After Leaving DJI and Yunjing, He Transforms Traditional Hardware Products with AI and Raises Over 10 Million in Crowdfunding | Exclusive Interview by Hardcore Kr

Author | Zhang Ziyi

Editor | Yuan Silai

In this year's primary market, Zettlab is definitely a "unique" project. It has the qualities that investors love the most right now: the team members come from DJI, it focuses on vertical AI NAS hardware, and its technology has barriers. The key is that it also achieved good crowdfunding results.

However, the mood of the founder, Guo Yanan, no longer fluctuates greatly with praise or doubts.

In 2023, when the financing situation in the primary market was at its worst, Guo Yanan quit his job at Yunjing. Almost everyone showed confusion, but he believed that "everyone is on the same starting line, all swimming naked. (At this time) the opportunities must be the best."

Guo Yanan dived into the rather niche and traditional category of NAS. NAS is short for Network Attached Storage. Simply put, it is a private cloud. Its basic function is storage. It can be connected to the Internet and has certain computing capabilities.

In the entire market, the demand from enterprise - level users accounts for about 60% of the total NAS market size. Manufacturers in Taiwan, China have a relatively stable market position. The remaining 40% of the market is mainly contributed by C - end consumers and small and medium - sized studios. The market is relatively fragmented and is divided up by enterprise - level NAS manufacturers and hard - disk manufacturers.

However, NAS has never been associated with AI. When the team conducted user research, NAS users thought they didn't need AI. When advertising in the overseas market, some people would scold them: "You've said a bunch of fancy words, but it's actually useless." The negative feedback and doubts were quite intense, and almost no one could see the prospects of AI NAS.

Guo Yanan firmly believes in his intuition as a technical engineer. As early as 2014, he started to get in touch with AI. He debugged autonomous driving algorithms in the freezing Mohe River and built his first driverless car during his student days. From DJI to Yunjing, Guo Yanan has been in charge of AI - related businesses and has witnessed the increasing functionality of AI. In his view, it is against common sense for NAS not to use AI.

As a private data collection point for users, NAS is one of the best places to host local AI Agents. New - generation AI products will lead to an ever - increasing amount of data. How to use data will be a challenge in the future. In the concept of Zettlab's products, AI can process the huge amount of video data on the edge side: search, organize, interpret, and even edit.

The market has indeed proven the correctness of his judgment. On the crowdfunding platform Kickstarter, Zettlab received more than $1.4 million in support, and 80% of users bought its products because of AI. Zettlab also recently completed its third - round financing, raising tens of millions of yuan. The investment institutions are Xinglian Capital, Vertex Ventures, and Fengqiao Capital.

We reviewed the entrepreneurial story of Guo Yanan, the founder and CEO of Zettlab, and also discussed the judgment of the AI NAS market. The following is the transcript of the interview, edited for content:

01 Two Principles for Finding a Track

Yingke: The NAS track is quite niche. Why did you consider entering it at that time?

Guo Yanan: In the past decade, AI has been "fragmented." The locks, cameras, and TVs in your home all have AI. A small chip costing just a few dollars can achieve some functions, but what they can do is very limited.

The current generation of AI is different. To truly function, it requires a "central node" with computing power, bandwidth, and storage. So we saw the opportunity here - NAS is naturally a small - scale data hub with the ability to support the aggregation of these AI capabilities.

It can be understood that from the previously "scattered" intelligent systems working independently, we are evolving towards an "integrated" intelligent system where everything works in coordination. This is exactly the reason why we decided to enter the NAS track.

NAS itself is niche and focused on storage. We hope to do a good job in storage first and then add AI and computing capabilities. This group of users pays more attention to local storage and privacy, so it's easier for us to attract them.

Yingke: Is your entrepreneurial direction also related to your background? Could you tell us about your background?

Guo Yanan: I graduated from Beijing Institute of Technology, majoring in mechanical engineering both as an undergraduate and a postgraduate. Later, I joined DJI in 2016 and started working on AI there. Actually, I was gradually transitioning before joining DJI. Around 2014, there were some autonomous driving competitions both overseas and in China. At that time, I built the first driverless car among Chinese students.

It was very difficult to work on AI at that time. To run a detection, recognition, or obstacle - avoidance model, you had to carry a server in the car. We were in Mohe, where the temperature was minus thirty or forty degrees Celsius. The computer would run out of power quickly if used and would shut down if not used because it was not designed for such an environment. But now, you can see that autonomous driving algorithms can run on a chip as small as a fingernail.

