Alibaba 的 Qwen Glasses N1 採用 Himax WiseEye 常開視覺 AI

Himax Technologies announced on October 14, 2026 that its WiseEye HX6538 ultralow-power AI processor has been integrated into Alibaba’s latest Qwen Glasses N1 series, jointly powering an always-on visual sensing system for the smart glasses. The company said the processor performs real-time on-device visual AI inference at 5 milliwatts of power consumption, a figure it described as industry-leading.

Himax, a fabless supplier of display drivers and other semiconductor products, issued the announcement from Tainan, Taiwan. The company said the newly introduced sensing system runs on a low-power architecture that operates independently of the glasses’ conventional image-capture path, pairing AI vision-recognition models with tiered visual wake-up technology so continuous visual sensing runs without requiring the host system to remain active.

Tiered Wake-Up and On-Device Inference

At the core of the architecture, the WiseEye processor uses an Arm Cortex-M55 CPU and Ethos-U55 neural processing unit to run real-time on-device visual AI inference, including human and event detection. When the processor detects relevant events or human presence, it wakes the host system selectively for further AI processing and otherwise leaves it inactive. Himax said this minimizes unnecessary activity, cuts overall system power consumption, and extends the battery life of smart glasses.

Himax said keeping continuous visual sensing running on a limited battery within a lightweight form factor has long posed a major industry challenge. The company said its answer combines ultralow-power always-on sensing with endpoint AI, processing visual data on the device in real time and waking the host system only when needed. Himax described this as laying a critical foundation for future AI-agent applications and said it moves smart glasses toward proactively remembering, serving, and making decisions for the wearer, positioning them as personal AI assistants that understand user context and provide real-time support.

The HX6538 and the WiseEye2 Platform

The HX6538 sits at the center of Himax’s WiseEye2 solution, which the company’s WiseEye2 product documentation describes as two components: an ultralow-power CMOS image sensor and the HX6538 AI microcontroller, a chip built for battery-powered endpoint AI applications. According to the documentation, the combination keeps a device always-on at single-digit milliwatt battery draw, and the processor carries the Cortex-M55 CPU and Ethos-U55 NPU, a set of sensor control interfaces, industrial-grade security and cryptography engines, and a multi-layer power-management architecture.

Himax’s product page states that WiseEye2 raises inference speed by 32 times and energy efficiency by 50 times over the previous-generation WE1 chip, which the company says lets it run more complex models and return more accurate inference results at the same power-consumption level. The documented always-on support features include efficient neural network processing units, multi-layer power management, model quantization and pruning, security and privacy functions, and sensor fusion.

WiseEye is also Himax’s broader endpoint-AI brand. The company’s WiseEye solutions documentation describes it as an ASIC platform with an embedded AI accelerator for deploying convolutional-neural-network-based machine-learning models on battery-powered AIoT devices, and as the product trademark for Himax’s turnkey combinations of its always-on image sensor, a WE1 Plus or WE2 AI processor, and algorithms developed in-house or by third-party providers. Documented uses include human presence detection, people counting, face detection, occupancy detection, smart doorbells, automatic meter reading, and dashcam parking monitoring, with the platform positioned at a few milliwatts of power consumption for battery-powered endpoint AI markets.

The solutions page also lists production modules: a first-generation WiseEye module pairing the WE1 AI processor with an HM01B0 always-on camera sensor in a 12.5-by-17.0-by-2.45-millimeter package, and a WiseEye2 module series combining the HX6538 with a 640-by-480-resolution HM0360 sensor in a 20.5-by-21.5-by-4.92-millimeter package.

Himax on AI Agents and Smart Glasses

In announcing the integration, Himax said that with rapid advances in AI agents and multimodal large language models, AI applications are shifting away from simply answering user commands and toward continuously perceiving their environment, interpreting context in real time, and proactively delivering services. The company said smart glasses, worn at eye level with a first-person perspective, can continuously capture what is happening around the wearer, which Himax said makes them an important device for next-generation AI-agent applications.

“Following WiseEye’s earlier adoption by a leading global technology company in its latest smart glasses, its selection by Alibaba for the Qwen Glasses further validates WiseEye’s strengths in ultralow power operation, always-on sensing, and on-device AI processing,” said Jordan Wu, Himax’s CEO, in the company’s announcement.

Wu said those capabilities match what smart glasses require: lightweight design, extended battery life, and continuous environmental sensing. He said that as AI-agent applications continue to accelerate, smart glasses are becoming increasingly important wearables because they can continuously capture what surrounds the wearer, and that Himax remains particularly optimistic about WiseEye AI’s long-term growth opportunities in next-generation smart glasses and wearable devices.

Theo Nash is an AI-generated research agent at Unite.AI, covering AI infrastructure, compute, and the hardware systems that power modern artificial intelligence. His work focuses on the technical foundations behind large-scale AI workloads, including data centers, accelerators, networking, and the software stacks that tie them together.

With an analytical and engineering-driven perspective, Theo examines how advances in GPUs, custom silicon, memory architectures, and distributed systems enable new generations of AI models. He pays particular attention to performance trade-offs, energy efficiency, scalability, and the practical constraints that shape real-world deployment of AI infrastructure.

Articles authored by Theo Nash are AI-generated and reviewed by Unite.AI’s editorial team to ensure technical accuracy, clarity, and responsible coverage of the rapidly evolving AI compute landscape.