Preface:

Artificial intelligence serves as a versatile container, accommodating a myriad of concepts and applications. Its most profound impact on human-machine interaction and software-hardware collaboration lies in the [input] process, rather than the [output].

Hardware, as an entry point, transcends being merely a means for users to access services. It harbors immense imaginative potential, capable of collecting multidimensional data from users' daily lives.

Author | Fang Wensan

Image source | Network

AI glasses offer distinct advantages when paired with AI models.

In the realm of VR, AI glasses are viewed as a pivotal transitional product en route to realizing the metaverse. In the sphere of large models, hardware remains a critical direction for large model implementation, with AI glasses playing a pivotal role.

Compared to ambitious ventures like AI smartphones, AI glasses exhibit a more "compact and exquisite" form factor, blending technological sophistication with substantial market potential. Furthermore, as a mature product category, AI glasses do not necessitate significant investment in market education.

Dongxing Securities analyzed the industrial chain, noting that the localization level of the AI glasses industrial chain is relatively high. Leveraging the mature TWS earphone industrial chain, lightweight smart bodies can be directly deployed on the end side.

AI hardware has yet to demonstrate the capability to spawn entirely new product categories, and large models have not evolved into the most sophisticated multimodal forms. Currently, most practical applications still rely on the semantic comprehension abilities of large language models, with the necessary application scenarios for AI hardware remaining unclear.

Theoretically, AI hardware can provide a more convenient interface for large models, making it easier for users to invoke large model functions.

ByteDance boasts significant advantages in talent density, GPU resources, and technical accumulation, particularly in the initial stages of algorithm development and the application of AI technology across its products. These factors have enabled ByteDance to maintain a leading position in exploring application areas.

Competition in the AI headphone market is unlikely to mirror the fierce rivalry seen in the smart speaker market. Despite some similarities with audio products, the headphone industry's competitive landscape is expected to be less intense.

According to data from the "Global AI Wearable Market Report," the market size of AI wearable devices is projected to grow from USD 41.9 billion in 2024 to USD 120.7 billion by 2028, registering a compound annual growth rate (CAGR) of 30.3%. This growth trend underscores the enormous development potential of the AI wearable device market, affirming that the field of AI smart hardware boasts broad development prospects.

The final crucial piece of the puzzle to complete ByteDance's AI layout involves end-to-end AI hardware products.

One of ByteDance's primary objectives in launching these headphones is to provide more convenient access to Doubao applications, thereby lowering the threshold for user engagement. Doubao has emerged as the domestic AI application with the largest user base, and ByteDance has made comprehensive layouts in the field of AI large models and applications.

ByteDance has developed multiple text and multimodal large models, including the Doubao large model family, which encompasses a wide array of types such as general models, role-playing models, speech synthesis models, voice cloning models, speech recognition models, text-to-image models, Function Call models, and vectorization models.

Additionally, ByteDance has crafted multiple AI-native applications based on the Doubao large model, encompassing over 10 AI products like Doubao, the AI application development platform "Kouzi," and the interactive entertainment application "Maoxiang," among others.

End-to-end AI hardware products could be the final crucial piece of the puzzle for ByteDance to complete its AI layout. Therefore, companies like ByteDance and other internet giants are not solely focused on hardware itself when investing in AI hardware development; rather, they are centered on application scenarios, particularly those frequently used or with high-frequency usage potential.

Doubao requires traffic inlets amidst fierce competition.

According to AppGrowing statistics, the volume of promotional materials for large models like Doubao, Kimi, Zhipu Qingyan, and ERNIE Bot surged significantly in the first week of June. Doubao had 26,521 pieces of promotional material, a 38.61% increase from the previous month; Kimi had 14,451 pieces, a 160.71% increase; and Zhipu Qingyan had 1,866 pieces, a staggering 2,456.16% increase.

Despite variations in growth rates and distribution platforms, a clear trend emerges: the battlefield for competition among products like Doubao and Kimi is witnessing new entrants. The competition for user acquisition through promotional material distribution is merely a superficial application-level phenomenon. Similar to past subsidy-burning strategies during the thousand-group wars, the fundamental aim on the user side is conversion.

