Why in-vehicle physical robots are erupting in a concentrated wave in 2026

[[**In 2026, in-car physical robots will erupt in a concentrated wave, with Xiaomi Smart Display 2, Geely Eva, Roewe Doubao Screen and other products each taking different paths. Their essence is to use a perceptible persona to establish differentiation amid the homogenization of smart cockpits, aligning with the rising demand for emotional value, but their long-term value depends on whether they can truly integrate into the interaction ecosystem.**

**Key Points**

**1. The driving force is cockpit homogenization, not AI maturity**

Smart cockpits are increasingly similar in screens, chips and voice; physical robots turn invisible software capabilities into a physical image users remember, becoming a vehicle for differentiation; but they are not necessarily smarter than voice assistants—currently they are mostly preset animations plus lightweight interaction.

**2. Consumer demand for emotional value has already emerged**

The young cohort willing to pay for Pop Mart-style emotional value has entered the auto market, and small-ticket "companionship consumption" is heating up amid the economic downturn; during the 2026 Spring Festival travel rush, several in-car AI robots priced at around 200 yuan on Taobao reached "10,000+" in sales, showing that real demand already exists in the aftermarket.

**3. Three commercial paths coexist**

Xiaomi Smart Display 2 takes the route of an in-car/out-of-car peripheral and ecosystem extension, but there are no landed cases of deep smart-home integration yet; Geely Eva validates demand with standalone hardware at 1,099 yuan, compatible with some models; Roewe Doubao Screen incorporates the cockpit assistant into the whole-vehicle solution, with higher pre-installation integration responsibility and validation costs.

**4. Product design needs to be differentiated by model scenario**

On the Geely Xingrui, it can be a "snarky-friend-type co-pilot," highlighting personality and fun; on the Lynk & Co 900, it is better suited as a family companion, carrying on children's habit of using smart hardware at home while giving parents control over content and duration.

**5. Short term is about speed; long term is about making users unable to do without it**

This article is from the WeChat public account: Chengzhu Dongjian, author: Zhu Dawei

This year, in-car physical robots have suddenly become hot.

Xiaomi Smart Display 2, Roewe Doubao Screen, Geely Eva, and various aftermarket AI pets... all of a sudden, it has become trendy to place in a car a little creature that can turn its head, blink, and make expressions.

But NIO's NOMI was released as early as 2017, and after years of iteration it has become a highly recognizable signature design for the brand. China's auto industry has never lacked followers, so why did this category wait nearly nine years before suddenly ushering in a concentrated wave?

I once heard a three-question framework said to be used by Alibaba's Daniel Zhang when reviewing new project proposals: What is it, why, and why us?

What is it? What is the essence of this proposal, and what capability or service does it provide?

Why? Why is it needed? Who does it serve? Why now?

Why us? Why are we able to do it well? Why can we beat the competition?

These three questions seem simple, yet they can break a bustling market phenomenon down into three levels: product value, user demand and commercial competition.

Today, I will try to use this framework to see why in-car physical robots suddenly became hot in 2026, and how far this wave can really go.

The most respectable explanation is: large models have come onboard, so physical interaction naturally follows.

But the real driving force is probably not that romantic.

Over the past few years, competition in smart cockpits has revolved around screen size, chip computing power, UI design and voice response speed. Parameters are still growing and features are still increasing, but for consumers, cockpit experiences across different brands are becoming more and more similar.

When everyone uses similar large models, similar interaction paradigms, and even offers similar features, where exactly can differentiation between products still be established?

Physical robots offer a new vehicle for differentiation: they turn originally invisible software capabilities into a physical image that users can see, remember, and even be willing to form an emotional connection with.

A larger central control screen can hardly make people remember a brand by appearance alone; but a little robot that turns its head, blinks, and can react to your behavior may become the user's most intuitive memory of this car.

It is not necessarily smarter than a voice assistant, but it has a chance to be more present than a voice assistant.

In September 2026, Xiaomi launched Smart Display 2, featuring a full-screen design and supporting head-turning and multiple expressions; Geely Eva entered the market as standalone aftermarket hardware; Roewe Jiayue 07 further incorporated the Doubao Screen and Doubao cockpit assistant into the whole-vehicle solution.

Three different product paths point in the same direction: making smart interaction no longer happen only in screens and sound, but also on an image that can be seen.

Of course, a physical robot does not equal a true AI Agent.

A piece of hardware that can make expressions and turn its head may still be backed only by preset animations and simple voice-trigger logic. How anthropomorphic the appearance is and how strong the intelligence is are not the same thing.

A true smart cockpit Agent needs to understand user intent, invoke vehicle capabilities, coordinate different systems, and take action at the right time. The large model is only one part of it; open vehicle interfaces, system permissions, safety constraints and execution feedback are equally important.

If these capabilities have not yet been connected, the robot can only chat with you and make expressions, and it will be hard for it to truly help you complete complex tasks.

