Amanda Brown
Active member
As the AI wave sweeps across industries, television, as one of the earliest sectors to explore intelligence, is no exception.
As a typical mature saturated market, after several rounds of industry battles, the television sector seems to be experiencing a resurgence thanks to AI.
On one hand, giants like Google, Microsoft, and OpenAI have successively announced the integration of large models into TV platforms; on the other hand, at the AWE exhibition, domestic TV giants such as Hisense and TCL showcased comprehensive AI implementations from underlying chips and large models to terminal user experience... Even Apple was recently reported by media to be attempting to launch a new self-developed smart TV distinct from Apple TV.
Reviewing the changes in the TV industry over the years, from early years of hardware price wars to the internet TV era relying on software for revenue (advertising and content subscriptions), and now aiming to reshape core value and terminal experience through AI, can TV giants and industry players leverage AI to turn the tide this time?
1. The Ones Who Gained the First Wave of AI Dividends? Not the TV Giants.
Judging solely by the 2024 financial fundamentals, TV giants seem to have reaped the initial benefits of AI, and each company highlighted its AI achievements in its financial reports.
Taking Hisense and TCL as examples, after facing pressure on net profit attributable to the parent company in the first three quarters of 2024, Hisense Visual Technology achieved positive full-year growth, with TV revenue exceeding 46.6 billion yuan. Meanwhile, TCL Electronics' 2024 revenue reached 99.32 billion Hong Kong dollars, a year-on-year increase of 25.7%, and its net profit more than doubled year-on-year.
However, the recovery in the TV giants' performance was not primarily due to AI.
On one hand, this benefited from the TV giants' "premiumization + large screen" strategy in recent years. Moving from competing on low prices to competing on large screens, players in the 100-inch era actively initiated this consumption upgrade. They focused on bringing "100-inch giant screens" into millions of homes. By optimizing product structure and avoiding the old path of price wars, they drove up average selling prices, preventing a return to the difficult times of "falling volume and falling prices."
On the other hand, the relative underperformance of Japanese and Korean manufacturers, coupled with the technological upgrades and globalization efforts of domestic TV giants over the years, the increasing proportion of overseas revenue also somewhat alleviated the pressure of domestic market competition.
What is surprising is that, the ones who gained the first wave of AI dividends this time were the system service providers who previously relied on the TV giants.
Coocaa recently released "Super AI Agent" Image source: Company official
"(Coocaa's AI Agent) has already signed agreements with operator clients and car manufacturers in the first quarter, covering medium and large screens, such as operator set-top boxes," Coocaa Technology CEO Wang Zhiguo, a veteran of the TV industry, recently told "Silicon Base Lab".
As a significant player in the internet TV era, Coocaa, which recently announced its transformation into an "AI-native company," has collaborated with well-known enterprises in multiple fields such as operators, automotive, hospitality, and office, utilizing its self-developed Super AI Agent, and has launched new products like learning machines and "buddy screens" for the market.
"Silicon Base Lab" learned that unlike the early AI commercialization attempts in hardware like TVs, **there are two main reasons why system service providers like Coocaa achieved rapid commercial adoption in this AI wave:**
First, controllable operating costs and high computing power cost-effectiveness. Wang Zhiguo mentioned that Coocaa's Super AI Agent was developed based on FPGA architecture and has currently transitioned to a 32B smaller model.
Second, scene data advantage. Unlike downstream hardware manufacturers like TV companies, enterprises like Coocaa are positioned midstream, integrating hardware, systems, content, and services. Therefore, they have accumulated vast amounts of data over the past, including user data, device data, and also significant data accumulated in vertical scenarios such as audio-visual, education, and lifestyle.
More critically, AI has also brought new business models for companies like Coocaa.
In the internet TV era, smart system service providers largely followed the monetization logic of "befriending hardware manufacturers and doing business with software." However, as large models deepen software-hardware synergy, **companies like Coocaa are gradually shifting from "selling software" to "selling services."** For instance, Coocaa currently offers four commercialization paths: software sales, device licensing, PaaS services, and ecosystem co-prosperity.
Ultimately, the reason why "Coocaas" captured the first wave of AI dividends is simple: by doing business beyond the "TV screen" and deeply entering various industries to sell services powered by AI, their user base grew, and their own AI ecosystem was established.
2. AI TV: The Competition Remains Fierce
However, for traditional TV giants, AI TV is not just a gimmick.
The intelligence of televisions is not new, and AI has long been applied to enhance picture quality and other aspects. As early as around 2016, major brands had already launched smart TVs equipped with features like voice interaction and intelligent Q&A.
