EN / 中文

Imaging Hardware Enters Phase 3: AI + Robotics, Shifting from Clear Footage to Ready-to-Use Clips

by leifengwang·November 12, 2026

Author: Xu Zhi | Editor: Liu Wei

A food content creator wants to slowly push the camera from a wide shot of the dining table to a close-up of a bowl of noodles. This action is not difficult; what is hard is that he has no second person to help him push, nor is there space to set up a slider on the table. In the end, he switched to a fixed camera position. The image was clear, but it lacked a sense of breathing.

This table represents the real-world scenario for much of today's content creation. Product reviews, food videos, beauty tutorials, live streaming, and e-commerce sales—vast amounts of shooting take place within just a few square meters.

The equipment itself is already quite good: smartphones can shoot in 4K, pocket gimbals can stabilize the image, and camera quality has long exceeded most people's needs. However, once the lens needs to "move"—requiring pushes, pulls, orbits, or lifts—the standard gear in ordinary hands falls short.

Cameras are good enough, but camera movement remains a craft requiring years of training. What ZENITHIC Innovation is doing is turning this into an action executable by a machine.

Founder Zeng Cheng has experience working on film sets, and his master's research focused on robot trajectory optimization. The company was established in 2023, initially developing large-scale cinematography robots for film crews. Over time, they progressively reduced weight, shifting their technical roadmap to five-kilogram-class desktop products.

Supporting this transition is a precise understanding of constant loads in specialized scenarios: the camera, as the end-effector load, has a constant weight. After balancing, gravity is neutralized, so the motors do not need to lift the camera but rather move it. This led to the development of zero-gravity direct-drive technology, which serves as the foundation for product lightweighting. Combined with full degrees of freedom and AI director decision-making, this constitutes ZENITHIC Innovation's product path—the first generation serves professional creators, while the second moves toward the consumer market, potentially entering the price range occupied by DJI and Insta360.

The imaging industry is reaching a node where new answers are needed. The ceiling for parameter competition has appeared, and portable experiences have been pushed to near limits. The next phase of competition will no longer focus on "shooting clearly," but on helping users actually complete the content they want to shoot.

How does a thirty-person startup carve out a slice of the cake alongside giants? Why does ZENITHIC Innovation start with a desk when both DJI and Insta360 are pitching "director cameras"?

In the following dialogue, Zeng Cheng provides answers from perspectives including technical trade-offs, product definition, industry judgment, and competitive landscape.

1. From Film Set to Desktop: Why Did the Robot Need to "Downscale"?

Leiphone: How did ZENITHIC Innovation first enter the imaging field? What pain points did you see at the time?

Zeng Cheng: My master's thesis was on robot trajectory optimization. I also ran self-media channels and participated in behind-the-scenes work on film sets, allowing me to understand robotics technology while grasping the know-how and pain points behind lenses in different shooting scenarios.

At the time, I saw on film sets that renting a single robotic arm cost tens of thousands of yuan per day, and it required two engineers on-site for programming. Equipment had to be booked in advance, pre-programmed before shooting, and if the director changed a requirement on-site, it could take hours to adjust. Efficiency was low, and prices were high.

We then developed a product that primarily optimized the operating software, reducing a workflow that originally took 6 hours to 1 hour. But as we progressed, we realized the ceiling wasn't in technology, but in the scenario. The price and size of large equipment meant it could only serve a very small segment of people. However, the market capacity and the real incremental growth in content creation were not on the same scale.

Leiphone: Where is the real incremental growth in content creation?

Zeng Cheng: It lies with individual creators and small teams. Because the channels for video distribution have changed. Ten years ago, it was traditional converged media and TV ads; now, much of the distribution is completed by self-media.

The global active content creator base exceeds 200 million, with more than half of the shooting occurring indoors, specifically on desks—such as unboxing reviews, food, beauty, live streaming, and e-commerce sales. They have creative needs; their lenses need to move, and the images need a sense of breathing, but current equipment cannot adequately meet these needs. Their primary devices are sliders and stabilizers, which offer limited degrees of freedom and are time-consuming. A small portion of top-tier creators can use million-yuan-class large equipment, but most creators' workspaces simply cannot accommodate such devices.

So we completely re-engineered our technical roadmap, moving from single-degree-of-freedom to full degrees of freedom, and from hundreds-of-kilograms large equipment to desktop-level.

Leiphone: From hundreds of kilograms to desktop level, what specifically happened in between?

Zeng Cheng: The earliest prototype was an industrial robot plus a base, weighing 500 kg. Later, we switched to collaborative robots, reducing the weight to 120 kg. Then we used carbon fiber and smaller motors, bringing it down to 70 kg, then 40 kg, continuously reducing the weight.

