Why are you reluctant to wear AI glasses? Do you believe it can replace smartphones?
Let me ask you a question: Would you be willing to wear a pair of AI glasses? Do you believe that AI glasses will replace mobile phones in a few years?
By the end of 2025, when we look back at the AI smart glasses industry this year, we will find that it is experiencing explosive growth, which has excited everyone. Its impressive data has also made the market and the media increasingly optimistic.
According to the latest research and judgment by IDC, the smart glasses sector is experiencing unprecedented popularity in 2025, with global annual shipments expected to climb to 12.8 million units. Among them, the Chinese market is expected to top the global list, with an expected sales volume of over 2.75 million units and an explosive growth rate of 107% year-on-year.
(Image source: IDC)
In the industry, the topic of "replacing smartphones" has never stopped, and everyone is anticipating their own "iPhone moment". However, if we look back at the history of mobile internet, the path to the popularity of smartphones is not simple - it relies on a mature, excellent, and stable hardware foundation, as well as the strong driving force of "killer apps".
Looking back at the present, the AI glasses, which have been buzzing for a whole year, are revealing their weaknesses - the industry has yet to come up with a decisive scenario that makes users feel "indispensable", and there is still a lack of core motivation to break out of the niche circle. The absence of comfort also makes it difficult to see people wearing them casually and naturally on subways, in shopping malls, and in offices.
This contrast of "cold and hot temperature differences" constitutes the final chapter of AI glasses in late 2025. After generative AI surpassed XR to become the new darling of the tech industry and raced around the world for three years, AI smart glasses failed to seamlessly integrate into people's lives like TWS earbuds and smartwatches.
1
The dilemma of the current situation is emerging: the "false heat" in data
Looking through all the PPTs from AI glasses manufacturers I received this year, it feels like I'm back in 2016, the moment known as the "first year of VR". But when I turn my attention to the back-end data of e-commerce, I see an imperceptible "trial and error game".
Image source: JD.com
According to Sina Finance, the return rate of mainstream AI glasses in China this year has remained high at around 30% on traditional e-commerce platforms such as JD.com and Tmall; however, on interest-based e-commerce platforms like Tiktok and Kwai, this figure even soared to 40%-50% at one point.
This figure is somewhat alarming. It implies that for every two pairs of glasses sold by the manufacturer, one pair will be returned within seven days, accompanied by strong disappointment. This high "abandonment rate" reveals, to some extent, the industry's "dilemma", that is, traffic has not been converted into retention.
Based on our observation, many consumers are drawn to futuristic concepts like "AI assistants" and "holographic displays", but they leave disappointed due to the significant gap between expectations and reality. This "false hype" not only drains consumers' patience but also exposes manufacturers to the vicious cycle of high marketing costs and low user engagement.
2
Cognitive dissonance and pain points: Jarvis or Bluetooth accessories?
The public's imagination of AI glasses has been largely "misled" by Hollywood blockbusters.
Consumers are anticipating Jarvis from "Iron Man" - an all-around assistant capable of independent thinking, real-time traffic analysis, and hands-free operation. However, many mainstream products delivered to the market in 2025 still retain the impression of being "mobile phone accessories".
The pseudo-independence of "being unable to be separated from mobile phones" - this is the biggest "flaw" of current AI glasses, and also the most heavily criticized area by users. Many mainstream products on the market, priced between 1500 yuan and 2500 yuan, lack powerful computing power and network connectivity, and are more like "Bluetooth headphones with cameras and microphones".
To be specific, many manufacturers emphasize AI reasoning and voice-to-text conversion, which rely on the SOC and APP processing on the mobile phone side. The glasses are only responsible for data input and feedback display. Once the mobile phone signal is poor (such as in a subway or elevator), or Bluetooth connection is interfered with, the glasses will instantly become "dumb". And when the glasses are running out of battery, even the most basic voice activation will fail.
Social awkwardness - Ideally, AI should be like an invisible secretary. However, in reality, it often brings moments of "social death" to its wearer, especially in our cultural context.
Imagine you're in a crowded elevator, trying to wake up your glasses to check your schedule. However, due to the noise interference, you have to shout "Hey, XX" into the air. In the end, your glasses even disclose your privacy in public. Fundamentally speaking, although AI has made great progress in multi-turn dialogue capabilities, its sound pickup and comprehension abilities in noisy environments are still weak.