Wang Zhiguo also told "Silicon Base Lab" that, on the user side, the biggest problem facing the TV industry is actually the decline in daily active users (DAU). According to the "2024 China Smart TV Interaction New Trends Report" released by the Qianzhan Industry Research Institute, **since 2016, the TV power-on rate in China has plummeted from 70% to 30%, meaning fewer and fewer people like watching TV.**
This is also why TV giants are increasing their investment in AI TV: to enhance the TV large screen's power-on rate, viewing duration, and user activity by reconstructing the human-machine interaction experience through AI.
Currently, the TV giants' focus in AI TV development is primarily concentrated on three major areas: display, interaction, and interconnectivity.
The display domain is an "old battlefield" for AI implementation and represents the "lower limit" of AI TV competition.
Starting from 2018, TV manufacturers such as Samsung, Hisense, and Huawei began combining AI algorithms and display technologies to enhance TV picture quality and performance, which is also a major selling point for TV premiumization.
Samsung previously launched its latest 8K AI picture quality chip NQ8 AI Gen3, combined with QD-OLED technology, to further improve picture quality and audio clarity. Hisense also applied its new generation XinXing AI picture quality chip and LED backlight chip to its RGB-Mini LED TVs launched at CES, enabling features like AI analysis and recognition of display scenes, and AI adaptive backlight control.
The implementation of AI in the display domain is, to some extent, determined by the TV manufacturers' past technological trajectories and has not escaped the competitive logic of the "TV technology parameter war."
In contrast to the display domain, interaction and interconnectivity represent the "upper limit" of this wave of AI TV, as the competitive approaches of different manufacturers vary.
In the interaction domain, domestic TV manufacturers like Hisense and Skyworth exhibit a clear centralized scene thinking. Their typical approach is to launch AI agents based on their self-developed large models, providing intelligent assistant services for large-screen home demands such as education, gaming, and video search, making the AI agent the new center of TV interaction.
Samsung, however, takes a different approach, with a more apparent "decentralized" characteristic. Although Samsung has also launched AI agents, its emphasized Vision AI focuses on interconnectivity with its own IoT ecosystem and smart home services.
Neither path is inherently superior or inferior, as fundamentally, they have not yet fully moved beyond the "TV screen." The AI TV race has just begun, and ultimately, who can successfully tell the AI story will be decided by user adoption.
As a typical mature saturated market, after several rounds of industry battles, the television sector seems to be experiencing a resurgence thanks to AI.
On one hand, giants like Google, Microsoft, and OpenAI have successively announced the integration of large models into TV platforms; on the other hand, at the AWE exhibition, domestic TV giants such as Hisense and TCL showcased comprehensive AI implementations from underlying chips and large models to terminal user experience... Even Apple was recently reported by media to be attempting to launch a new self-developed smart TV distinct from Apple TV.
Reviewing the changes in the TV industry over the years, from early years of hardware price wars to the internet TV era relying on software for revenue (advertising and content subscriptions), and now aiming to reshape core value and terminal experience through AI, can TV giants and industry players leverage AI to turn the tide this time?
1. The Ones Who Gained the First Wave of AI Dividends? Not the TV Giants.
Judging solely by the 2024 financial fundamentals, TV giants seem to have reaped the initial benefits of AI, and each company highlighted its AI achievements in its financial reports.
Taking Hisense and TCL as examples, after facing pressure on net profit attributable to the parent company in the first three quarters of 2024, Hisense Visual Technology achieved positive full-year growth, with TV revenue exceeding 46.6 billion yuan. Meanwhile, TCL Electronics' 2024 revenue reached 99.32 billion Hong Kong dollars, a year-on-year increase of 25.7%, and its net profit more than doubled year-on-year.
However, the recovery in the TV giants' performance was not primarily due to AI.
On one hand, this benefited from the TV giants' "premiumization + large screen" strategy in recent years. Moving from competing on low prices to competing on large screens, players in the 100-inch era actively initiated this consumption upgrade. They focused on bringing "100-inch giant screens" into millions of homes. By optimizing product structure and avoiding the old path of price wars, they drove up average selling prices, preventing a return to the difficult times of "falling volume and falling prices."
On the other hand, the relative underperformance of Japanese and Korean manufacturers, coupled with the technological upgrades and globalization efforts of domestic TV giants over the years, the increasing proportion of overseas revenue also somewhat alleviated the pressure of domestic market competition.
What is surprising is that, the ones who gained the first wave of AI dividends this time were the system service providers who previously relied on the TV giants.
