Finance and Economics APP learned that Huajin Securities released research reports stating that Oladance, the leader in OWS technology, was acquired by ByteDance for 100% equity, becoming its wholly-owned subsidiary. Currently, ByteDance has formed a complete set of AI products including big model products, AI basic development, AI social, and AIGC creation, all of which rely on the technical support of Beanbag big models. The increasing demand for smart wearable devices and the rise of generative AI big models are pushing AI audio glasses into a wider market. In the short term, given the unclear user profile of AR glasses, smart audio glasses, acting as a replacement for headphones, with glasses (vision correction/decoration) as the carrier, combined with AI to improve the experience and selling points, are expected to open up a new audio track.

Acquiring Oladance to lay the foundation for entering the AI audio glasses market.

(1) Audio: Oladance, as the leader in OWS (OpenWearableStereo) technology, was acquired by ByteDance for 100% equity, becoming its wholly-owned subsidiary. In September 2021, as a leader in OWS technology, Oladance publicly announced the OWS technology concept for the first time, opening up a new track for wearable devices. OWS headphones are more integrated, smaller, and smarter, aiming to highly integrate small audio functions, positioned as wearable personal audio. By dynamically monitoring and calculating through EQ algorithm, forming synthesized sound sources with multiple sound source points, dynamically processing the phase relationship of the synthesized sound sources, transmitting sound to the ears, and then using patented smart bass compensation technology to suppress the diffusion of external sound and enhance the detailed performance of in-ear sound transmission.

(2) AI: Currently, ByteDance has formed a complete set of AI products including big model products, AI basic development, AI social, and AIGC creation, all of which rely on the technical support of Beanbag big models. In terms of big model products, there are already 9 types, such as Beanbag universal model pro, Beanbag universal model lite, Beanbag role-playing model, Beanbag voice synthesis model, Beanbag sound replication model, Beanbag document-image model, Beanbag speech recognition model, Beanbag vectorization model, and Beanbag Functioncall model. In multiple public evaluation sets and professional third-party evaluations, Beanbag universal model pro has consistently shown outstanding performance, ranking the highest among domestic big models. By July 2024, Beanbag's daily average tokens usage of big models has exceeded 500 billion.

MetaRay-Ban's sales exceed a million units, and the competitive landscape in the AI glasses sector may undergo a reshuffle. The core factors of MetaRayban's success are: the strong support of the Ray-Ban brand, design, and channels, the high cost-effective experience of the POV camera + OWS headphones combination, and the imagination space brought by multimodal AI. Ray-Ban's empowerment in sales channels and design guarantees that the product is wearable by users, ensuring a minimum sales volume, while Meta's iterations in audio, video, AI, and other areas are key to MetaRay-Ban's ability to create higher sales.

In addition to Meta, several manufacturers have already made layouts in the AI glasses sector:

(1) Huawei: Huawei Smart Glasses 2, this product is equipped with Huawei HarmonyOS 4 operating system and has access to Huawei Pangolin AI large model. Users can wake up Xiaoyi by tapping the leg of the glasses to achieve all-weather intelligent experience, supporting intelligent broadcasting of weather, flights, schedule, taxi, and other information.

(2) Beehive Technology (a Xiaomi ecological chain company): Jiehuan 1 is equipped with multiple large models, and AI cross-platform unified VUI experience is achieved through voice interaction. It has functions such as AI notification broadcasting, AI chat, AI pinching, and face-to-face translation.

(3) Li Weike: In the first half of this year, it launched its first AI glasses - MetaLensChat, priced at 699 yuan. This glasses focuses on AI voice interaction capabilities. Users can use voice interaction to let AI help solve work, schedule and other functions. According to the data from I Love Audio, the global smart audio glasses market shipment volume in 2022 reached 3.43 million pairs, a year-on-year increase of 58.6%, and the market size reached 0.91 billion US dollars. The compound annual growth rate (CAGR) of global audio glasses shipment volume from 2022 to 2030 will reach 13.6%, and it is estimated that by 2025, the global audio glasses market shipment volume will exceed 5.03 million pairs, and by 2030 it will reach 9.51 million pairs. In the future, if ByteDance launches AI glasses, with its advantages in AI, audio technology, brand awareness, etc., it will quickly attract market attention and bring competitive pressure to other AI glasses manufacturers or break the original competitive pattern.

Investment Recommendation: The demand for smart wearables and the rise of generative AI large models are pushing AI audio glasses towards a broader market. In the short term, with the unclear user profile of AR glasses, smart audio glasses can enter the market by replacing the function of headphones, using glasses (vision correction/decoration) as a carrier, and combining AI to improve the experience and selling points, thus opening up a new audio track.

Suggested focus on entering various product supply chains or manufacturers with relevant technological reserves. For assembly: Goertek Inc. (002241.SZ), Longqi Technology (603005.SH), Yidao Information (301314.SZ), Cosonic Intelligent Technologies (300793.SZ), etc.; for storage: Gigadevice Semiconductor Inc. (603986.SH); for cameras: Sunny Optical Technology (02382), Will Semiconductor (603501.SH), etc.; for SoC: Bestechnic (Shanghai) Co., Ltd. (688608.SH), Juxin Technology (688049.SH), etc.; for structural parts: Shenzhen Jame Technology Corp. (300868.SZ), Shenzhen Everwin Precision Technology (300115.SZ).

Risk Warning: Risk of undefined technology evolution trajectory and industry ecosystem; Risk of weak sensitivity to the industrialization process of forward-looking key technologies; Risk of lower-than-expected content ecosystem construction; Risk of lower-than-expected downstream demand.