Open Source Securities: AI Technology Permeating to the Terminal Acoustic Hardware Industry Chain Seizes the Opportunity of Quantity and Price Increase

Open-source Securities released a research report stating that as generative AI technologies such as DeepSeek penetrate the end market, voice interaction is evolving from "passive response" to "active intelligence." According to Canalys, the global shipment volume of personal intelligent audio devices is expected to reach 533 million units by 2025, an 8% year-on-year increase. The increase in interaction demand will drive hardware upgrades, leading to an opportunity for the acoustics hardware industry chain to see simultaneous increases in quantity and price. End-side AI audio devices bring real-time noise reduction, AI voice interaction, and other features, enhancing consumer experience and potentially driving the upgrade of existing devices. Recommended investment targets include: Shanghai Awinic Technology (688798.SH) and MEMSensing Microsystems (Suzhou, China) Co., Ltd. (688286.SH). Key points from Open-source Securities are as follows: AI interaction capabilities require matching hardware upgrades. In the AI era, audio terminals such as headphones are no longer just audio playback devices but are being enhanced with audio functions, noise cancellation, and improved sound quality to become crucial mediums for human-machine interaction. Systematic upgrades in hardware capabilities involve reconstructing the "sensory nervous system" of headphones in terms of signal acquisition and information processing, which requires breaking through the performance thresholds of traditional devices. Upgrades are focused on two main directions: (1) Signal acquisition - High signal-to-noise ratio: AI-driven active noise cancellation and voice interaction rely on high-precision sound-field modeling, and headphones with high signal-to-noise ratios can support interactions in noisy environments, expand application scenarios, and improve user experience. According to Omdia, the compound annual growth rate for MEMS microphones with a signal-to-noise ratio higher than 64dB in the consumer sector is expected to reach 8.7%, with sales volumes approaching 3 billion units by 2027. (2) Information processing - High computational power and low power consumption: AI-driven real-time audio processing sets new requirements for computational power, necessitating advancements beyond the capabilities of traditional DSPs, such as integrating NPUs to enhance efficiency. Furthermore, improved computational power must be accompanied by reduced power consumption to ensure battery life. AI-side innovations are driving a new cycle of consumer electronics demand. The penetration of generative AI into the end market may reshape audio terminal demands. On one hand, AI audio devices on the end side bring features such as real-time noise reduction and AI voice interaction, enhancing the consumer experience and potentially driving the upgrade and replacement of existing devices. According to RUNTO Technology data, in August 2024, although AI headphones accounted for only 1.4% of total headphone sales on Chinese online e-commerce platforms, their growth was remarkable, with sales volume increasing by 763.3% compared to the same period last year, and sales revenue increasing nearly 14.5 times. It is predicted that by 2024, the sales volume of AI headphones in the Chinese e-commerce market is expected to exceed 200,000 pairs, with a year-on-year growth rate estimated at 488.7%. On the other hand, AI interaction entrances are expanding into various forms, and the AI glasses industry is entering a period of intensive new product releases. AI glasses, as wearable devices that integrate visual, auditory, and other important perception and interaction methods, are poised to become important new products in the "artificial intelligence + consumer" market and may become the next consumer electronics terminal to achieve a shipment volume exceeding one billion units. Beneficiary targets: Shanghai Awinic Technology (688798.SH): A global leader in audio amplifier chip supply, such as its flagship product AW88396for Digital Smart K, which integrates rich sound effects and protection algorithms DSP, has extremely low background noise and static power consumption, and combines algorithm innovation to achieve high-quality audio experiences. MEMSensing Microsystems (Suzhou, China) Co., Ltd. (688286.SH): The third-largest MEMS microphone shipment volume globally, high signal-to-noise ratio microphones with a 68dB signal-to-noise ratio have begun sampling, with mass production scheduled to begin in 2024. Products with a 70dB signal-to-noise ratio are at the sampling stage, and the company is actively discussing potential application scenarios and selecting new products with industry-leading customers. Risk warning: Lower-than-expected downstream demand and slower-than-expected technological iterations.