/

AR glasses are the "complete form" that fulfills people's future imagination, while AI glasses are a "transitional product".

Taking advantage of the momentum of AI, giants have flooded into the AR glasses market.

Undoubtedly, Meta's second-generation smart glasses product, RayBan Meta, with its price of $299, weight of only 49 grams, and shipments in the millions, has demonstrated the feasibility of AI terminal scenarios for the public. Although Apple's VISION Pro encountered setbacks, it at least dispelled the notion that wearable devices can only be used for entertainment.

In addition to Meta and Apple, several domestic internet giants have also entered the market this year. On August 8, Xiaomi ecosystem company Beecomb Technology released AI audio glasses; on September 3, ByteDance acquired Oladance in full to pave the way for entering the AI audio glasses market; on November 12, Baidu released Duer AI glasses equipped with a large Chinese language model.

Even Samsung is joining the smart glasses battle. According to a foreign media report on November 18, the extended reality (XR) smart glasses jointly developed by Samsung and Google are expected to be released in the second half of 2025, with an initial production run of 500,000 units. These glasses will use Qualcomm's AR1 chipset, similar to those produced in collaboration with Ray-Ban and Meta.

With a relatively small user base for AR glasses, leading manufacturers have focused on AI glasses, which have become the mainstream product direction in the industry due to their lower technical threshold and higher user acceptance.

A research report by CITIC Securities points out that AI smart glasses are currently recognized as the most cost-effective hardware implementation solution for AI in the industry, and 2024 marks an inflection point in the field of AI smart glasses.

At this juncture, to quickly capture C-end users, in addition to the fierce competition among giants, the so-called "AR Four Little Dragons"—Thunderbird Innovation, XREAL, INMO, and Rokid—are also not to be outdone, engaging in fierce competition in products, prices, and ecosystems—frequently launching new products, enriching product lines, and collaborating with traditional glasses brands to increase influence on the demand side, even engaging in price wars down to the thousand-yuan level.

For example, Beecomb Technology has collaborated with GrandVision to create co-branded products and enter its affiliated Baoda Optical stores, while Thunderbird Innovation has established a joint venture with Bose Optical. Recently, Rokid launched its latest AI+AR glasses collaboration with BOLON glasses, the Rokid Glasses, priced at 2499 yuan.

Is AI+AR a sexy story? Which of the new products rolled out in succession by "Rokid and peers" can "conquer" and occupy the minds of the masses?

-Business Show-01 Price war in AI glasses market is "fierce"; "killer" products are absent

Since AI glasses came to prominence, price wars have intensified within the industry.

Initially, the unit price of AI glasses ranged from 2000 to 5000 yuan, and the "Four Little Dragons" also continued to focus on the price range below 5000 yuan.

As industry competition intensifies and lightweight products become popular, a large number of products priced in the thousand-yuan range have emerged in the market. Taking the smart audio glasses segment as an example, online monitoring data from Runto Technology shows that the average online market price of smart audio glasses in 2023 was 1167.8 yuan; among them, the market share of the 1000-1999 yuan price range reached 55.3%, the highest proportion; the proportion of the 400-699 yuan price range increased to 24.1%, a year-on-year increase of 14.7 percentage points.

From large companies to startups, it seems that everyone is "rolling prices" at the expense of cost.

Recently, Baidu launched its Duer AI glasses. According to reports from Kuai Technology, insiders revealed that the sales cost of Duer AI glasses may be lower than the $299 (approximately 2136 yuan) of Meta smart glasses.

Previously, this price may have been competitive, but now, with players represented by Thunderbird Innovation and XREAL having brought the prices of such products down to the thousand-yuan level, Duer AI glasses may lose their price advantage.

Previously, Thunderbird Innovation's Air 2s AR glasses attracted users with a pricing of 2698 yuan, but at the end of October this year, the latest AR glasses, Air 3, were priced at only 1699 yuan. Five months ago, XREAL launched the Beam Pro, which was also highly competitive. However, among the multiple versions available now, the price of the Beam Pro 6GB+128GB WiFi version is as low as 1299 yuan.

This is nothing short of a blatant "price war." For players, only by adhering to a low-price strategy can they attract potential customers in the mass market and thereby increase product shipments.

This round of price wars coincides with the point where AI glasses have become the mainstream product direction.

The product iteration of AR glasses has been controversial. Over the past few years, the "Four Little Dragons" have gradually advanced the integration of AR technology and smart glasses, but the vision of becoming the "next-generation smartphone" has not yet been realized. The overall size of the industry is still too small, and it is still far from the "next iPhone moment."

"Helmet-style" AR glasses still require high hardware configurations. For example, many AR glasses require at least 5-7 cameras to realize functions such as shooting, perception of reality, and eye tracking. One of the core highlights of Meta Orion is that it has 7 cameras.

Higher hardware configurations also mean higher cost inputs and purchase thresholds. The production cost of each pair of Meta Orion glasses is approximately $10,000, making it known as the "most expensive AR glasses in history." Meta plans to produce about 1000 Orion units, which translates to a hardware investment of $10 million.

