Exclusive Interview with the CEO of Plaud China: We Only Do What "Must Be Done" and What "We'll Die If Not Done"

Written by Zhou Xinyu

Edited by Su Jianxun

In terms of commercialization, Plaud is undoubtedly the most successful startup model among this wave of AI hardware companies.

With a card - shaped AI meeting note - taking product less than 3mm thick, this team that has been established for less than three years has sold its products to 170 countries, with a shipment volume of over one million units.

In an interview with Forbes in September 2025, Xu Gao, the co - founder of Plaud, mentioned that the total revenue in 2025 is expected to reach $250 million.

This revenue includes both the hardware products and the AI software Plaud Intelligence installed on them. The software adopts a subscription - based payment model, with the annual fees for the overseas version ranging from $99 to $240.

On September 22, 2025, Plaud held a new product experience event in Shenzhen, launching two card - shaped AI meeting note - taking products, Plaud Note and Plaud Note Pro, as well as a wearable note - taking product, Plaud NotePin S — This is also the first time Plaud has launched products in the Chinese mainland.

△ From left to right: Plaud Note Pro, Plaud NotePin S, Plaud Note. Image source: Company official

First, test the waters through overseas crowdfunding. After the product gains a certain level of attention, start selling it in the domestic market. This is the standard approach for current AI hardware startup teams.

But Plaud's story is more typical. In the early days, it didn't rely on a star - studded team or huge amounts of financing to stand out. Xu Gao, the co - founder, rarely made public appearances. Eventually, it achieved its current scale through product word - of - mouth.

What exactly is the secret to Plaud's success? One day before Plaud's products were launched in China, we had a conversation with Mo Zihao, the partner of Plaud and the CEO of the China region.

△ Mo Zihao, partner of Plaud and CEO of the China region. Image source: Company official

His answer is simple: Plaud has only one barrier, which is the extreme dedication of the team.

This definitely includes the focus on the matter of "AI meeting note - taking". Mo Zihao divides all choices into three categories: Can do, must do, and will die if not done. "We only do the latter two."

Before Plaud entered the Chinese market, big - name players such as DingTalk, Mobvoi, and 360 had already launched similar products with lower prices and more benefits.

This didn't bring any sense of crisis to Plaud. In fact, when entering the Chinese market, the pricing of Plaud's hardware products remained unchanged, still on par with overseas prices.

"The determination and concept of big companies' investment depend on how much they believe in this matter," Mo Zihao said. "We have ten times the manpower and ten times the dedication of others. The products we create must be significantly different."

Two days after DingTalk launched its AI voice recorder, the DingTalk A1, Mo Zihao pinned a WeChat Moments post: The appearance can be imitated, but the concept cannot be replicated.

Now, Plaud's global team has over 400 people, far exceeding the employee scale of most similar products. Mo Zihao told us that from the recruitment stage, these hundreds of employees must align with one thing: believing that Plaud can expand the boundaries of users' intelligence.

Plaud's concept stems from the emphasis on "conversation". Both Xu Gao, the co - founder, and Mo Zihao like to quote the famous words of the philosopher Ludwig Wittgenstein: "The limits of our language mean the limits of our world."

The team found that the work of many organizations and enterprises is not driven by "writing documents" but is completed through a large number of conversations.

The user needs for "writing documents" have been met by DingTalk and Lark. Plaud aims to fill the gap in the market for AI meeting note - taking tools for conversation - driven professions.

The difference in product concepts makes Plaud's design quite different from similar products.

For example, Plaud's newly launched "Ask Plaud" function allows users to have Q&A sessions with AI about the recorded content. Mo Zihao mentioned that this function should actually be called "Plaud Ask". "Others let you ask questions to AI, but we let Plaud ask you questions."

△ The newly launched "Ask Plaud" function of Plaud Intelligence can conduct in - depth Q&A sessions with users based on the recording. Meanwhile, all answers can be traced. Image source: Author's trial

Besides aligning with the concept, achieving "ten times the investment of others" also requires focus and a clear boundary between "what to do" and "what not to do".

Internally, Plaud has clearly defined three areas not to engage in: Not to develop document - related functions, not to develop modal conversion functions such as PPT generation, and not to develop lifestyle - related functions. This is to avoid resource dispersion and distraction, concentrating all efforts on doing a good job in AI meeting note - taking.

Before joining Plaud, Mo Zihao had nearly a decade of experience in mobile Internet startups. He doesn't agree with the practice of piling on functions in products and also tries to avoid using empty words such as "ecosystem" and "empowerment".

He told "Intelligent Emergence": "We want to return to the purest way of product development ten years ago: Find the element that brings about the greatest change and then provide users with the greatest value based on this element."

Mo Zihao defines both Xu Gao and himself as "people who have experienced failure". Before founding Plaud, Xu Gao had three not - very - successful startup attempts. Mo Zihao's first startup related to public welfare was, in his view, acquired because "it wasn't done well enough".

"Many times, success may not be because you did something right, but because you didn't do something wrong," Mo Zihao summarized. "So far, Plaud hasn't made any major mistakes."

The following is the conversation between "Intelligent Emergence" and Mo Zihao, the partner of Plaud and the CEO of the China region. The content has been sorted and edited:

Plaud is not a 'work - helper gadget'

Intelligent Emergence: Why did Plaud choose to launch its products in the Chinese market now?

Mo Zihao: Actually, we've been preparing for the Chinese region for quite a long time.

This idea took shape at the end of 2024. Judging from the data, the capabilities of the models, and people's expectations of our company, the team thought that if we wanted to achieve something big but didn't serve 1.4 billion people, it wouldn't make logical sense.

I joined Plaud on May 6th this year and have been preparing intensively for three or four months.