Coocaa recently released "Super AI Agent" Image source: Company official
"(Coocaa's AI Agent) has already signed agreements with operator clients and car manufacturers in the first quarter, covering medium and large screens, such as operator set-top boxes," Coocaa Technology CEO Wang Zhiguo, a veteran of the TV industry, recently told "Silicon Base Lab".
As a significant player in the internet TV era, Coocaa, which recently announced its transformation into an "AI-native company," has collaborated with well-known enterprises in multiple fields such as operators, automotive, hospitality, and office, utilizing its self-developed Super AI Agent, and has launched new products like learning machines and "buddy screens" for the market.
"Silicon Base Lab" learned that unlike the early AI commercialization attempts in hardware like TVs, **there are two main reasons why system service providers like Coocaa achieved rapid commercial adoption in this AI wave:**
First, controllable operating costs and high computing power cost-effectiveness. Wang Zhiguo mentioned that Coocaa's Super AI Agent was developed based on FPGA architecture and has currently transitioned to a 32B smaller model.
Second, scene data advantage. Unlike downstream hardware manufacturers like TV companies, enterprises like Coocaa are positioned midstream, integrating hardware, systems, content, and services. Therefore, they have accumulated vast amounts of data over the past, including user data, device data, and also significant data accumulated in vertical scenarios such as audio-visual, education, and lifestyle.
More critically, AI has also brought new business models for companies like Coocaa.
In the internet TV era, smart system service providers largely followed the monetization logic of "befriending hardware manufacturers and doing business with software." However, as large models deepen software-hardware synergy, **companies like Coocaa are gradually shifting from "selling software" to "selling services."** For instance, Coocaa currently offers four commercialization paths: software sales, device licensing, PaaS services, and ecosystem co-prosperity.
Ultimately, the reason why "Coocaas" captured the first wave of AI dividends is simple: by doing business beyond the "TV screen" and deeply entering various industries to sell services powered by AI, their user base grew, and their own AI ecosystem was established.
2. AI TV: The Competition Remains Fierce
However, for traditional TV giants, AI TV is not just a gimmick.
The intelligence of televisions is not new, and AI has long been applied to enhance picture quality and other aspects. As early as around 2016, major brands had already launched smart TVs equipped with features like voice interaction and intelligent Q&A.
Wang Zhiguo also told "Silicon Base Lab" that, on the user side, the biggest problem facing the TV industry is actually the decline in daily active users (DAU). According to the "2024 China Smart TV Interaction New Trends Report" released by the Qianzhan Industry Research Institute, **since 2016, the TV power-on rate in China has plummeted from 70% to 30%, meaning fewer and fewer people like watching TV.**
This is also why TV giants are increasing their investment in AI TV: to enhance the TV large screen's power-on rate, viewing duration, and user activity by reconstructing the human-machine interaction experience through AI.
Currently, the TV giants' focus in AI TV development is primarily concentrated on three major areas: display, interaction, and interconnectivity.
The display domain is an "old battlefield" for AI implementation and represents the "lower limit" of AI TV competition.
Starting from 2018, TV manufacturers such as Samsung, Hisense, and Huawei began combining AI algorithms and display technologies to enhance TV picture quality and performance, which is also a major selling point for TV premiumization.
Samsung previously launched its latest 8K AI picture quality chip NQ8 AI Gen3, combined with QD-OLED technology, to further improve picture quality and audio clarity. Hisense also applied its new generation XinXing AI picture quality chip and LED backlight chip to its RGB-Mini LED TVs launched at CES, enabling features like AI analysis and recognition of display scenes, and AI adaptive backlight control.
The implementation of AI in the display domain is, to some extent, determined by the TV manufacturers' past technological trajectories and has not escaped the competitive logic of the "TV technology parameter war."
In contrast to the display domain, interaction and interconnectivity represent the "upper limit" of this wave of AI TV, as the competitive approaches of different manufacturers vary.
In the interaction domain, domestic TV manufacturers like Hisense and Skyworth exhibit a clear centralized scene thinking. Their typical approach is to launch AI agents based on their self-developed large models, providing intelligent assistant services for large-screen home demands such as education, gaming, and video search, making the AI agent the new center of TV interaction.
Samsung, however, takes a different approach, with a more apparent "decentralized" characteristic. Although Samsung has also launched AI agents, its emphasized Vision AI focuses on interconnectivity with its own IoT ecosystem and smart home services.
Neither path is inherently superior or inferior, as fundamentally, they have not yet fully moved beyond the "TV screen." The AI TV race has just begun, and ultimately, who can successfully tell the AI story will be decided by user adoption.