When we reached 40 kg, I realized that following the weight-reduction logic of industrial robotic arms had hit a dead end. Its structural and load logic were designed for industrial scenarios like gripping and handling, not for mounting cameras for shooting. Without a breakthrough design, we couldn't go further; none of the metrics—performance, volume, or price—could reach where we truly wanted to go. So starting last year, we aggressively redesigned the entire technical roadmap.

Leiphone: Switching roadmaps and reducing weight means giving up something. What did you give up?

Zeng Cheng: We gave up empty-load scenarios. When you remove the camera, the robotic arm becomes unusable. Like gimbal logic, you need a camera to weigh it down first. We also sacrificed a certain degree of precision. The second-generation product's robotic arm has a precision of 5 wires (0.05 mm), which is half the width of a human hair—the error between two movements to the same spot is less than one hair. But shooting doesn't require such high precision; instead, it requires continuous motion. This is the trade-off we made.

Leiphone: How is the continuity required for shooting guaranteed?

Zeng Cheng: In the industry, six degrees of freedom are called full DOF. Six DOF allow for any position and orientation, but they introduce singularities. When the robotic arm moves near certain configurations, it may become unable to continue and must detour. Industrial scenarios emphasize the "result" of movement—it's okay to detour or pause—but shooting emphasizes the process; interruption during shooting is unacceptable. Therefore, we needed to solve the singularity problem.

We designed a special configuration combined with proprietary control algorithms to cleverly avoid singularity issues during camera movement, ensuring trajectories can be completed continuously.

2. From Parameters to Finished Footage: The Competitive Logic of Imaging Hardware Is Shifting Gears

Leiphone: The question of how the machine moves has been answered. Returning to imaging products themselves, what has been the iteration logic of this industry in recent years? Where will the next phase of competition occur?

Zeng Cheng: I divide imaging products into three stages. The first stage was the era of Canon and Nikon, where everyone competed on image quality; whoever had better parameters sold better. In the second stage, after parameters reached a certain threshold, increasing them further became imperceptible to users, so the industry started focusing on experience, such as pocket cameras—a category pioneered by DJI—which offered excellent portability and decent image quality. In the third stage, continuing to upgrade pocket camera parameters isn't as meaningful. So I believe the third stage uses AI plus robotics to ultimately enhance the user's experience of obtaining finished footage.

Leiphone: Why "finished footage experience"? What is lacking in current devices?

Zeng Cheng: Take videos shot by Yingshi Jufeng (Media Storm) using a Pocket camera as an example. If a novice uses the same device, they can't produce similar results. The fluidity, breathing room, and overall rhythm of the visuals cannot be captured by ordinary people. We studied the reasons deeply—professional creators like Yingshi Jufeng understand the three stages of video production (pre-production, production, and post-production) and can execute them at a high level.

Ordinary users often lack a complete creative workflow. There is no planning before shooting; they shoot whatever they see. During shooting, they don't know how to select focal lengths, compose, or move the camera. Finally, they have collected one or two hundred clips but can't edit them into anything. This is the key point—the emergence of AI plus robotics can solve this problem.

Current products on the market are either purely about shooting or purely about generation. Our focus is: placing AI capabilities at the highest-value point to assist users in shooting. AI handles part of the director's role—AI acts as the director, the robot as the cameraman, the user participates in the process, and the final result enjoyed is a high-quality finished piece.

Leiphone: How exactly does this "AI Director" work?

Zeng Cheng: Actually, the AI Director isn't the crux of the issue. The director's intent must be translated into execution to guarantee creative quality. AI can complete script creation, while the execution of the script is handled by the robot, and final editing is done via AI one-click compilation. That is roughly the workflow. Product capability should span the entire creative process—pre-production, production, and post-production—rather than being intelligent only in isolated local functions.

3. Selling First to the Most Picky Buyers: Gen 1 Anchors Mindshare, Gen 2 Defines the Category

Leiphone: What is ZENITHIC Innovation's product plan?

Zeng Cheng: Two generations of products, targeting different user groups.

The first generation is for professional creators, primarily serving the mid-stage execution phase. The true scarcity for these creators lies in motion control, not lens creativity. The second generation will be smaller, capable of serving as the main device for novice users, and incorporating cinematic styles into shooting.

The first generation can be understood as putting Hollywood on the creator's desk. The second generation puts cinematic language in your pocket.

Leiphone: What form will the second-generation product take?

Zeng Cheng: The second-generation product will be more portable and user-friendly, with lower requirements for professional expertise. Many technologies are still under validation and are currently classified as confidential internally.

Leiphone: Why target the professional track first instead of going directly for the consumer market?

Zeng Cheng: There are two reasons. The first is brand. The tone set by the first product determines where users position you. Letting the most professional and pickiest users adopt it first allows us to accumulate capabilities while establishing a brand anchor. Experience learned from professional users can also be passed on to consumer-grade users, enhancing experience and building trust.

