The color TV industry is currently in a period of deep adjustment: industry sales continue to decline, TV activation rates are falling, and corporate earnings reports reflect the severity of the situation. Where is the key to breaking through? Recently, an interview was conducted with Wang Zhiguo, CTO of Skyworth Group and CEO of Coocaa Technology. During the half-hour conversation, Wang Zhiguo presented the "three keys" to solving the current downturn in the TV industry—AI, picture quality, and sound. In his view, if the TV industry wants to sustain growth from its current sales volume of approximately 32.89 million units and rise again, AI-driven development is the critical factor. (Data source: Lotu "China TV Market Brand Shipment Report," published January 2026)
During the conversation, he also shared his thoughts on the TV industry's "membership nesting" problem and corporate organizational reform. The following is a transcript of the dialogue (slightly abridged):
"Three Keys: AI, Picture Quality, and Sound Quality—Finding a Path to Break Through for the TV Industry"
Question: Given the overall decline in color TV sales and falling activation rates, how do you view the current situation? What methods are there to break through?
Wang Zhiguo: The color TV industry currently has several characteristics. First, there is concentration toward the top players—although overall sales are declining, the top few companies are still seeing sales growth. Second, all companies are looking for ways to transform—TV hardware is getting thinner, larger, and clearer, driven by more cost-effective screens and technological iteration to promote upgrades. This is the path that has been continuously pursued. Actually, AI is now an even more important factor. If the color TV market is to reverse course and truly rise again, it must be through improving TV usage frequency and user usage stability—these are the biggest keys to TV transformation. In this process, there is an unavoidable question: at home, can elderly people use the TV conveniently? Can children use it healthily? How do young people use it? Using AI to truly understand users and serve users is our most critical exploration right now—lowering the barrier to use and reducing operational complexity. There have been many knots that couldn't be untied in the past—because of copyright, membership fees must be charged. These knots should have breakthrough solutions in the AI era, and we are currently discussing this with upstream content partners. The core question is: how to use the big TV screen to serve the family again and make the family feel better. From what we can see now, exploration has reached deep waters, and we are at the doorstep—some positive changes at the data level are already visible. But there are still some key challenges to break through—including the original business model, content and interaction methods, and the completeness of the entire AI ecosystem.
AI is expected to become an important variable in solving the existing pain points of the TV industry and driving it toward a new stage of development.
Question: AI is one key, and the industry also considers picture quality to be another key. Skyworth is also making chips, but investment in this area is relatively high, and everyone is competing. Does the input-output ratio work out?
Wang Zhiguo: Skyworth's AI chip is an AI picture quality chip, and I am the one driving it. Why invest so heavily and still do this? First, the real advantage of TV is giving everyone a reason to watch TV instead of looking at their phones—you must be immersed in its audio and visuals, otherwise why would I use such a big TV? As screens get bigger, content clarity can't keep up with the changes in screens. How do we make content that isn't very clear, or old films that many people still watch, look clearer on larger screens, without degrading the experience just because the screen is large? These are the things we are working on. Second, high-end TVs in the industry place great importance on picture quality chip capabilities—chips play a key role in color, transparency, and overall visual presentation. In high-end retail settings, users directly compare the picture quality of multiple high-end TVs. If our products cannot deliver sufficient visual experience in color saturation, contrast, and screen transparency, it will be very difficult to win over high-end users. This is also an important reason why we insist on developing our own AI picture quality chips.
Question: Have we achieved this?
Wang Zhiguo: This is the goal we are striving to achieve. Currently, Skyworth's related product sales have achieved growth, which is the result of combined capabilities in picture quality, sound quality, industrial design, and other aspects.
Question: Picture quality can be called the second key. Is the third key audio?
Wang Zhiguo: You could say that. Our requirement is that when people watch TV, listen to music, or watch films, they should have an immersive and emotionally engaging auditory experience. In the market, some audio products are priced higher than TVs themselves—this in itself demonstrates the importance of sound, and shows that sound experience has strong penetration and emotional impact for users. Vision and hearing are very important sensory experiences for humans—vision is picture quality, and the big screen is an important advantage of TV, so we put vision first. From the perspective of user product experience habits, hearing is one of the key determining factors in whether users choose to use the TV to watch films. But the TV's function is not just for watching films—when the TV screen is off, if the audio is good enough, the home effectively has both a TV and a good sound system for the price of one TV. So the penetration of high, mid, and low frequencies in audio is very important, making the TV product experience more comprehensive. Skyworth has spent a lot of time and R&D resources on tuning audio products. As TVs get thinner, the sound cavity and many other things are affected, so Skyworth's high-end wallpaper TV has a separate box underneath that houses the audio, providing panoramic sound and various high, mid, and low frequencies—making the audio effect very expressive even when used independently.