Later, when I joined DJI to work on consumer electronics, I found that AI functions and algorithms could be integrated into very small devices like mobile phones and drones. Around 2021, I went to Yunjing, where I was in charge of the AI team and some products. Based on these past experiences, I think there are great opportunities at the device end (edge side). I also have the ability to implement cross - border projects like "AI + hardware + storage."

Yingke: A good team is also very important for hardware development. What's the situation of your internal team?

Guo Yanan: The members of our team come from DJI, Huawei, ByteDance, and Li Auto. They have experience in making drones and sweeping robots, writing algorithms, building system architectures, and also in integrated marketing and international markets. They are a good fit for the hardware + AI combination.

Moreover, there is a common understanding among us - we like technologies and products that can be truly implemented and felt by users.

Yingke: However, starting a business in 2023 seemed to be in a tough environment, right?

Guo Yanan: 2023 was the worst year for the investment environment. Everyone advised you not to start a business at that time. What's the point of starting a business in such a situation?

Yingke: Then why did you still decide to start a business and choose such a field?

Guo Yanan: Because OpenAI fired a shot, telling everyone that a new era has begun. Everyone is on the same starting line. Whether it's a large company or a small company, they have no experience. They are all swimming naked. Starting a business in such a state, the opportunities must be the best. In such a stage, new species have more chances to emerge.

Later, we summarized two principles: First, there should be short - term benefits and long - term exponential growth; Second, the technology should have barriers, and we should have an advantage.

If you filter according to these two principles, there aren't many things you can do. We tried some other categories but finally gave them up - either the competition was too fierce, or the growth logic was not valid.

02 Everyone Said "No Need," but I Became More Certain

Yingke: Then you settled on NAS?

Guo Yanan: NAS is a direction that's relatively easy to figure out. First of all, data will definitely become more and more important - all devices will eventually converge to a "data hub." NAS naturally has computing power, bandwidth, and storage. It is an ideal carrier for local AI.

Yingke: How did you discover the NAS track at that time?

Guo Yanan: I talked to many friends, and quite a few of them recommended NAS. I've also been a NAS user for more than a decade, having used products like ZSpace, QNAP, and Synology. Another key point is that there are some NAS users in DJI's user group, so I could quickly find a group of users for interviews and analyze their pain points.

At that time, there was a very counter - intuitive phenomenon: When you asked NAS users if they needed AI, everyone told you no.

Yingke: I was also very shocked when I first heard about the AI function in NAS.

Guo Yanan: I was looking at discussions in the Reddit community at that time. Almost everyone said that "NAS doesn't need AI." Even many long - time NAS users were naturally resistant to AI, thinking it was just a marketing gimmick.

But from a common - sense perspective, if you can't use the increasing amount of data, then the data is just garbage and has no value. AI can help you use and manage data better, and even help you generate data better. This must be valid from a human - nature perspective. It's just that there was too much negative information at that time.

Yingke: I think video creators are likely to be the first group to feel the pain of "storage." Could you give an example to elaborate?

Guo Yanan: For example, a video creator often doesn't shoot according to a timeline.

They might shoot at home in the morning, go outdoors in the afternoon, and then return home at night. There are a lot of scattered materials. AI can help them understand and classify the materials again, align videos, pictures, and texts automatically, greatly improving post - production efficiency.

Also, many times creators need to find "previously shot footage." In the past, they had to rummage through hard drives and material libraries. Now, they can find what they need with a natural - language search. This difference in experience is revolutionary.

Yingke: So, does Zettlab use AI for video search? Can't previous NAS products do this?

Guo Yanan: Let me give you an example. If you're a sports event photographer and you've shot a two - hour game, and you need to immediately create a short video to find the exciting moments after you get back. How do you find them? You have to watch the two - hour video. But with a natural - language search, it can help you edit the video in less than a minute. This is a huge improvement in efficiency.

Yingke: Overseas users attach great importance to privacy. How do you balance AI capabilities and privacy protection?

Guo Yanan: So we chose an edge - side AI architecture. All core functions can run independently locally without relying on the cloud.

Connecting to the Internet is just for the convenience of remote access. For example, users may want to retrieve materials when they are on a business trip.

However, all searches, recognitions, and intelligent processing are completed locally on the device. Data synchronization between devices is also through P2P direct connection without going through a third - party server. In this way, the capabilities of AI and privacy protection are no longer in conflict but can coexist.

For most users, their data only exists in their own devices, and the security is controllable and trustworthy.