The second reason is data. Professional creators know what makes a good shot, when to push, when to orbit, and which focal length matches which scene size. These judgments are real creative decisions, not something written on a spec sheet. While serving them, the first-generation product accumulates this decision data, which is used to train the intelligence capabilities of the second-generation product.

So the relationship between the two generations is clear: The first generation serves professional users, responsible for anchoring the brand and collecting capabilities; the second generation targets ordinary users, translating professional decisions into ready-to-use shooting plans.

Leiphone: What types of customers are actively reaching out now?

Zeng Cheng: There are several categories. One group includes BBC, Hollywood, and CCTV. Another includes self-media bloggers, MCNs, and internal e-commerce teams within companies.

Previously, the director of photography for CCTV-9 contacted us wanting to shoot a cultural relic documentary. Since the relics couldn't be moved, they had to shoot on-site. They drove two trucks into Tibet; their current supplier's equipment required four large air freight cases, taking up half a truck. Our device fits into a carry-on suitcase. He contacted us multiple times, but we were still in the early stages of technical validation and missed their project. He said once we released, we must tell him, and every team would equip one. BBC also emailed to inquire.

Another category is corporate brand teams. Any company with its own products and brands that wants to do e-commerce needs its own shooting department, often investing hundreds of thousands or millions in equipment. These users have sustained needs for creating content.

Leiphone: What unique technical accumulations do you have in motion control?

Zeng Cheng: ZENITHIC Innovation's motion control is developed specifically for the scenario, ensuring the user experience feels effortless. Motion control from previous industrial scenarios couldn't be used directly; editing a trajectory program took hours. We aim to shorten this process to minutes. To achieve this experience, both the body and motion control require dedicated design. Combined with user feedback after product launches, iterative optimization naturally leads to greater technical accumulation.

4. Hundred-Billion Market and Next-Gen Products: What Does the Imaging Industry Really Lack?

Leiphone: The second-generation product might compete directly with DJI and Insta360. How do you view this competition?

Zeng Cheng: The imaging product market is large enough to accommodate more companies. Just as Apple phones are excellent, they haven't eaten up all the mobile phone market share.

User needs are not a single point or a line, but a spatial volume composed of countless points. No single company can satisfy all demand points; everyone must make trade-offs. Discarded points don't mean they have no value, but rather that value must be concentrated on the retained points. The opportunity for startups lies in identifying value points different from those of large companies, making different trade-offs, and thoroughly addressing the retained demand points to create excellent products.

Leiphone: How large do you estimate the imaging market to be? What unsatisfied scenarios remain?

Zeng Cheng: The predicted market size for consumer-grade smart imaging hardware will not be less than CNY 100 billion.

The imaging market is vast; it occupies the two most important human senses: vision and hearing. Video includes sound and accepts the most information, coming in various forms. There are too many demands to explore, and many remain unsatisfied. For example, ball sports in motion scenarios, World Cup matches, tennis games—those perspectives are impossible for consumer-grade users to achieve, yet those are the most beautiful perspectives, capturing highlight moments.

Leiphone: Do you have an initial estimate of how many players have entered the imaging track? Who are the real competitors, and who are passersby?

Zeng Cheng: The imaging track is large enough to attract many talents and startup teams looking for new opportunities. First, startup teams need to find genuine user needs, deeply understand scenarios, define products well, and deliver good products and after-sales service. Teams that can do all these things are qualified competitors. As for who will be passersby, users will vote with their feet.

Leiphone: From the demand side, what exactly is missing in current imaging products?

Zeng Cheng: For professional users, what's missing is easy-to-use creative tools. For consumer-grade users, what's lacking is a complete creative workflow. ZENITHIC Innovation aims to start with professional scenarios to productize creative capabilities, providing users with an intelligent product featuring complete creative abilities.

Leiphone: In the more distant future, where will imaging go?

Zeng Cheng: Human creation of video content is essentially for expression. Whether through shooting or generation, any form that helps with expression will survive. Only imaging products that truly serve content production and help creators express themselves better will survive.

5. Afterword

Most progress in imaging hardware over the past decade or so has occurred on the "recording" end: higher image quality, smaller size, better stabilization.

Between "what I want to shoot" and "what I actually shot," there lies a whole set of capabilities that require long-term training to acquire.

An untrained person, even holding a sufficiently good camera, still doesn't know where the lens should start, where it should go, or where it should stop.

What ZENITHIC Innovation wants to do is deconstruct this capability, productize it, and break down professional judgment into steps that machines can execute and users can invoke.

If this path works, it will change not just shooting equipment, but the definition of "who is qualified to capture good images." Whether this path holds up will be answered by the market after the product launch next year.

But it at least points to a long-neglected problem: when the device itself is no longer scarce, what is scarce is the ability to make the device effective. Whoever can make this capability invokable may find their place in the next phase of competition.