Due to the common challenges faced in terms of cost, mass production, and technological research and development, project development delays or suspensions occasionally occur. Foreign media outlet The Information once reported that Apple's Vision Pro will cease assembly as early as November this year and will be discontinued subsequently.

The product form of AI glasses is entirely different. Characterized by being lightweight, compact, and easy to carry, it focuses more on integrating AI technology to enhance the user experience. It can be quickly implemented at this stage and provide consumers with a novel experience.

If AR glasses are the "complete form" that fulfills people's future imagination, then AI glasses are a "transitional product".

As a result, many AR manufacturers have begun to claim that they are pursuing an "AI+AR" approach. In reality, no one can accurately predict which type of product or technology route will dominate in the future. AR manufacturers can only choose a two-pronged approach, focusing on the research and development of both AR glasses and AI glasses.

However, no one is willing to give up on AI glasses, a "transitional product". Li Hongwei, the founder and CEO of Thunderbird Innovation, once mentioned that in the past one or two years, the sales volume of AI glasses will be much larger than that of AR glasses.

This is indeed the case. A research report released by Wisdom Information shows that global AR sales in the third quarter of 2024 were 106,000 units, a year-on-year decline of 4%, marking the first year-on-year decline in AR glasses sales in recent years.

The main reasons for the overall sluggish sales of AR glasses include weakened channel promotion and investment, shrinking demand for B-end AR glasses, and the rise of AI smart glasses, which is quietly taking a share of the market "cake".

However, a phenomenal product has not yet emerged. From a practical perspective, it is time for players in the AI glasses market to take up positions in advance and increase investment. Cruelly, to win the favor of mass consumers, there is still a long way to go in enriching functions and application scenarios.

-Business Show-02 Weak willingness of mass consumers to pay

Despite the prevalence of AI glasses, existing products cannot escape the issue of homogeneity in terms of product features and niche scenarios, making it difficult to truly meet the needs of the masses.

Over the past few years, AI glasses products have undergone several iterations but have maintained relatively "lightweight hardware" characteristics because consumer usage habits have guided product development—an AI glasses product that possesses AI functions but overall resembles ordinary glasses is more popular among most users.

To make them easier to wear, most products have reduced their overall weight, but when actually worn, there is still a significant gap compared to ordinary glasses, and they cannot be completely unnoticeable.

AI products launched by Thunderbird Innovation, such as the Thunderbird X3 and Thunderbird Air 3, have balanced weight and functionality, with product weights ranging from approximately 60 to over 80 grams.

Li Hongwei once mentioned in a conference, "The Thunderbird X3 really looks like a pair of glasses... weighing around 60 grams, which is the weight of a normal pair of sunglasses that everyone wears." However, in reality, nearsighted glasses and sunglasses worn by consumers generally weigh around 30 grams.

There are also other lighter products on the market, such as the Rokid Glasses, which weigh 49 grams, making it a relatively obvious lightweight advantage in China. For comparison, the popular foreign product RayBan Meta can only achieve a weight of around 48 grams.

In addition, the user experience of most mainstream applications is limited, and more complex functions cannot be developed due to hardware constraints. Although consumers have little patience for the interaction process and prefer foolproof operations, they still hope that products can provide a complete smart experience.

Both the Duer AI glasses that have not yet officially gone on sale and the Rokid Glasses integrate functions such as first-person shooting, AI chat, and translation, with extremely high repetition rates. In addition, most products come with mainstream applications such as navigation, movie watching, and game screen casting; in niche scenarios, most players mainly develop functions around fields such as gaming, culture and tourism, education, and entertainment.

Specifically, battery life is a major "pain point." AI glasses have small battery capacities and short battery life, unable to meet the needs of long-term use, resulting in most products only being used in fixed locations.

For example, the Rokid Glasses have a battery capacity of 220mAh, with a continuous daily usage time of 4 hours and a working time of only 40 minutes during shooting; Duer AI glasses do not mention specific values, but the continuous playback time is 5 hours; Thunderbird Air 3 claims to have a built-in battery capacity of up to 4000mAh, but under normal use, the battery life can only reach 6-8 hours.

In terms of processors, some products adopt the Qualcomm AR1 solution, such as the Rokid Glasses; others choose to built-in Qualcomm Snapdragon 6 Gen 1, a chip for the low-end market, such as the XREAL Beam Pro and Rokid AR Lite, which have been questioned as "sacrificing core capabilities to reduce costs."

Due to hardware limitations, in scenarios such as gaming, movie watching, and shooting, many AI glasses experience issues such as laggy images and screen casting disconnections, preventing an immersive user experience.

It can be seen that most leading AR manufacturers with hard technology capabilities have not adopted the same "aggressive" approach as when developing AR glasses. The development of AI products is more conservative, placing greater emphasis on the preferences of the mass market.

