had not imagined that one day it would open itself up to financing.
So, as mentioned earlier, many of the well-known investment heavyweights only recently managed to get in touch with the company. Of course, Liang Wenfeng had long avoided meeting investors. After news of the fundraising emerged, many institutions did not even try to compete for an investment opportunity.
But judging from the outcome, some secured a ticket through so-called synergy, some through persuasion, some through branding, and some simply through persistence.
Elsewhere once tried asking several investors about the deal. Most did not want to say much. The sensitivity of the transaction was only one reason. More than that, there seemed to be a kind of self-conscious restraint: because they felt honored to be part of it, they wanted to preserve its low-key nature.
One participating investor said that throughout the transaction process, he kept every document carefully, hoping that one day he might have the chance to frame them.
Another investor repeated twice the 16-character line released with the preview version of DeepSeek V4:
“Unmoved by praise, unafraid of slander, follow the Way, and uprightly correct oneself. 不诱于誉,不恐于诽,率道而行,端然正己。”
Related Reads
China's self-sufficiency in 35 critical "stranglehold" technologies
"Stranglehold" technologies are critical industrial capabilities where dependence on foreign suppliers creates strategic vulnerabilities. Think semiconductors, advanced lithography machines, or aircraft engines—components so complex that few nations can produce them independently.
DeepSeek is a soul-searching moment for China's struggling VC industry
DeepSeek is a wake-up call for the US. But in a more subtle way, it’s also a wake-up call for China, especially China’s venture capital industry.
Can China catch up with the US in AI?
This article is based on a recent lecture I gave at Peking University to a large group of European students studying in the dual degree programs between PKU and LSE / Sciences Po. My lecture centers on the US-China tech competition, covering 3 vectors: AI, data, and venture capital. My thesis on data and venture capital has been well-covered by our previous articles (on





