AR to the left, AI to the right, smart glasses at a crossroads

12/09 2024 508

Author | Chen Wen

Source | Insight New Research Institute

From Google Glass to Vision Pro, human exploration of smart glasses has made progress, but none has been truly successful. It was not until the emergence of Ray-Ban Meta that the industry regained confidence in smart glasses. Since last year, with the deepening application of large AI models, smart glasses have a new story to tell.

In the second half of this year, Baidu and Rokid, a leading domestic smart glasses manufacturer, released AI-enabled smart glasses. In addition to Baidu and Rokid, Xiaomi and Apple are also rumored to be entering the smart glasses development field. Meanwhile, OPPO, vivo, Huawei, Tencent, ByteDance, and other technology giants are evaluating AI glasses projects.

Is it to continue down the AR path or blaze a new AI trail? Today's smart glasses stand at a crossroads.

01 The Battle of a Hundred Mirrors

In 2012, Google Glass, dubbed Google's "moonshot project," was launched. In Google's plan, the ultimate mission of Google Glass was to "replace mobile phones" and become a "smartphone" that people can wear on their eyes every day.

Compared to mobile phone screens, Google Glass used prism projection technology to display emails, text messages, navigation information, etc.

Rivaling mobile phones' multimedia entertainment functions, Google Glass could capture first-person footage through its built-in 5-megapixel camera and replay videos through prism projection.

Matching mobile phones' internet capabilities, Google Glass supported voice search interaction, allowing users to inquire about weather, stock prices, and other knowledge-based questions.

The reality was harsh; Google Glass did not achieve the expected sales success, let alone "replacing mobile phones." In 2015, Google suspended sales of Google Glass.

The industry attributed Google Glass's failure to issues such as marketing strategy, lack of application scenarios, and market acceptance. Among them, the American technology media Verge clearly pointed out that many functions provided by Google Glass (such as voice photography and navigation) were also available on smartphones, but the experience on smartphones was better and more convenient.

Since the path of "replacing mobile phones" was not viable, smart glasses manufacturers focused on the original function of glasses—providing a differentiated visual experience. Therefore, they persisted in display technology, and AR/VR glasses became the mainstream in the post-Google Glass era.

To achieve better display effects, the design of AR/VR glasses was very different from that of ordinary glasses, often adopting exaggerated and sci-fi styles. The weight of the glasses remained an unresolved issue in the industry. Coupled with a poor application and content ecosystem, AR/VR glasses have not caused much commotion in the consumer market.

Apple gathered over 2,000 designers and engineers, spending billions of dollars over eight years to develop Vision Pro, which is a typical example.

In terms of product capabilities alone, Vision Pro has indeed achieved the pinnacle of smart glasses. The two Micro OLED display screens can achieve a display effect close to 8K for both eyes, with a pixel density of an astonishing 3400 PPI, far exceeding the PPI resolution limit within the viewing distance of smart glasses.

Before the release of Vision Pro, Apple fans were eagerly anticipating it. However, after the product's launch, the market voted with its feet, and production was discontinued less than a year after its release, making it Apple's "shortest-lived" product.

After AR/VR glasses, the industry ushered in an AI glasses boom activated by Ray-Ban Meta.

This smart glasses collaboration between Meta and Ray-Ban, launched in September last year, has sold over a million units in just over a year. Annual sales in 2024 are expected to exceed 2 million units. A research report released by Guotai Junan Securities predicts that when Ray-Ban Meta's sales network is established globally, its annual sales potential could rise to over 6 million units.

Ray-Ban Meta has clearly provided the industry with great demonstration and stimulation. Among Chinese manufacturers, Baidu's "Xiaodu AI Glasses" launched this year and Rokid's "Rokid Glasses" in collaboration with the fashion eyewear brand BOLON have attracted considerable attention.

Baidu's entry into AI glasses stems from the ERNIE Bot large model, known as the "most proficient in Chinese," which can seamlessly integrate with Baidu's application ecosystem, including Baidu Maps, search, and encyclopedia.

Rokid's newly launched "Rokid Glasses" incorporates the algorithmic capabilities of Alibaba's Tongyi Qianwen large model, enabling functions such as question answering and real-time translation.

In addition to Baidu and Rokid, many other technology companies are also deploying smart glasses products.

In April this year, the AR brand "Li Weike" released the AI smart glasses Meta Lens Chat. In August, Xiaomi ecosystem company Beecomb Technology launched the Jiehuan AI Audio Glasses. In the same month, Thundercomm Innovation and Bosshi Optical established a joint venture, with the first product expected to be released by the end of the year...

According to public information and channel news, at least 50 companies in China are currently advancing smart glasses projects. These include leading enterprises such as Xiaomi, Huawei, Baidu, and ByteDance, as well as emerging manufacturers that rose during the previous wave of smart glasses, such as Thundercomm, INMO, and Rokid. Among overseas manufacturers, giants such as Apple, Samsung, and Google are also eager to participate, creating a landscape reminiscent of the "Battle of a Hundred Mirrors."

