[骁龙峰会] 高通发布骁龙音效Elite Gen 2
高通在骁龙峰会上发布下一代音频可穿戴平台骁龙音效Elite Gen 2。高通XR、可穿戴与个人AI业务高级副总裁兼总经理Ziad Asghar出席发布。该平台面向音频可穿戴设备，属于芯片原厂的新品发布，与元器件供应链中的可穿戴芯片环节相关。
原文深度解读
本文报道高通在骁龙峰会2026上发布下一代音频可穿戴平台骁龙音效Elite Gen 2（Snapdragon Sound Elite Gen 2）。高通XR、可穿戴与个人AI业务高级副总裁兼总经理Ziad Asghar于9月23日在毛伊岛介绍该平台，称其为面向个人AI时代的首个高端音频平台。该平台为系统级芯片（SoC），面向无线耳机、头戴式耳机和音频眼镜等形态，集成端侧AI处理、云端连接AI能力、微功耗Wi-Fi 6E、aptX Lossless、XPAN、第五代自适应ANC等技术。关键数字包括：AI算力较上一代骁龙S7 Gen 1音效平台提升最高两倍，峰值AI性能达128 GOPS（250 MHz、256 MACs），上一代为64 GOPS和128 MACs；Wi-Fi连接空闲模式目标电流0.6 mA，较上一代1.0 mA降低最高40%；SoC合计面积由44.22 mm²缩小最高30%至31.18 mm²。高通提供参考设计，并与Cleer、Bose、HP合作。该发布属于可穿戴芯片原厂新品，与电子元器件供应链中的可穿戴SoC、Wi-Fi 6E、音频编解码、麦克风与传感器等环节相关。
- 高通发布骁龙音效Elite Gen 2，定位为面向个人AI时代的首个高端音频平台，由Ziad Asghar于2026年9月23日在毛伊岛骁龙峰会2026上介绍。
- 该平台是SoC，支持无线耳机、头戴式耳机和音频眼镜等智能音频可穿戴（hearables）形态，集成高端音频技术、端侧AI处理、面向安全的云端连接AI能力和微功耗Wi-Fi 6E。
- AI算力较上一代骁龙S7 Gen 1音效平台提升最高两倍：峰值AI性能128 GOPS（250 MHz、256 MACs），上一代为64 GOPS和128 MACs，可同时运行多个端侧AI工作负载而不延迟。
- 功耗与芯片尺寸下降：Wi-Fi连接空闲模式目标电流0.6 mA，较上一代1.0 mA降低最高40%；SoC合计面积由44.22 mm²缩小最高30%至31.18 mm²，便于在耳机和眼镜等紧凑产品中加入个人AI功能。
- Snapdragon Sound Apps & Agents软件与AI代理生态通过Wi-Fi 6E直连云端，无需经智能手机路由，音频可穿戴可独立访问AI助手、流媒体等服务；平台还包含实时调音技术、aptX Lossless、XPAN、第五代自适应ANC、cVc、多麦克风架构和Snapdragon Audio Sense Solution。
- 高通向制造商提供参考设计以快速构建原型并验证产品商业化，活动中预览了与Cleer合作开发的智能音频可穿戴产品，并正与Bose和HP合作；Asghar称研究显示最高70%用户希望在可穿戴设备中获得先进AI能力。
供应链影响
- 该平台将AI算力、Wi-Fi 6E和音频处理集成于单颗SoC，可能推动可穿戴设备对高集成度芯片、微功耗Wi-Fi 6E相关射频与电源管理元器件的需求结构变化，具体取决于终端厂商采用节奏。
- SoC面积缩小最高30%、空闲电流降低最高40%，可能使耳机和音频眼镜等紧凑形态更易导入个人AI功能，从而可能影响麦克风、传感器和音频编解码等外围元器件的用量与规格选择。
- 平台支持通过Wi-Fi 6E直连云端而无需经智能手机路由，可能改变音频可穿戴的连接架构，进而可能影响蓝牙/Wi-Fi组合芯片、天线和射频前端等环节的供应链配置。
- 高通提供参考设计并与Cleer、Bose、HP合作，可能加快采用该平台的终端产品开发与验证节奏，从而可能带动相关元器件在可穿戴SoC生态中的配套需求，但实际放量取决于终端产品上市与市场接受度。