Question: If the three keys were to be ranked, how would you rank them?
Wang Zhiguo: If we are solving the problem of current market competitiveness, picture quality comes first—what you see at first glance is more important, and everyone definitely notices picture quality first when they go to a store. Next is sound—with such a big screen, it's primarily for watching, and secondarily for listening. Finally, AI. From the current market perspective, no one yet believes AI is a major factor in choosing a TV, because what everyone thinks of first when buying one is still watching TV. But from the perspective of the TV's future development—if we hope the TV industry can break through its current scale bottleneck and further unlock market space, AI will be a very important driving factor. With AI, beyond picture quality and sound quality, users will also care about whether the TV is equipped with a high-definition camera, memory size—just like the hardware and software of a phone, the TV experience can be continuously updated and upgraded, and users' pursuit of better product experiences will become a reason for them to upgrade their TVs. So for the TV industry to break through, AI will be a very important driving direction, and we are all working toward this direction.
"Breaking the Membership Nesting Problem: Using AI to Optimize the Family Big-Screen Content Consumption Experience"
Question: Everyone says the smarter the TV, the worse the "membership nesting" problem, and the less elderly people know how to use it.
Wang Zhiguo: First, for a period of time in the past, everyone just wanted to push more content to users, which led to overlapping memberships—"so-called 'membership nesting' refers to users being forced to stack layer upon layer of nested memberships in order to watch different content." Now with AI, we can identify what kind of content users like and push the most suitable membership and the most suitable content to them, so that users get real value for their money. After pushing it to them, we then deeply explore that membership's content and continuously give them good content to watch, so they can fully enjoy the membership—if the content they watch every day is enjoyable, won't everyone feel the membership was worth buying? So the first thing is to reduce unnecessary memberships, not let users buy memberships for no reason, first explore user habits, and push the most suitable options based on what they like. Second, after users buy a membership, we need to use AI to serve them better. This process is the change we hope users can perceive.
Reporter: So AI as the key to turning on the TV is mainly about improving the membership model? Letting users get the most out of every membership?
Wang Zhiguo: Actually, that's just a small point, because it only solves the problem of "watching TV." The real big direction is for the TV to become about "using the TV." There are many needs in the home. Although phones are already very convenient—you can just take out your phone and ask an AI tool—but many times, for example, I just want to know what major events are happening right now. Can the TV stay on standby 24 hours a day, like my TV butler, so that when I habitually ask, it can immediately answer? So AI has tremendous value in improving TV interaction. It's just that how these scenarios become user habits and how they truly meet user needs still requires a lot of testing. We have launched many features and found that many users use them, but stickiness is weak—they use it once and then stop—which shows we haven't done well enough. So we improve and iterate, look at user feedback, until we make it sticky. There is also an "intermediate transition" problem: take businesspeople as an example—businesspeople often need to check and book flights, but how often and for how long do businesspeople use TV scenarios like checking flights and booking tickets? Just like the transition from button phones to touchscreen phones back then, not everyone can immediately get used to it—there is a habit-forming process. And this process precisely requires us to comprehensively improve the AI experience, ultimately making it truly usable for everyone.
Reporter: After using AI, has Skyworth's TV sales or activation rate improved?
Wang Zhiguo: Skyworth's sales are now rising, and the active rate is also rising in sync because of AI. In pilot tests of AI audio-visual search and personalized content recommendations conducted on some Skyworth TV models, product active rates have improved significantly. In some test samples, the active rate increased by more than 10%, with the pilot peak reaching up to 20%. We are still continuously iterating and optimizing.
"From Passive Intelligence to Proactive Service, Redefining the Next Generation of Home Appliances"
Question: Looking ahead three years, what major home appliance product innovations with practical applicability under AI empowerment do you predict? AI TVs, air conditioners, or home robots?