Let's look at another scenario. In a supermarket, you, who love trying new things, want to pay with your glasses: it can be used, but under the gaze of others, you awkwardly shout out, "XX, payment 50 yuan." In the eyes of many people, this "social pressure" far outweighs the "tool value" that the glasses can provide, which will make some users choose to abandon it or even not try it.
3
The shackles of physics: technology compromises with reality
Chinese manufacturers have strong R&D and imitation capabilities, but they can't make perfect AI glasses? Because there is no magic in physics, and there is currently no perfect model to imitate.
The predicament of the entire industry, fundamentally speaking, is that manufacturers are engaging in an extremely painful exercise of "robbing Peter to pay Paul" across four dimensions: computing power, battery life, weight, and heat dissipation. It is normal to receive some negative feedback. Given the background of manufacturers (technology) compromising with reality, we at least see three shackles that are difficult to break through in the short term:
The red line of weight - ordinary glasses typically weigh between 20-30 grams. Once AI glasses exceed 50 grams, prolonged wearing can cause pressure on the nose bridge and a noticeable foreign body sensation. To keep the weight under 50 grams, manufacturers have no choice but to drastically reduce the battery capacity.
The shortcoming of battery life - Currently, mainstream products often have a continuous usage time of only 3-4 hours. If video recording or high-intensity AI conversation is enabled, it may not even last for an hour. This makes it impossible for it to become an all-day wearable device.
The paradox of heat dissipation - we desire AI to respond extremely quickly, but this requires the chip to operate at high speed. However, with the temples of the glasses pressed against the temples, even a slight temperature rise of a few degrees can cause a "brain-burning" discomfort to the user. To avoid burning the user, the system often forces a frequency reduction, leading to stuttering and eliciting negative feedback from the user.
All the shortcomings in user experience ultimately stem from the red lines drawn by physics.
To cram smartphone-level computing power and battery life into a device weighing 50 grams is inherently a challenge against the laws of physics. Although the outcome is still uncertain, the cost-performance comparison between different types of products remains a natural gap that smart glasses manufacturers must confront before smartphones fade away.
By the end of 2025, a pair of AI glasses with decent experience will still be priced firmly within the range of 1499 yuan to 2499 yuan. With the same budget, consumers can either purchase a domestic smartphone with excellent configuration and experience (full-featured and without obvious shortcomings), or a pair of top-notch noise-cancelling headphones.
In comparison, purchasing a semi-finished product with "short battery life, heat generation, and occasional inability to connect to a phone" for 2,000 yuan or even higher is not a wise decision for the average consumer.
4
Market prediction: Victory of the "Waiting Party"
The price war in 2026
Since we've mentioned consumer psychology, it's necessary to talk about the C-end mass market: this year's wage earners are smarter than ever before, because apart from houses, everything seems to be rising in price.
In the field of consumer electronics, after experiencing rounds of advancements in VR headsets and the metaverse, players are now facing a frenzied surge in memory and SSD prices at the end of the year. Given these circumstances, the traditional skill of "waiting for the next generation" is currently at its peak. Their logic is impeccable: "Electronic products should be bought new, not old, especially when AI technology iteration is calculated on a monthly or even weekly basis. This year's flagship product will become next year's trash."
This mindset gave rise to two interesting phenomena in 2025:
Firstly, "Xianyu Leasing" has become popular. Yes, many users don't buy new devices at all, but instead rent them for a week for a few dozen yuan on Xianyu. Their mindset is: "I just want to try if this thing works. Since it's likely to be useless, why spend thousands of yuan?"
Next, the expected emergence of "Huaqiang North alternatives" has been observed. During a business trip to Shenzhen, the author noticed that in the market of China's 3C holy land, a large number of imitations priced at only two or three hundred yuan have appeared. Their appearance is highly reminiscent of those from major manufacturers, and although their functions are rudimentary, they are sufficient to meet the demand of "showing off on social media". According to IDC data, their sales volume accounts for the majority of the industry's total sales in 2025.
This is not a healthy industry ecosystem. If this mentality persists, we might boldly predict that the AI glasses industry will face a fierce "price war" and a "reshuffle period" from 2026 to 2027.