But this does not mean it has no value.

On the contrary, physical robots may be a pragmatic product path: first use relatively mature hardware and lightweight interaction to build user perception, then gradually expand the services it can provide as software capabilities improve.

2. Why? The Pop Mart generation has grown up, but the key is that emotional value now has an interactive vehicle

Why do users need a little robot sitting in the car?

One very intuitive explanation is this: the generation that loves Pop Mart and is willing to pay for emotional value has already entered the car consumption market.

In the past, buying a car mainly revolved around functional needs such as space, power, fuel consumption, and features. Today, especially for young consumers, a car is increasingly like a form of personal expression: it not only has to be easy to drive and easy to use, but also has to have its own style, and ideally make life a little more fun.

But blind-box consumption can only explain why this kind of product has a market; it cannot fully explain why it is specifically an in-car robot. After all, people who like buying trendy toys do not necessarily need to put another toy in their car.

What makes the in-car robot special is that it is not just an object to be looked at, but an object that can respond to you.

It will turn its head to look at you, react to your voice, and may also change its expression according to music, driving scenarios, or the content of a conversation. These actions themselves are not necessarily very complex, but they change the relationship between the user and the object.

An ordinary ornament will not look up at you when you get in the car; an ordinary screen will not respond as if it had its own emotions just because you said something. When an object begins to show attention, emotion, and personality, people are more likely to treat it as an object they can interact with, rather than just a decoration in the car.

This also explains why some products that seem to have little practical function can still make people willing to pay.

During the 2026 Spring Festival travel rush, multiple in-car AI robots priced at about 200 yuan on Taobao showed sales of "10,000+" on their product pages. Some consumers call them "in-car companions" or "electronic pets." This set of data cannot represent the size of the entire market, but at least it shows that in-car companionship is not only a story told by automakers; observable consumer demand has also appeared in the low-priced aftermarket.

One noteworthy background is this: during an economic downturn, people's demand for emotional value actually rises. This is not an easy causal relationship to prove, but judging from the continued growth of the pet economy, trendy toy blind boxes, and in-car companionship hardware, when spending on big-ticket items becomes more cautious, people are more willing to pay for small purchases that "make the present feel a little better." An in-car robot costing a few hundred yuan happens to fall right within this psychological range.

3. On what basis? In the short term, it depends on speed; in the long term, it depends on insight

If demand for this category does indeed exist, the next question is: who can do it well? And in what way will it enter cars?

Judging from the current product forms, there are at least three different business paths.

1. Xiaomi Smart Display 2: from in-car peripheral to ecosystem extension

Xiaomi Smart Display 2 adopts a full-screen design, can turn its head, display expressions, and is compatible with models such as Xiaomi SU7, SU7 Ultra, and YU7. It is first of all an independent in-car peripheral, and users can obtain a new in-car interaction experience by purchasing it.

From a product design perspective, Xiaomi also has a condition that other manufacturers may not have: it has multiple terminals such as phones, speakers, and smart home devices.

If in the future the in-car display can form a more coherent interaction experience with these devices, it will not have to be just an isolated electronic pet, but can become a new entry point for Xiaomi's smart ecosystem in the automotive scenario.

The key here is not how many expressions are added to it, but making users feel familiar: at home, on the phone, and in the car, they are facing a smart ecosystem with a sense of family identity.

However, it should be noted that in currently public information, there are not yet clear implementation cases of deep linkage between Display 2 and Xiaomi's home ecosystem; ecosystem continuity is more of a product design direction than an established fact.

2. Geely Eva: first use independent hardware to validate demand

Geely Eva opened pre-sales in May 2026, with a reported price of 1,099 yuan. It uses a mounting handle and connects to the head unit via Bluetooth BLE and Wi-Fi, and can display corresponding animations in scenarios such as welcome upon entering the car, voice listening, navigation, and charging.

The compatible models listed in the report include some versions of the Galaxy Xingyao 7.

The advantage of this model is straightforward: it can first use a relatively independent product to validate demand, without waiting for a complete vehicle project to finish development, and without making it a cost that all models must bear from the start.

If users are willing to buy it, it shows that this experience is attractive to them; if users think it is not worth it, it will not affect the basic use of the vehicle.

Discussion around Eva also shows two typical attitudes: some feel the emotional value is maxed out, while others feel it is useless and not worth spending money on.

But interestingly, around NOMI, users have already developed behavior of buying hats, decorations, and other peripherals.

From the logic of traditional functional configuration, this seems hard to understand; from the perspective of user management, however, it shows one thing: when users begin to identify with a product image, what they consume may no longer be just hardware, but the decorations, interactions, and personal expression built around that image.