Wang Zhiguo: All products are possible, but the biggest marker is proactive intelligence. In the past, everyone talked about smart home appliances, but are these appliances really smart? Being controllable by phone or voice doesn't make something smart—it still has a usage barrier. So I believe the most important landing point for the next step in smart home appliance innovation is whether your appliances can understand you and provide proactive services. We shouldn't be learning to use appliances; rather, appliances should proactively adapt to us and serve us better. Ultimately, what will truly enable all home appliances to rise again is the degree of intelligence and proactive service capability of the appliances. The form factor doesn't need to be debated—it could be a TV, an air conditioner, a robot vacuum, or a movable companion screen. Every form has its value and its own role to play. Drawing an analogy to the L1-L5 classification of autonomous driving, I personally divide the evolution of home appliance intelligence into several stages: simply achieving networked control can be seen as the initial stage; being able to automatically adjust temperature and lighting based on algorithms is a step further; being able to understand user voice intent and complete automatic adjustments is a higher stage; integrating large models to achieve capability generalization is the next stage; the ultimate ideal form is when appliances can proactively perceive and proactively provide services to users. Of course, this is just my general understanding of industry evolution and does not equate to the current national standard classification definitions for smart home appliances.
Question: So what level are home appliances at now?
Wang Zhiguo: I estimate between L3 and L4—rapidly integrating large models, but large models haven't been used particularly maturely yet.
Question: Many home appliance companies are now pushing home robots. Does Skyworth have any layout in robotics?
Wang Zhiguo: Skyworth's manufacturing system already uses many robots—robotic arms, transport robots, and so on. But my view on (household) robots is this: everyone may hope to have a robot to accompany them and help them do things, but we view this matter relatively cautiously and will first think about where the necessity lies. For example, if we make firefighting robots that can help more firefighters avoid danger, I think that is valuable and definitely a major direction. Take cooking robots as an example—from the current stage of home scenarios, I personally hold a cautious attitude toward the pace of large-scale adoption of such products. Cooking itself carries family emotions and life's pleasures; even if users don't want to cook, food delivery and group meals already provide mature solutions. Of course, this is just a judgment based on the present and does not deny the possibility of development for household cooking robots after future technological iteration. Home robots may arrive someday, but there is still a long way to go. So we will cautiously reduce related investment in these areas. Again, it's about value—what social pain points can I solve, what user pain points can I solve? Starting from the most painful pain points is the only way to achieve rapid improvement and a sound input-output process.
"AI Restructures Enterprise Organization and Management, Returning to User Value"
Question: The interim reports just disclosed by various companies are quite grim. How do you view the current situation? How can everyone break through?
Wang Zhiguo: Overall, leading companies still face a grim situation, whether in terms of growth or competition. But I think the essential problem is not in ideas and approaches. After AI arrived, I talked with many entrepreneurs, and everyone's ideas and approaches are correct: in a market characterized by intense competition, everyone knows they need to create new value for users, hope to improve and optimize the efficiency of the path to deliver value to users, hope to connect with users as quickly as possible rather than the traditional model of being done once sold and getting no user feedback. Everyone's thinking actually returns to the values of the early startup period. The real problem is in enterprise management—our professional management system is KPI-oriented. Everyone's value creation is aimed at rapidly increasing sales. So can our management system be restructured on the basis of AI? How to make enterprises more efficient? I believe this is the real problem enterprises face. If enterprises can truly use AI for management and use AI to enhance their management transformation capabilities, what changes will it bring? In the past, not many management transformations succeeded. Leading companies are still doing well today, and their transformations have been relatively successful. But many companies—everyone works very hard, yet transformation isn't considered successful, because they haven't truly reorganized their value chain. What changes does AI capability bring? AI can break through the traditional layer-by-layer transmission in enterprises from strategy to execution. Why is execution poor now? When enterprise decision-makers have strategic ideas, the information is transmitted to different organizations within the company for research and gets distorted layer by layer, and by the time it reaches execution, it becomes "this is a task assigned by the boss." No one thinks about what value what they're doing has for users. AI can directly penetrate management to every individual: using AI's vast knowledge base, integrating strategy with AI and rapidly reaching every position; using AI to determine what each position should do; using AI to rapidly improve everyone's capabilities and guide everyone to use AI to be more efficient and do things well; using AI to clarify whether what everyone is doing is correct. So we often say: do the right things, do things right, and do things to perfection. These three things can no longer rely on traditional layer-by-layer organizational transmission—AI can directly penetrate thinking to every individual and enable good division of labor and layout.
Question: Is Skyworth also making changes in this area? Will you provide services externally?
Wang Zhiguo: On September 1, Coocaa fully launched its self-developed AI-native enterprise operating system, and we also provide supporting AI-native enterprise organization and management solutions externally. What Skyworth is doing aligns with the state of most traditional enterprises that want to transform—we understand the difficulties involved: you can't change everything all at once, nor can you wait to change slowly. You need to find a path that allows the organization to transition smoothly. So we created something called happy work, an AI OS, through first practicing internally, continuously refining the product, and empowering enterprise management—but in fact, we are already empowering external parties now.
Comments