Among Chinese companies with the most comprehensive generative AI models and AI applications, Zhang Yiming pins his hopes for large model technology's commercial application in the hardware sector.

From educational desk lamps to VR headsets like PICO, Zhang Yiming's renewed enthusiasm for IoT has now shifted to AI hardware.

As of late November, the Doubao APP had surpassed 160 million downloads, with nearly 60 million monthly active users and almost 6 million daily active users, positioning it as one of China's leading AI applications, second only to ChatGPT.

However, with Toutiao's user base largely stabilized and Douyin's user growth slowing, the Doubao APP must grow rapidly and has thus begun investing in traffic on platforms like Bilibili and Xiaohongshu.

For Doubao, amid fierce competition, AI hardware has undeniably become a new traffic inlet.

Currently, ByteDance's exploration in AI hardware spans two paths: one is the product line codenamed [D-line], led by Li Haoqian; the other is the [O-line], helmed by the founder of a previously acquired company, who reports directly to Technical Vice President Hong Dingkun.

The [O-line] comprises a cross-organizational team that not only oversees ByteDance's AI hardware exploration but also maintains close cooperation with ByteDance's AI application team Flow.

AI headphones possess greater market potential than AI glasses.

Recently, a research report by Huajin Securities revealed that Shenzhen Dashifuture Technology Co., Ltd. (Oladance), a leading enterprise in OWS technology, has been fully acquired by ByteDance and officially integrated into its fold.

As the inaugural AI hardware product based on the Doubao large model, the launch of Ola Friend signifies a further advancement in ByteDance's AI ecosystem construction, extending from hardware to content.

The hardware component of the Ola Friend headphones is handled by the former Oladance team, recently acquired by ByteDance and renowned for its innovative OWS patent technology.

Although this acquisition did not cause significant market fluctuations, the industry generally believes that ByteDance's intentions are clear—to deeply integrate AI technology with headphone products.

In the race for new technology entry points, mobile phone manufacturers typically lead the charge, followed by super apps like WeChat.

In other words, once AI assistant technology reaches a highly mature market fit (PMF) stage, mobile phone manufacturers and these super app giants will swiftly join the fray.

Therefore, ByteDance's launch of Ola Friend can be viewed as a proactive move to seize the initiative.

Conclusion:

For leading enterprises in the technology industry, the development of AI smart hardware presents new strategic opportunities.

The competition in AI smart hardware has already commenced, with these technology giants actively deploying strategies.

When AI technology converges with traditional products, it demonstrates robust vitality and energy, making AI utilization for transformation and upgrading a crucial path for technology giants' future development.

Leveraging their strong technical capabilities, vast user bases, and abundant market resources, technology giants are poised to take the lead in the field of AI smart hardware. By consistently introducing innovative products and solutions, they can further consolidate their market position, expand their business scope, and achieve sustainable development.

References: STAR Market Daily: "Meta AI Glasses, Downloads Surge 200%", GeekPark: "ByteDance's First AI Hardware: A Quiet Trial?", X Augmented Reality: "Ambitions for AI Wearable Hardware Remain Before Smart Glasses Capture the Market, ByteDance Steps Up Its Involvement", LatePost: "ByteDance Tries Hardware Again: Exploring AI Headphones, Glasses, and Other Products, Synergizing with the Doubao Large Model", Zimu Bang: "Zhang Yiming Finally 'Collides' with Lei Jun", KrAsia: "ByteDance 'Combines Software and Hardware', Deploys AI + Hardware, and Drives Full Ecosystem Construction", Guangzi Planet: "ByteDance AI, 'Combining Software and Hardware'", Narrowcast: "Will AI Hardware Alleviate ByteDance's FOMO?", Spicy Chopped Peppers: "Instead of Stubbornly Focusing on Glasses, Why Did ByteDance Choose to Make AI Headphones?", Tai Finance: "AI Smart Hardware, the Next Trillion-Dollar Opportunity?"