Transcript of Liang Wenfeng's Four-Hour Investor Meeting

By: rootdata|2026/07/23 02:44:00

Author: elsewhere

Last month, "elsewhere" reported on the financing story of DeepSeek. Among the most discussed topics was the legendary four-hour investor meeting.

In the past month, various quotes from Liang Wenfeng have circulated widely, and we have gathered some of the content from those discussions.

During the meeting, Liang Wenfeng repeatedly emphasized the word "no": not a genius, not pursuing unreasonable profits, not chasing user numbers, not closing source code, not engaging in 3D/video generation/world modeling, and not creating the next super app. He stated that restraint is a strategy used to increase the probability of achieving AGI.

From the limited materials we gathered, the following high-frequency terms emerged: model, cost, AGI, time, open source, etc.

Most of the time, Liang Wenfeng's tone left room for interpretation, using simple and straightforward language. Only when discussing a few issues he cared about did he show some sharpness: "As long as I can maintain the stability of the team, I can definitely achieve AGI. It's that simple."

Here are 52 statements we collected. Some phrases may slightly differ from the original while retaining their meaning.

DeepSeek Has Only One Main Line

Now is not the time for maximizing product profits. The path to AGI requires passing through the product stage, but we do not need to invest too much thought and energy into C-end or B-end products. Approaching lower-level technologies from a high technical standpoint is a dimensionality reduction attack. Products are a byproduct on the road to AGI.
Many things are not on our main line, such as 3D, video generation; similarly, world modeling has little to do with the limits of intelligence.
Multimodality is important for products and C-end users, but it is just a component, not the main line or intelligence itself.
The hallucinations of large models can certainly be addressed, but it is a long-term proposition. Internally, we categorize hallucinations as a product issue that we will address, but it is not a priority.
At this stage, the most important thing is still the Coding Agent. Given the domestic situation, the most reasonable approach should be to fully focus on a general-purpose Agent, while the priority for other Agents in finance, healthcare, etc., is lower.
If the AI era produces many trillion-dollar companies, it would be great if DeepSeek is one of them.

First Achieve Continuous Learning, Then AI Self-Iteration, the Endpoint is Embodied Intelligence

AI currently lacks taste and intuition; what it lacks is the ability for continuous learning.
Humans can continuously learn, but for the same task, AI must be given all the context. This is almost impossible, so AI cannot replace employees. Therefore, the next generation of models must have the ability for continuous learning to be called the next generation of models.
We hope the next generation of models can assist in our own development. To put it simply, our goal for the models we create is not for everyone to use them well, but for us to use them well. This is the fastest way to achieve AGI.
Currently, no good methods have been found worldwide because "learning" consists of many components.
DeepSeek's long-term vision is AGI. If we compare the path to realization to climbing stairs, last year's step was CoT (Chain-of-Thought), and this year's step is Agent. The problem to solve after Agent is continuous learning.
After achieving continuous learning, we may reach a gradual singularity: models can accomplish all tasks that humans can do, including developing more advanced AI models themselves. That is, AI can accelerate AI research. Only after this step is completed can we achieve embodied intelligence.
The endpoint of intelligence may be embodied. Because for a normal person, their need is not for computers, but for human power.

A Complete Shift to Commercialization is Still Far Away

We only earn a reasonable profit, not a profit-maximizing price.
Initially, we were worried that demand for one of our models would be too high, so we set the price relatively high, but later reduced it to a quarter, and many people in the company cheered. Because this is the purpose of our hard work in developing this model: to allow everyone to use it fully.
Low cost is a result; our models have always been architected to lower costs. We also hope that costs are affordable, especially in the context of limited computing power.

Another reason: the lower the cost, the larger the model we can afford. When computing power is limited, high computational efficiency allows for training larger models. Large companies can solve this by adding resources, while we prioritize cost efficiency.

From the outside, it may seem that we have chosen a very difficult model. But in fact, we are doing it very easily. Lowering prices is definitely not good news for our competitors; they are certainly not cheering. I don't think selling APIs is that attractive. I just need to have a few people maintain this API, and I don't even need customer service or sales; users will come on their own.
We have always been commercializing, just not with commercialization as our goal. The time point for DeepSeek to completely shift to commercialization should be very far away.
I don't even need to consider taking a position in it at that time. As long as there is such a large business opportunity, there will definitely be a way. DeepSeek is a product of an era, a reflection of reality, not a result of imitation.

Open Source is the Sweet Point for Companies of Our Scale

Restraint is a strategy: sacrificing some things in exchange for more of others. Open source is a way to give back: internally, employees gain a sense of achievement, and the company has cohesion; it also benefits society, as peers or ordinary people will be happy.

I have no doubt that AGI will have tremendous commercial value. Based on this, my priority is not to take a larger share but to increase the probability of success.

To make AI successful in business, open source is beneficial. This may sound counterintuitive. Historically, a software company might generate tens of billions of dollars in the market in a year, but once it goes open source, it disappears. However, AI is large enough that it may eventually account for ten percent of human society's GDP. If we want to monopolize this benefit, we will definitely be abandoned by history. This is an objective law, a historical perspective.
The open-source models we provide are the same as the models we deploy ourselves. We will not open source a subpar model while using a better model for our own deployment.
I am not worried about others deploying our models to compete with us. Not every company has the willingness and capability to achieve this goal. Startups that are too small lack the strength to do this; large companies find it difficult to organize. This is a sweet point for companies of our scale.
Open source has no impact on our business model, provided that we "only earn a certain profit." If you want to earn a hundred times the profit, open sourcing will indeed affect you.
We are unwilling to become competitors with any large or small internet companies. Under this premise, we are very willing to assist and help anyone, including Alibaba, Zhizhu, and Yuezhianmian, to do better.

