Is the companion robot completely out of the picture? It was a hit at CES but has now vanished without a trace.

As a leading indicator of the global consumer electronics industry, we can see the industry's "expectations" for the coming year at the annual CES. For example, at CES 2025 which ended earlier this year, the reporting team of Lei Technology found that almost the entire industry was focusing on AI smart glasses and AI companion robots. At that time, Lei Technology also provided relevant interpretations of these two product categories. In the blink of an eye, more than half a year has passed, and the two product categories that were highly regarded at that time are now showing completely different market development trends.

Image source: Lei Technology

With the launch of Xiaomi's AI smart glasses, the smart glasses category and the head - mounted AR devices it represents have witnessed significant development in China. Established players in the glasses industry such as Rokid, Thunderbird, Xingji Meizu, and xreal are constantly innovating and refreshing their product capabilities; new forces like Xiaomi are closely following, building their brand recognition with rich ecological applications. Except for the failed Shanjipai Paijing, almost every player can find their niche in the smart glasses market.

However, Lei Technology found that the situation of AI companion robots after half a year is relatively more complicated.

The "Disappeared" AI Companion Robots

Compared with smart glasses, the popularity of AI companion robots has cooled down very rapidly. At CES 2025 earlier this year, many brands, from well - known ones like TCL to lesser - known ones like Looi, exhibited companion robots of various sizes and forms, trying to seize the opportunity in this new market segment.

It should be noted that the companion robots discussed here are not functional robots like vacuum cleaners and lawn mowers that focus on "completing tasks", but emotional robots that emphasize interaction with humans and provide emotional comfort: they neither sweep the floor nor carry goods, but try to fill the emotional gap of people and provide users with "emotional value" through expressions, conversations, etc.

Image source: Lei Technology

As of now, Ropet, which made its debut at CES 2025, can be regarded as a leader in the field of emotional companion robots. Simply put, Ropet is an AI companion that sits on the desk and cannot move independently. Its main function is to chat with users and "amuse" them with expressions. Ropet even has a heating function, providing a better tactile experience when users "stroke" it. However, judging from the market response, even a leading brand like Ropet, which has achieved good results in the North American and Japanese markets, has not effectively penetrated the domestic market.

As for AI companion robots of other brands, they can even be described as having disappeared.

In Lei Technology's view, this "disappearance" is not accidental. On the one hand, the hardware form of AI companion robots is complex, integrating multiple sensors and modules for motion, visual recognition, voice interaction, etc. The cost is much higher than existing products such as smart speakers. In scenarios that mainly focus on voice and dialogue interaction, users are more inclined to use lower - priced and more integrated AI mobile phone assistants or smart speakers to meet their basic needs.

Image source: Xiaomi

On the other hand, the biggest selling point of AI companion robots is "emotional value", but this is also the most difficult function to scale up and the most likely to cause controversy. Even though the industry has evolved from smart speakers to companion robots, the "human - like interaction" of the latter is still in its infancy. They are often just content - playing devices with mimetic expressions and voices, and it is difficult to build a real emotional relationship. For most consumers, the motivation to pay thousands or even tens of thousands of yuan for such robots is quite limited.

Why Is "Emotional Value" No Longer Popular?

More precisely, the fundamental reason for the cold reception of emotional companion robots is that emotional value has not yet become a widely recognized consumer category in society.

Most of the past consumer electronics hits were based on two types of logics: either solving efficiency problems, such as the supplementary role of smart watches in health monitoring and the value of vacuum cleaners in liberating hands; or meeting "quantifiable" living needs, such as clearer photos, larger screens, and longer battery life. However, positioning "loneliness" as a selling point for companion robots seems rather awkward in the context of mass consumption.

Image source: Lei Technology

Especially in the Chinese market, this kind of consumption behavior of "paying for emotions" is still regarded as a symbol of "impulse consumption" and "consumerism" and has not truly become mainstream, and may even trigger social aversion. Users are willing to use services like Kimi to relieve stress and practice oral English because Kimi and other AI services solve "quantifiable" problems. However, when physical hardware with "emotional support" as a selling point appears on the desk, consumers often become extremely resistant.

Emotional companion robots are neither as "dispensable" as toys nor as clearly positioned as white goods. This in - between positioning directly leads to a contradictory state among potential users of companion robots: they are interested but dare not buy, or they buy but rarely use.

At the same time, some manufacturers' understanding of emotional value remains at a very shallow level: they copy the designs of JellyCat or Wall - E, pair them with local AI models or directly connect to the ChatGPT API in the cloud, and give them a nice name, thinking that a companion robot is completed.

Image source: Lei Technology

But in fact, these are far from providing emotional support. In Lei Technology's view, a device with real emotional projection ability needs to learn users' behaviors and preferences over a long period and be able to actively provide understanding and comfort at critical moments - which is far beyond the value that current smart hardware focusing on "one - time sales" can provide.

After all, emotional companionship is essentially a "long - term service", but the current product logic is still that of "smart hardware" - most companion robots adopt a one - time sales model, without a content subscription mechanism or the ability to continuously train and update. Once the hardware capabilities become outdated and the models are no longer updated, these "smart companion hardware" will soon be forgotten by users, and their usage frequency may even be lower than that of the AI assistant on the phone.

From this perspective, companion robots seem to have chosen the wrong path "from the very beginning".

Companion Robots Must Learn to Be "Proactive"

Of course, from an industry perspective, when I say that companion robots "chose the wrong path from the beginning", I am not denying their significance. But before brands rush into the market, I hope they can truly understand the real meaning of "emotional companionship" in today's individualized society, rather than simply launching a "talking ornament".

In Lei Technology's view, if brands want to truly make an impact in the "AI companionship" field and create products that truly attract users, they must accelerate the transformation of companion robots from passive to proactive.

Most companion robots on the market still require users to initiate commands actively. The only difference is that the "wake - up words" have changed from "Xiaomi AI" or "Hey Siri" to "Hello", "I'm so bored", or "Talk to me". However, real emotional companionship should be that the robot actively senses whether you are in an abnormal emotional state or have been silent for too long and decides when to start an interaction. Behind this active emotional support are the support of sensing hardware such as microphones and cameras and algorithms such as emotion recognition, semantic analysis, and situational emotion perception.

Judging from this standard, products that truly have the opportunity to succeed in the field of smart companionship should at least achieve the ability to be "readily available and proactively appear". Even just nodding when the user is changing shoes before going out or reminding the user to rest early when working overtime late are ways to actively establish emotional interaction with users. Or to put it more directly, an excellent smart companion product must "proactively do the right thing in the right situation" rather than just being a "responsive" smart speaker.

From a psychological perspective, emotional companionship is not a false proposition. As social beings, humans' need for emotional value will always exist. However, if products for interest companionship remain at the "hardware" level rather than the service level, these products are doomed not to go far.

This article is from the WeChat official account "Intelligent Pro", written by Tianxing, and is reprinted by 36Kr with authorization.