Intelligent Emergence: This year, DingTalk and Mobvoi have both launched AI meeting note - taking products similar to Plaud. Have you paid attention to this news?

Mo Zihao: Actually, I haven't even tried them.

But to be honest, we're quite different from them. Everyone is talking about helping users better summarize meetings and generate to - do lists, but we hope to help users save time and create more value.

Intelligent Emergence: Will the entry of big companies bring a sense of crisis to you?

Mo Zihao: I started working in the mobile Internet in 2013. I don't think these things are unexpected. If you have Product - Market Fit (PMF) or if your track has been proven successful, it's abnormal if there are no followers in the same track.

Intelligent Emergence: Are they your competitors?

Mo Zihao: I don't think so. Our concepts are different, the services we provide are different, and the resources we invest are also different.

Intelligent Emergence: What are the specific differences?

Mo Zihao: We hope to break through the boundaries of human intelligence, help you make better decisions, and provide you with intellectual value. This is completely different from "efficiency tools" or "work - helper gadgets".

We also don't want our product to play the role of an "assistant". We'd rather provide an equal "adviser" who understands you well and is very smart. You can discuss decision - making points with it.

So we've developed a different function called "Ask Plaud". Actually, it should be called "Plaud Ask". Others let you ask questions to AI, but we let Plaud ask you questions.

Compared with humans, if humans score 40, 60, or 80 points, Plaud provides a value equivalent to 40, 60, or 80 points, or even 42, 62, or 82 points. So our starting point is different from many large - model manufacturers.

Intelligent Emergence: This is the first time Plaud has entered the domestic market from overseas. Have you encountered any difficulties?

Mo Zihao: First of all, when providing services in the domestic market, we can't use overseas models, overseas payment methods, or overseas storage. We have to rebuild everything. So we need to make many service - related changes.

In the domestic market, we mainly use models such as Doubao 1.6 Pro, Qianwen's super - large version (Qwen3 - 235B - A22B), and DeepSeek.

Another issue is the more intense price competition. Many people may try to capture the market with prices that are half or one - third of Plaud's, but they don't provide particularly high - value services in some aspects.

Intelligent Emergence: Does the product need to be localized in the Chinese market?

Mo Zihao: The greatest opportunity for product developers in this era is that the general and truly user - valuable things have similar demands in different countries and cultures. This is brought about by the leverage of large models, which has provided additional intelligence and improvement.

We've already sold our products in 170 countries. We'll make local adaptations in terms of compliance according to privacy requirements, but the product design and functions remain the same.

The same goes for the Chinese market. Another very important task is to spread Plaud's concept.

Intelligent Emergence: What kind of concept are you spreading?

Mo Zihao: I think there's a lot of "noise" in the domestic market. Some people on the Internet say that Plaud is a "work - helper gadget". We never think of it that way. In fact, few people who sit in the office typing documents use Plaud.

You can understand that our scenarios are not limited to single - purpose meetings or office work. Instead, we emphasize work and conversation itself. So we need to spend more time and effort on brand communication.

There are many people in the world whose work doesn't involve writing documents. Conversation may be their main way of creating and transmitting value.

Many non - document - driven organizations naturally need other things to serve them because many things in daily conversations can't be structured. We've identified this market gap.

Intelligent Emergence: Which industries do this group of people mainly belong to?

Mo Zihao: Many doctors, lawyers, salespeople, consultants, and investors. These are the people who need to complete their work through a large number of conversations every day.

Plaud's barrier is the team

Intelligent Emergence: What do you think is Plaud's barrier?

Mo Zihao: There's only one, the team.

Actually, the determination and concept of big companies' investment depend on how much they believe in this matter. We have ten times the manpower and ten times the dedication of others. The products we create must be significantly different.

Intelligent Emergence: Can the product be replicated?

Mo Zihao: It may take others three months or half a year to replicate an identical function. During this time, we won't be idle, right?

Moreover, the product functions and stability we deliver at this stage can't be achieved by factories in Huaqiangbei. For example, many of our team members come from big companies. If the products we make can also be made in Huaqiangbei, then there's no point in having big companies.

Intelligent Emergence: Previously, Su Hao from the University of California said that it's very difficult to break the barrier between software and hardware teams. Is there any difficulty in the communication and collaboration between Plaud's software and hardware teams?

Mo Zihao: No, we have full trust in each other. We think the things they do are the best, and they think the same of us.

This trust is established starting from the management level. If the boss thinks this way, the employees will too.

Intelligent Emergence: What kind of founder do you think Xu Gao is? Which of his concepts will influence the company's culture?

Mo Zihao: He's a relatively simple person, and this is also a major starting point for the startup. Additionally, one of the values written in our company's creed is: Empower others.

Because both Xu Gao and I have had unsuccessful startup experiences before. No one wants to go through that pain again. No one can do everything alone, so we must empower others, and based on this, there will be trust among us.

Intelligent Emergence: What was your failure experience?

Mo Zihao: Maybe choice is more important than effort. The market you enter and the scale of this market are more important.

Also, the character of employees is very important. Raising the recruitment standards will make things simpler.

Intelligent Emergence: What was the opportunity for you to join Plaud?

Mo Zihao: Among our generation, only a very small number of people have truly built successful mobile Internet companies.

In 2013, after I was recommended to Peking University, I started my startup journey. But at that time, the Internet industry was on a downward trend, and we didn't catch the big wave of the era.

At that time, I worked on large - model projects in my previous - previous company for a year. The most important thing about this business was that the large model didn't change what I was selling. I was engaged in sales, and the large model only made the sales strategy better, but the product remained the same.

I thought this wasn't fundamental enough. So at that time, I wanted to see what the so - called "geniuses" were doing. So in January 2025, I joined Baichuan. At that time, Baichuan knew that there might be bottlenecks in model training, so we had to think about what else we could do.

I'm