- 原文称最高70%用户希望在可穿戴设备中获得先进AI能力，若该需求成立，可能推动可穿戴芯片向更高端AI与音频能力演进，进而可能影响上游晶圆、封测和音频器件等环节的产能与规格分布，但存在不确定性。
系统解读，仅作为判断线索，不构成备货、出货、涨价或投资建议。
Ziad Asghar，高通高级副总裁，发布骁龙音效Elite Gen 2。（照片：THE ELEC）
高通发布了骁龙音效Elite Gen 2（Snapdragon Sound Elite Gen 2），这是其下一代音频可穿戴平台。
高通XR、可穿戴与个人AI业务高级副总裁兼总经理Ziad Asghar于9月23日在毛伊岛举行的骁龙峰会2026上介绍了该平台。“我们正在看到由AI代理、音频和用户语音融合形成的一类新型智能可穿戴听觉设备（hearables）的开始，”他说。“骁龙音效Elite Gen 2是首个为个人AI时代设计的高端音频平台。”
骁龙音效Elite Gen 2是一款系统级芯片（SoC），旨在支持智能音频可穿戴设备，即hearables，覆盖无线耳机、头戴式耳机和音频眼镜等一系列形态。
该平台集成了高端音频技术、端侧AI处理、面向安全的云端连接AI能力以及微功耗Wi-Fi 6E。设备可以收集语音信号和周围数据，快速分析并自然响应用户需求。
AI计算能力较上一代骁龙S7 Gen 1音效平台提升最高两倍。峰值AI性能在250 MHz下达到每秒1280亿次运算，即128 GOPS，每时钟256次乘加运算，即256 MACs。这使多个端侧AI工作负载能够同时运行而不延迟。上一代提供64 GOPS和128 MACs。
功耗和芯片尺寸也有所降低。Wi-Fi连接空闲模式下的目标电流消耗为0.6 mA，较上一代的1.0 mA降低最高40%。SoC合计面积由44.22 mm²缩小最高30%至31.18 mm²。更小的芯片和更低的功耗要求使设备制造商更容易将个人AI功能添加到耳机和眼镜等紧凑产品中。
面向无线hearables的骁龙音效Apps & Agents软件和AI代理生态通过Wi-Fi直接连接云端。（照片：THE ELEC）
面向无线hearables的骁龙音效Apps & Agents软件和AI代理生态通过Wi-Fi 6E直接连接云端，无需通过智能手机路由连接。音频可穿戴设备可以独立访问AI助手、流媒体服务和其他内容。
该平台还包含可根据听者及周围环境变化实时调整声音的技术。它集成了高通aptX Lossless——一种旨在提供高分辨率无损音频且无失真的蓝牙音频编解码器，以及高通扩展个人局域网（XPAN），后者将连接范围显著扩展到传统蓝牙范围之外。XPAN可实现向微功耗Wi-Fi连接的无缝切换。
噪声控制和通话质量技术也得到了改进。该平台集成了第五代高通自适应主动降噪（ANC）、高通清晰语音捕获（cVc）技术和先进的多麦克风架构。这些技术可消除环境噪声并更清晰地分离佩戴者的语音。它还包含骁龙音频感知解决方案（Snapdragon Audio Sense Solution），该方案将高性能麦克风与感知算法相结合，以精确识别人声和环境声音，提高AI助手的准确性和响应速度。
高通正在向制造商提供参考设计，以帮助他们快速构建原型并验证产品以进行商业化。在活动中，高通预览了一款与Cleer合作使用该平台开发的智能音频可穿戴设备。它还在与Bose和HP合作。
“我们的研究显示，最高70%的用户希望在可穿戴设备中获得先进的AI能力，”Asghar说。“骁龙音效Elite Gen 2通过实现实时翻译、情境辅助、个性化听力和同时处理语音与视觉信息的多模态能力，为音频可穿戴设备开辟了新的可能性。”
Ziad Asghar, senior vice president at Qualcomm, unveils Snapdragon Sound Elite Gen 2. (Photo: THE ELEC)
Qualcomm has unveiled Snapdragon Sound Elite Gen 2, its next-generation audio wearable platform.