There are many similar stories. Xia Yongfeng once mentioned a story in an interview with AI Light-Year. In mid-2021, the company he founded, Beecomb Technology, received two rounds of funding in succession. Many investors hoped that the company would incorporate dual-screen displays in its products, but he rejected them all.

He believes that the core need of users for glasses is the ability to wear them comfortably for extended periods rather than various complex display functions in front of the eyes. Currently, AI technology does not create new user needs but rather enhances the user experience.

However, if mass consumers are to pay for AI glasses, product development cannot simply lower the usage threshold or "castrate" product functions. Instead, it should advance technological research and development, providing a feasible solution that balances performance, cost, and usability while improving product functions and application scenarios.

When players do not focus on creating user needs, it is difficult to produce a highly innovative product that disrupts the market.

-Business Show-03 The inflection point has arrived; who will break the deadlock?

With the rapid development of technology, AI+AR glasses have achieved preliminary commercial implementation and product applications, but they are still insufficient to change the lives of mass consumers.

The market is still in an early stage of development with relatively low penetration. According to GIR data, global smart glasses shipments climbed from 193,800 units in 2014 to 6.7553 million units in 2023. However, the current global penetration rate of smart glasses is less than 1%.

However, with the arrival of a new wave, the entry of more giants means that they are willing to concentrate their efforts on and harvest this niche market.

However, this is not good news for "veterans" in the sector.

At a stage with low barriers to entry and market penetration, "veterans" are more likely to seize the market and become leaders. Once competition intensifies or even enters a reshuffling phase, "veterans" without strong capabilities and continuous financial support are likely to be the first to be eliminated.

Undoubtedly, the development of the AI+AR glasses sector requires significant financial support.

This year, the capital market has once again witnessed an investment boom in AR. Thunderbird Innovation completed three rounds of funding within half a year, with the most recent being hundreds of millions of yuan in B+ and B++ round funding jointly invested by institutions such as Jiaxing Nanhu Keying. To date, it has raised over 500 million yuan in cumulative funding within half a year. Rokid, one of the "Four Little Dragons," has received nearly 500 million yuan and 100 million yuan in consecutive funding in January and September this year, respectively, bringing its total funding to over 2 billion yuan after completing 13 rounds of funding.

But as everyone knows, internet giants like Baidu, Xiaomi, and ByteDance, as Chinese technology companies with the most cash and a significant investment portfolio, have their hands in almost all popular sectors. Once they enter the market, it will inevitably lead to more intense capital competition within the sector.

In the AI+AR sector, the "Four Little Dragons" must possess capital sufficient to compete with giants if they want to reach the end.

At the initial stage of the market, burning money is not a "get-out-of-jail-free card" for players. Ultimately, players need to transform their huge investments into insurmountable competitive barriers—involving not only technology, products, and markets but also the integration and cooperation of upstream and downstream industries.

At this stage, the "Four Little Dragons" possess a large number of implemented products, forming a rich product matrix, which is conducive to their cooperation with upstream and downstream industries in the industrial chain and completing market education. For example, products like Thunderbird X2 have begun to be compatible with the WeChat Mini Program ecosystem, expanding product influence through social software with a large user base; Beecomb attracts offline market consumers by collaborating with traditional glasses companies to create products, especially penetrating user groups that have already formed wearing habits.

However, the current market size has not yet taken off, and the expansion of online and offline channels is actually premature.

In contrast, the technical accumulation around AI algorithms and models is more likely to exert influence among consumer groups through strong innovation, thereby stimulating market growth.

In this regard, players with in-depth layout in the AI industry have more advantages. For example, Baidu's Duer AI glasses, equipped with a large Chinese language model, can provide more powerful AI chat functions. However, players still need to accumulate technology in core capabilities such as speech recognition, image recognition, and natural language processing to provide intelligent suggestions and functional optimizations for AI glasses.

An open ecosystem is also a highlight of AI+AR glasses. Players need to build an open AI glasses ecosystem to attract more developers, partners, and users to join.

On the one hand, provide commonly used applications such as social media, short videos, shopping, etc., and encourage developers to develop more applications and services for AI glasses; on the other hand, realize the interconnection between AI glasses products and more intelligent hardware devices. Mobile phone manufacturers, including Huawei, Xiaomi, etc., can transfer their accumulated experience in hardware and software to AI glasses.

Overall, the giants have just begun their layout, but they already have their own accumulations and advantages in core areas. The giants' increased investment in niche segments is often a signal for a new round of competition to begin.

For startups, as long as they delve deeply enough into vertical areas, they are unlikely to be eliminated. In the long run, if the "Four Little Dragons" fail to increase investment in core technology barriers such as AI and improve their ecosystem capabilities, their future may very likely involve "working for" the giants.

Who will become the "game-changer" of the industry?

Everyone is eagerly anticipating the emergence of a C-end "killer" product within the industry—just like WeChat in social software or smartphones in hardware devices, becoming a necessity in consumers' lives.

And before the dawn arrives, players still need more patience and determination to perfect their products, ecosystems, and technologies. [End]