In September 2026, SAIC Motor introduced the Doubao cockpit assistant and Doubao screen solution for the Roewe Jiayue 07. Official materials state that through in-car perception, multi-zone voice recognition, and whole-vehicle SOA service interfaces, the model connects natural language understanding with navigation, cockpit entertainment, and other vehicle control capabilities.

Compared with aftermarket products, the advantage of pre-installation integration is that hardware, head unit, sound, power supply, installation, and interaction can be considered in a unified way during vehicle development.

If the product experience is mature enough, it will no longer be a toy that users need to additionally purchase and install, but part of the overall vehicle experience.

But pre-installation integration also means higher product responsibility.

Once it becomes a vehicle configuration, users will naturally expect it to run stably, work in concert with the head unit, and hold up over the long term in terms of user experience. Development costs, quality validation, after-sales maintenance, and model adaptation all need to be taken into account.

This is not a simple question of "should we put a robot in or not," but rather one that requires thinking about:

What kind of users does this car hope to add what value for, through the robot?

Suppose you are the product manager for an in-vehicle physical robot, facing a new project—where would you start?

My suggestion is: in the short term, first do expressions and lightweight interaction well.

If from the very beginning you try to connect vehicle status, vehicle control, navigation, intelligent driving, and various digital ecosystems, the project can easily get bogged down in complex system integration and safety validation, the development cycle drags on, and the market window may be missed as a result.

By contrast, expressions, movements, and simple scenario interactions are usually easier to use to quickly validate user reactions.

However, lightweight interaction does not mean casually making a few expressions, and even less does it mean copying the same set of animations onto every vehicle model.

What is truly worth thinking about is: what kind of robot does each different vehicle model actually need?

In the Geely Xingrui, it can be a "snarky-friend-type co-pilot."

If the target users are young people who value individuality more and enjoy the fun of driving, Eva does not have to always maintain a gentle, polite, and positive standard service posture.

In traffic jams, it can make a helpless "I'm annoyed too" expression; when playing favorite music, it can nod along to the rhythm; after completing a fun interaction, it can show a bit of a smug look.

It can even retain some slight "imperfections": yawning when it is first woken up, shaking its head twice when praised.

The value of these details is not in how many functions they add, but in giving it a more distinct personality.

The question it needs to answer is: can this thing make a person feel less bored while driving?

In the Lynk & Co 900, it may be more like a family companion.

The Lynk & Co 900 faces a different kind of need.

For families with children, especially those who often need to take children on long trips, the in-vehicle robot does not need to deliberately play the role of a cool tech toy, nor does it necessarily have to be a wisecracking "snarky friend" like the robot in the Xingrui.

It is better suited to being a companion that children are familiar with and parents can trust.

Here is a change worth noting: in-vehicle robots do not need to educate users from scratch, especially children.

Xiaodu speakers, Xiao Ai, and various children's story machines have already gotten many children used to talking to smart hardware. They know they can ask questions to an invisible voice, have it tell stories, guess riddles, play games, and chat with it for a few words when they are bored.

This means that when a robot with expressions that can turn its head and respond appears in the car, what the child faces is not a completely new way of interacting, but an extension of a familiar experience.

If they heard half of a dinosaur story at home, can they continue listening to it after getting in the car? If the child suddenly becomes curious about a bridge by the roadside, can the robot pick up on that question? On a long trip, if the child feels bored, can it start a guessing game suitable for playing in the car?

These may all look like small things, but they could constitute a new kind of product value: migrating the smart interaction habits children have already established from the home to the car, and making the experience between the two scenarios as continuous as possible.

Of course, the product cannot only design characters that children like; it also needs to give parents sufficient control: content and usage duration can be set, volume and interaction frequency can be adjusted, and features involving personal information should have clear permission boundaries. Interactions related to driving must especially avoid distracting the driver.

The question it needs to answer is: when children are already used to interacting with smart devices, can we bring that familiarity into the car, while also making parents feel that this journey has indeed gotten better?

Back to the three questions at the beginning.

What is it? It is an attempt by the smart cockpit to move from screens and voice toward embodied interaction, and a differentiated space discovered by automakers.

Why? The group of people who play with Pop Mart have grown up, society's acceptance of smart hardware has increased, and most importantly, during an economic downturn, people's demand for emotional value (carriers) has instead risen.

On what basis? In the short term, it is about who can make cuter expressions and who can iterate faster, but in the long term it still depends on whether one can truly understand users, integrate into the overall vehicle experience, and continuously create value through software iteration.

NOMI has used nearly nine years to prove that a small robot in a car can indeed become part of a brand.

And the new attempts by Xiaomi, Geely, and Roewe bring us to the next question: when every car has a robot, who can make users truly unable to do without it? Channel: Cars and Mobility. This content comes from the internet, original link. The views represent only the author and do not represent Huxiu's position. If there are copyright issues, please contact hezuo@huxiu.com, and we will verify and handle them promptly. Everyone who is changing and wants to change the world is on the Huxiu APP [ ]