Ziad Asghar, senior vice president and general manager of XR, Wearables and Personal AI at Qualcomm, introduced the platform at Snapdragon Summit 2026 in Maui on Sept. 23. “We are seeing the beginning of a new category of intelligent hearables formed by the convergence of AI agents, audio and the user’s voice,” he said. “Snapdragon Sound Elite Gen 2 is the first premium audio platform designed for the Personal AI era.”
Snapdragon Sound Elite Gen 2 is a system-on-chip (SoC) designed to support intelligent audio wearables, or hearables, across a range of form factors including wireless earbuds, headphones and audio glasses.
The platform integrates premium audio technologies, on-device AI processing, security-focused cloud-connected AI capabilities and micro-power Wi-Fi 6E. Devices can collect voice signals and surrounding data, quickly analyze them and naturally respond to users’ needs.
AI computing capability has increased by as much as twofold from the previous Snapdragon S7 Gen 1 Sound Platform. Peak AI performance reaches 128 billion operations per second, or 128 GOPS, at 250 MHz, with 256 multiply-accumulate operations per clock, or 256 MACs. This enables multiple on-device AI workloads to run simultaneously without delays. The previous generation delivered 64 GOPS and 128 MACs.
Power consumption and chip size have also been reduced. Target current consumption in Wi-Fi-connected idle mode is 0.6 mA, up to 40% lower than the previous generation’s 1.0 mA. The combined SoC footprint has been reduced by as much as 30% to 31.18 mm² from 44.22 mm². The smaller chip and lower power requirements make it easier for device makers to add Personal AI features to compact products such as earbuds and glasses.
The Snapdragon Sound Apps & Agents software and AI agent ecosystem for wireless hearables connects directly to the cloud through Wi-Fi. (Photo: THE ELEC)
The Snapdragon Sound Apps & Agents software and AI agent ecosystem for wireless hearables connects directly to the cloud through Wi-Fi 6E, eliminating the need to route connections through a smartphone. Audio wearables can independently access AI assistants, streaming services and other content.
The platform also includes technology that adjusts sound in real time to the listener and changes in the surrounding environment. It incorporates Qualcomm aptX Lossless, a Bluetooth audio codec designed to deliver high-resolution lossless audio without distortion, as well as Qualcomm Expanded Personal Area Network (XPAN), which significantly extends connectivity beyond conventional Bluetooth range. XPAN enables seamless switching to micro-power Wi-Fi connectivity.
Noise control and call-quality technologies have also been improved. The platform incorporates fifth-generation Qualcomm Adaptive Active Noise Cancellation (ANC), Qualcomm Clear Voice Capture (cVc) technology and advanced multi-microphone architectures. These technologies cancel ambient noise and isolate the wearer’s voice more clearly. It also includes the Snapdragon Audio Sense Solution, which combines high-performance microphones with sensing algorithms to precisely identify human voices and environmental sounds, improving the accuracy and responsiveness of AI assistants.
Qualcomm is providing reference designs to manufacturers to help them rapidly build prototypes and validate products for commercialization. At the event, Qualcomm previewed an intelligent audio wearable developed with Cleer using the platform. It is also working with Bose and HP.
“Our research shows that up to 70% of users want advanced AI capabilities in wearable devices,” Asghar said. “Snapdragon Sound Elite Gen 2 opens new possibilities for audio wearables by enabling real-time translation, contextual assistance, personalized hearing and multimodal capabilities that process voice and visual information simultaneously.”