The Gap Between China and the U.S. is Not in Talent

In the future, we want to rewrite the narrative of AI between China and the U.S.: using a fraction of computing power to shorten the gap to six months or three months.
The gap in AI between China and the U.S. mainly lies in resources. We believe in scaling; the larger the scale, the better the effect. Training such large models is not because I think such large models are enough, but because there are only so many resources.
There is almost no gap in talent------ it is the same group of people. There is no shortage of talent domestically. Talent shortages are temporary; historically, there has never been a long-term shortage of a particular type of person.

Model Factory Competition, Cost is the Top Priority

Anthropic is now surpassing OpenAI, but this is not long-term; it is temporary. OpenAI and Google are likely to alternate in rising in the future.
There are too many companies in China making models, and each one is doing the same thing, leading to a dispersion of resources. Eventually, it will converge, but it requires a process. If each company only takes reasonable profits, there is no need for so many people to make large models: perhaps two large companies and two small companies are enough.
I absolutely do not believe that large model companies can take most of the profits in the AI industry.
The final gap in large model competition will be reflected in three aspects: cost, time, and user experience.

Cost is the top priority------ for the same quality of service, what cost can you provide it at? The second is time; being a few months early or late makes a difference. User experience may have some stickiness and barriers, but it is not essential.

No Intention to Become the Next Super App

We do not want to become the next super app. Do we want to become the next ByteDance? The next Tencent? We have no such thoughts at all.
Not competing for this is because there are still watermelons behind; what is in front may only be sesame. Of course, that sesame may be relatively large, but I don't think it counts as big.
Last year, everyone was scrambling for chatbots and C-end traffic; this year, it is about grabbing To B revenue. But we do not think it is important; what the company is really concerned about is the roadmap for AGI and how to break through the next step in technology. It is strange that the most desired things are unattainable, while those that are not so valued are easier to obtain.
Last year's Spring Festival was a hit, but it was not in our script.

Maintaining Team Stability is Core

There is only one thing that cannot be compromised: we must maintain the stability of the team. This is also a significant risk we face. Of course, this risk has been significantly alleviated with this round of financing.
We do many things to maintain team stability. We do not want to become competitors with any large or small internet companies; we hope to empower and help them. We do not want to create enemies. This way, our own environment will also be better.
Some people think our organization is top-down, while others think it is bottom-up; I believe both are correct. Top-down is called "doing the right thing," and generally, we hope that "doing the right thing" does not occupy half of the employees' time. They still have half of their time to explore what they want to research, according to what they think is important, without preconditions.
We generally do not work overtime. The first reason is that research requires a relatively relaxed environment. The second reason is that we are very focused. Many products are not perfect, but we have not gone to fix them, which can also be seen as a culture of restraint.
Organizations are dynamic; they are not static. In the future, as the company grows, there may be some adjustments. However, it will not completely transform into a traditional hierarchical system, but there may be some necessary structures. The vision-driven approach will not change.

With Goodwill Towards the World

When we started this company, we did not intend to make a lot of money or go public in the capital market. The first few dozen people never thought this way. If they had, they would not have come. We are doing this with a great goodwill towards the world, believing it is useful for humanity.
"Achieving XX KPI" is not our way; we are a vision-driven organization. This has its advantages and disadvantages. In the future, we will find ways to leverage strengths and avoid weaknesses, but this is our characteristic.
The vision is not even written down; it is in the methods we use and our attitude towards the world. Perhaps everyone in our company has a different understanding of this vision, but in the larger direction, we are consistent.
About twenty years ago, the person I admired most in management was Jack Welch (former CEO of GE). Looking back now, most of what he said may no longer be correct, but he did say one thing right: the most important thing for a company is its vision.

The vision is not a slogan hung on the wall; it is not about what you say, but about what you do.

Restraint Enables Us to Achieve AGI

AGI is the most profitable. As for other things, if we have the energy, we will do them; if we don’t, we won’t. Restraint is part of our vision.
The AI matter is too big, and the benefits are too great. As long as we can achieve it, sharing a bit of the benefits is already significant. The more restraint you have, the more likely you are to achieve this.
I think this is intuitive------ at least it is intuitive to me. Besides the vision, we do not have many other advantages.
When this company was founded two years ago, there was not much money, no cards, no fame, and no appeal; we were just a group of very ordinary people. The narrative I like is "a group of ordinary people doing extraordinary things," rather than "a group of geniuses doing extraordinary things."
Open source is also part of restraint. In terms of pricing, we certainly do not start from the highest company revenue or profit. In the short term, a higher price means more revenue; but in the long term, it is hard to say. For me, restraint is a strategy.
Open source and low prices give employees a sense of achievement, which in turn fosters cohesion in the organization and benefits society, making other peers and ordinary people happy. Therefore, this kind of restraint, from a long-term perspective, can increase our probability of achieving AGI.
If your vision is to take more, you have already lost; you may face greater difficulties. This is how the world works.

-- Price

This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.