Analytically speaking, with the rapid maturity of upstream supply chains (especially domestic optical waveguide modules and low-power AI chips), hardware costs are plummeting. Consequently, the current era of profiteering is coming to an end, and low-end AI glasses may transition from being "standalone products" to becoming "phone giveaways" or "car accessories". This is not mere speculation; we are already seeing signs of it.
Firstly, there's the dimension-reducing attack from car companies. On December 3, 2025, Ideal Automobile officially launched its first AI glasses, "Livis", which bears some resemblance to the "Jarvis" series. Its pricing is set to fall directly within the 2,000 yuan range. Please note that Ideal does not define it as an independent consumer electronic product, but rather as an "automotive accessory" - it is an extension of the car's screen and a carrier for in-car voice interaction.
Secondly, there's the blatant strategy of mobile phone manufacturers. Just this year, a major domestic mobile phone manufacturer has begun experimenting with offering AR/AI glasses as a "privilege package" or "large discount gift" for purchasing flagship phones. Of course, this does not mean that their products are inferior, but rather a shift in the manufacturer's marketing approach, which may hint at subtle changes in product positioning.
In the future, it may become the norm to receive glasses as a gift when purchasing a car or a mobile phone. Will this spell doom for independent eyewear manufacturers who lack ecological support and rely solely on selling hardware for profit margins? As an industry media outlet, we do not wish to see the competition field become desolate.
5
"The Survivors" Are the Kings: How to Survive in the Price War?
In all areas of business competition, price wars are an unbeatable force. If hardware is destined to be priced like cabbages, how can manufacturers survive?
This brings us back to the ultimate question of the category: the competitive advantage of AI glasses has never been in the glasses themselves. In my opinion, in the future, only the following four types of players are likely to survive and win in the price war:
The first type is the cross-industry eco-giant.
For example, major mobile phone manufacturers or automobile companies. They don't expect to make money from smart glasses, but rather regard them as the "tentacles" and "glue" of their own ecosystems. The research and development and marketing departments actually sell services, cloud space, and driving experience. They have unlimited ammunition in the price war.
The second type can be classified as "special forces" in vertical scenarios.
Some companies do not produce general-purpose glasses, but specialize in specific scenarios. For example, they produce real-time subtitle glasses specifically for the hearing-impaired, and glasses with overlay drawings for industrial maintenance. In these fields, users are not sensitive to price, but they have extremely high requirements for accuracy, stability, and professionalism. It is difficult for general-purpose manufacturers to enter these markets.
The third type is an inevitable product of the AI era: a data and model driver.
Hardware is merely the entry point for data collection. The first-person video data (Ego-centric Data) collected through glasses is used to train more powerful and profitable multimodal models, which are then charged through software subscriptions (SaaS). This model is extremely challenging, but also boasts the highest barriers to entry.
The fourth type is industry leaders. One is Meta, another is Apple (recently, there have been reports about Apple's plan for smart glasses), and then there's Google, which has just returned to the race. There's not much to say about these companies; they strive to be role models in the industry.
6
Our perspective: From "trending topic" to "everyday life"
From the perspective of the ultimate form of human-computer interaction, AI glasses, with their physiological advantage of being close to the senses and the uniqueness of the first-person perspective, are indeed the theoretically most perfect AI entry point. Therefore, after delving into these deeply hidden details, if someone asks: Will the category of AI glasses still be a hot trend next year?
Our answer is simple: yes.
But it is no longer the opportunity where "even a pig can fly", but rather the airspace where "eagles fight". Before becoming a true next-generation computing platform, it must undergo a self-transformation - it must at least be a pair of comfortable, good-looking, and unobtrusive "glasses", and then it can become an "AI terminal".
The high return rate and market cooling-off period in 2025 are certainly not the end of the industry, but rather the build-up of momentum before a major explosion. As consumers, we are experiencing a period of transition from "trying new things" to "regular use".
On the morning when AI glasses truly integrate into my life, you will naturally encounter them in the armrest box of your friend's new car, in the packaging box of your phone, or even in your girlfriend's small perfume bag.
mail:wuping@csc-ic.com ； angel@csc-ic.com
right 2023. Chinaic Technology Co. ,Ltd. All Rights Reserved 粵ICP備13001812號