02 The Opportunity Brought by AI

After carefully studying the success factors of Ray-Ban Meta, all AI glasses manufacturers have listed photo recognition, audio recording and playback, and AI invocation as the three main functions of their products. That is, smart glasses are both cameras and headphones, as well as AI assistants that can translate and navigate for you.

The definition of AI glasses seems to have returned to the era of Google Glass. With the rise of large AI models, the industry needs other physical carriers for AI applications similar to mobile phones.

Looking back, Google Glass's product design concept was actually sound; it was just that the timing of the idea was off. As the theory of "leading by half a step" goes, "being one step ahead makes you a pioneer, but two steps ahead makes you a martyr."

The current technological maturity is vastly different from that of 12 years ago when Google Glass was released.

Taking the newly released Rokid Glasses as an example, unlike previous split designs, this version adopts an integrated structural layout with a weight of only 49g. The Rokid product team integrated the motherboard, battery, and other core components into a slender frame, making the overall shape closer to traditional glasses, greatly enhancing the wearer's comfort and portability. This represents progress in smart glasses industrial design.

With the support of light waveguide technology, this AR glasses can display a clear and delicate virtual interface through transparent lenses without affecting normal vision, achieving an information overlay display effect and thus opening up a new human-computer interaction mode. This represents progress in display technology.

In terms of functionality, Rokid Glasses can easily perform tasks such as object recognition, text translation, and even mathematical problem-solving by deeply integrating the algorithmic capabilities of Alibaba's Tongyi Qianwen large model.

For example, it can quickly and accurately calculate the calorie content of food and present it to the user instantly. Or, in cross-cultural communication scenarios, it can automatically capture the conversation content and translate it into the other party's language in real-time, facilitating communication between international travelers. This is a specific application under the technological progress of large AI models.

On the other hand, the maturity of the supply chain has greatly lowered the entry barriers for AI glasses.

In the opinion of industry insiders, pure AI glasses are lightweight products with relatively simple hardware components. AI glasses manufacturers only need to find a white-label glasses supplier, purchase cameras, acoustic modules, Bluetooth modules, batteries, and processors, and then develop an app that connects to an open-source large model. A basic AI glasses with essential functions is then basically formed.

In fact, such products have been available for a long time, even earlier than Ray-Ban Meta. The price of "AI Smart Audio Glasses" on Pinduoduo generally does not exceed 100 yuan, and the standby time can reach an astonishing 240 hours. This is one of the reasons why the industry has formed a competitive landscape resembling the "Battle of a Hundred Mirrors."

03 AR+AI=Future

It is not difficult to see from the above that the development of smart glasses has formed three technological routes: "smart glasses with AR functions," "AI glasses without display screens," and "audio glasses" focusing on audio experience.

Currently, the main difference between AR glasses and AI glasses, apart from their display capabilities, lies in their interaction methods. AI glasses use VUI (Voice User Interface), mainly interacting through voice, TTS (Text-To-Speech), and light feedback. AR glasses use GUI (Graphical User Interface), mainly interacting through display, touch control, voice, and TTS.

However, neither type of smart glasses has yet to overcome the "impossible triangle" of battery life, weight, and computing power at this stage. Like all smart devices, the more functions smart glasses have, the higher their energy consumption will be. Increasing battery capacity will affect the weight of the glasses.

In this regard, the pain points of AR glasses are particularly prominent. The cumbersome headset design of some products discourages many users from trying them out. Although "AI glasses without display screens" reduce weight to some extent, they are still heavier compared to ordinary glasses weighing 20-30 grams.

Currently, the weight of mainstream AI glasses on the market is around 45 grams. However, if equipped with lenses, the weight will increase by more than 10 grams depending on the lens prescription, making prolonged wearing uncomfortable.

Moreover, almost all current AI glasses have not solved the battery life issue. Compared to mobile phones, cameras, or other handheld shooting devices, Meta can maintain continuous shooting for up to three hours. This is also a pain point for most current AI glasses products.

This means that the product form of smart glasses has not yet been fully established. Whether AR glasses or AI glasses, there are still many technical issues to be resolved before the industry truly takes off.

For example, some AR glasses consciously integrate AI technology to broaden their application scenarios. Some AI glasses manufacturers are striving to enhance their display capabilities to compensate for their shortcomings.

One significant reason for this situation is that the industry generally believes that AR glasses are the ultimate form of smart glasses. "Humans receive 80% of information through vision. Without visual interaction, AI capabilities alone cannot become a new entry point for humans to obtain information and services," commented a professional.

Xu Chi, founder and CEO of XREAL, once said in an interview with the media, "AR is the best carrier for AI, and AI is the best interaction for AR. These are likely to be the two core technologies that can most change consumer electronics and our terminal experience in the next decade."

Regarding the future of smart glasses, AR and AI may merge. However, at present, both still need to face the same challenge of achieving scale and rapid penetration. It is worth noting that Ray-Ban Meta, the industry's only blockbuster, has only sold over a million units in a year, a number that smartphone manufacturers can achieve in just three days when launching new products.

Solemnly declare: the copyright of this article belongs to the original author. The reprinted article is only for the purpose of spreading more information. If the author's information is marked incorrectly, please contact us immediately to modify or delete it. Thank you.