Preface:

By the end of 2025, the dense (intensive) release of next-generation human-computer interaction interfaces represented by AI glasses marks the eve of the explosion from concept to large-scale application of on-device AI hardware. This is not just a hardware competition in the 'Hundred-Glasses Battle,' but a comprehensive war that reconstructs the chip and industry chain with real-world data as its core.

The 'Hundred-Glasses Battle' Erupts: A Trillion-Dollar Market Emerges

The tech circle has been abuzz with a 'Hundred-Glasses Battle' recently: Baidu has released its native AI glasses equipped with a Chinese large model, Alibaba's Kuake glasses integrate full-ecosystem services, Li Auto has entered the fray with its car-mirror linkage (linkage) scheme, and professional manufacturers like Rokid and Thunderbird are doubling down on 'AI+AR' technology... According to statistics, nearly 70 manufacturers have intensively launched new products, ranging from internet giants to smart car companies, from consumer electronics behemoths to cross-border newcomers, all vying for a slice of this impending trillion-dollar pie. IDC data shows that global smart glasses shipments reached 4.065 million units in the first half of 2025, surging 64.2% year-on-year. During the Double 11 shopping festival, category transaction volume on Tmall skyrocketed by 2500%, indicating that the industry has entered a vibrant stage akin to the 'feature phone late (end)' era.

The participants in this battle may seem diverse, but they can be categorized into five major camps, each with a clear core logic. Internet giants like Alibaba, Baidu, and ByteDance have the longest vision, viewing AI glasses as the 'ultimate entrance (gateway)' for large models and a 'data goldmine,' attempting to close the ecosystem loop with hardware. Alibaba's Kuake AI glasses exemplify this approach, featuring a dual-chip architecture of Qualcomm AR1 + BES2800 from Bestechnic, integrating full-link services like Taobao object recognition, Gaode navigation, and Alipay payments, complemented by a dual-battery balancing solution from Southchip Technology for 24-hour battery life, directly enhancing user experience with ecosystem advantages.

Consumer electronics manufacturers are seeking breakthroughs while on the defensive. Xiaomi and Huawei, for instance, regard AI glasses as an extension of smartphone functions, serving as both a second growth curve and a 'moat' for their hardware territory. Smart car companies have even more forward-looking ambitions; Li Auto's layout (layout) of car-mirror linkage aims to collect 'real-world data'—a core asset for future AI competition. Professional AR/VR manufacturers like Rokid and Thunderbird rely on technological prowess, leveraging accumulated expertise in waveguide displays to advance 'AI+AR.' Thunderbird X3 Pro's full-color arrayed waveguide technology achieves a balance between lightweight design and display performance. Cross-border players like Insta360 target niche segments like sports and security, with products like Light Intelligence Technology's smart swimming goggles gaining market traction through professional waterproofing and precise motion data.

Chip Industry Chain Restructuring: On-Device AI Becomes a New Growth Engine

The transformation of AI glasses from 'toys' to 'tools' hinges on their role as 'portable AI gateways.' In scenarios like sports, driving, and maintenance, their hands-free, real-time interaction capabilities are irreplaceable by smartphones. Functions like real-time translation, first-person navigation, and meeting minutes have already materialized. INMO Technology's translation glasses achieve 'bidirectional dialogue' closure through clip-on microphones, while engineers in industrial settings can perform precise maintenance via AR annotations. These scenarios are driving demand from novelty to necessity.

The demand explosion is directly reshaping the supply chain, offering opportunities across the entire chip supply chain. As the 'brain,' AI SoC chips have entered a dual-chip era, with Qualcomm AR1 paired with BES2800 coprocessors from Bestechnic becoming standard in high-end models. Industry trends favor 6nm processes and stronger NPUs. Memory chip demand is surging; Biwin Storage's ePOP series has penetrated supply chains of leading companies like Meta and Xiaomi, with AI glasses-related revenue projected to soar over 500% year-on-year by 2025. Power management chips are crucial for battery life; Southchip Technology's custom dual-battery balancing solution for Kuake AI glasses enables 24-hour operation. Sensors like bone conduction microphones and waveguide driver chips are creating new incremental markets. Notably, domestic manufacturers like Bestechnic and Actions Semiconductor are leveraging their IoT expertise to gain a favorable position in the on-device AI track (track), entering a growth window.

More critical than hardware is the ecosystem battle. Competition has shifted beyond parameter comparisons to ecosystem construction based on AI agents. The traditional 'app store' model is becoming obsolete; future users will simply need to speak a command, and the system will automatically orchestrate services like search, payment, and navigation to achieve 'service-to-user' delivery. Alibaba's ecosystem integration and Baidu's integration of a Chinese large model are early layout (moves) in this ecosystem war.

From Geek Toys to Mass-Market Staples: How Many Hurdles Remain?

Despite the soaring enthusiasm, AI glasses will take time to become a mass-market staple. The industry consensus is that 2028 may mark a critical inflection point, when non-display functional models need to drop below 1,000 yuan and AR glasses with displays need to fall below 2,000 yuan to trigger mass adoption. Technologically, achieving 30-gram weight and 8-hour battery life are core thresholds—current mainstream models weigh around 50 grams, but companies like VeriSilicon are tackling lightweight design through ultra-low-power GPU IP technologies, while Nika Optics' 0.6mm-thick volumetric holographic waveguide lenses offer weight reduction possibilities.

The market prospects are promising; IDC predicts global shipments will exceed 40 million units by 2029, with a CAGR of 55.6% from 2024 to 2029. However, bottlenecks persist: technologically, the 'impossible trinity' of weight, battery life, and display performance needs optimization, and AR display-induced dizziness remains unresolved. On the supply chain side, low waveguide yields drive up costs; although Gudong Technology has reduced single-unit costs to $10-30, they are still out of reach for the mass market. Ecosystem-wise, a lack of developers has led to a scarcity of applications, hindering user stickiness. More daunting are privacy concerns; continuous data collection by first-person devices has sparked trust crises, as seen in the backlash against Meta Aria Gen 2.

From the influx of nearly 70 manufacturers to the collective frenzy in the chip industry chain, the trillion-dollar market for AI glasses is emerging. The Greater Bay Area has formed a complete 'R&D-module-manufacturing' chain, with Shenzhen's 24-hour component sourcing capabilities accelerating innovation. However, this battle is destined to be a protracted one; the 'iPhone moment' for AI glasses may require 5-10 years of technological accumulation and ecosystem cultivation.

Epilogue

The 'Hundred-Glasses Battle' for AI glasses is far from the finish line; it is merely the starting point of the on-device AI hardware revolution. As the chip industry faces growth bottlenecks in the smartphone market, AI glasses have opened up a new realm of imagination for the semiconductor industry. For investors, the on-device AI chip track (track) is entering a 'golden window period'; for consumers, around 2028, we may witness the transformation of AI glasses from 'tech toys' to 'life essentials.'

Online References:

36kr: 'Not Just Li Auto and Kuake: Latest News on AI Glasses from 7 Major Manufacturers | Exclusive from Smart Emergence'

Shanghai Securities News: 'Decisive Battle on the Device Side: AI Applications Accelerate 'Hard' Landing'

TMTPost: 'AI Glasses Manufacturers Are Collectively Refusing to Pay 'IQ Taxes''

Zhitong Finance: 'CITIC Construction Investment: On-Device AI Explosion Expected; Domestic High-End Capacity Urgently Needs Breakthrough'