Alibaba Launched an AI-Powered Glasses

 

Alibaba's Head Office Image
Image Credit: Unsplash by Zhang Hui

Alibaba Launched an AI-Powered Glasses 

China's e-commerce giant Alibaba has launched its Quark AI Glasses to compete with Meta's Ray-Ban smart glasses. While Ray-Ban Meta glasses were designed for social media enthusiasts and content creators, Alibaba has a different focus for its AI-powered glasses. The glasses are being designed for Chinese tech business users and professionals within the Chinese tech giant's extensive ecosystem. These AI-powered glasses are Alibaba's first attempt in the emerging AI-driven device market. Meanwhile, HTF Market Intelligence reports that the wearable technology sector is projected to expand at a compound annual growth rate (CAGR) of 12.40% from 2024 to 2032. The market is expected to grow by 13.00% year-on-year, with North America leading the market with a value of 74.8 billion in 2024. This growth is in line with Alibaba's commitment to AI and cloud infrastructure development. The worldwide wearable technology market is seeing significant growth, expected to rise from around $72.5 billion in 2024 to more than $200 billion by 2033. This expansion is fueled by increasing demand for products such as smartwatches and fitness trackers, particularly in key regions including North America, Asia Pacific, and Europe. The market is divided into categories based on products (such as smartwatches and fitness trackers) and applications (like healthcare and fitness), with prominent companies including Apple, Google, and Samsung at the forefront.

On July 28, 2025, Alibaba launched the Quark AI Glasses, signaling the start of a new era of its AI-powered products. The main features of the Alibaba Quark smart glasses are as follows:

The Alibaba AI Glasses are powered by the Quark AI assistant and Qwen language model, offering two key functionalities: hands-free and voice-activated operation, as well as other features including :

Real-Time Translation: The glasses can translate conversations and texts in different languages, making them ideal for communication.

Audio and Communication: The AI-powered glasses include off-ear speakers and a multi-microphone array for clear, noise-free audio and calls.

Meeting Transcription: This feature converts spoken conversations into notes for professional and business settings.

Built-In Camera: The glasses feature a 12MP Sony sensor camera with a low-light mode for taking photos and videos.

Ecosystem Integration: The glasses are extensively integrated with Alibaba's main services, enabling users to:

• Navigate using AutoNavi Map with augmented reality directions.

• Make payments with Alipay's video payment features.

• Shop on Taobao by utilizing video product recognition and price comparison.

• Receive travel alerts from Fliggy.

Available in Two Models: A high-end AI+AR version, which has a micro-LED display for improved video, and a standard AI-powered version.

Dual-Chip Architecture: The glasses feature a Qualcomm Snapdragon AR1 processor and a BES2800 Bestechnic chip for advanced performance and long-lasting battery.

Pre-Order Available: The glasses could be pre-ordered in China from late October 2025, with shipping beginning in December.

Alibaba Partnership With Qualcomm on Quark Smart Glasses Processor

Alibaba had partnered Qualcomm for its Quark Smart Glasses processors. The glasses used dual-chip architecture featuring:

Qualcomm's AR1 Chip: A processor designed for compact smart glasses.

Bestechnic's BES2800 Chip: Used to balance processing performance and power efficiency. 

Qualcomm's AR1 Gen 1 Chip

The Qualcomm Snapdragon AR1 Gen 1 is a processor created for compact smart/AR glasses, allowing for low energy consumption. Its features include high-end camera support, an AI system, and a fast connection for smart glasses.

Designed for Smart Glasses: The AR1 is the first Snapdragon processor exclusively created for smart glasses, enhancing them to be light and practical.

Features:

• Multiple independent CPUs and GPUs

• A third-generation Hexagon Neural Processing Unit (NPU) that processes AI data directly on the device

• Dual Image Signal Processors (ISPs) for photo and video capture

• Support for Bluetooth 5.3 and Wi-Fi 7

Performance: It can process multiple computing tasks, dynamic images processing, and intelligent functions such as live translation and video search.

Power Efficiency: The processor is developed using advanced process technology, making it light and power-efficient as well as ideal for compact and battery-operated devices.

Applications: Ray-Ban Meta Smart Glasses also use AR1 processors.

Bestechnic's BES2800 Chip

The BES2800 chip is an energy-efficient, high-performance SoC (System-on-a-Chip) developed by the Chinese company Bestechnic. It is mainly used in smart wearables.

BES2800 Key Features

The BES2800 features multiple processors and wireless connectivity to provide a powerful yet efficient platform for wearable devices. 

Dual-Core Processing: It includes a dual-core Arm Cortex-M55 CPU subsystem and a dual-core BECO neural processing unit (NPU).

Low Power Consumption: The chip is fabricated on a 6nm process, which helps to substantially reduce power consumption and increase battery life.

Wireless Connectivity: It supports dual-mode Bluetooth 5.4 (including LE Audio) and Wi-Fi 6 for robust and high-speed wireless connections.

Multimedia and Graphics: Some versions, like the BES2800BP, feature a 2.5D GPU for advanced graphics and a display controller that supports up to 720p resolution.

Superior Performance: Compared to its earlier model, the BES2700, the BES2800 offers high processing power boosts, with doubled CPU computing power and quadrupled NPU power. 

Used in Other Applications: Beyond Alibaba's smart glasses, the capability of the BES2800 makes it ideal for a variety of wearable products. 

Smartwatches and Fitness Trackers: The chip can process health algorithms (e.g., heart rate, blood oxygen) significantly while managing power efficiently to achieve long-lasting battery life.

Wireless Headphones and Earbuds: It is developed for use in TWS (True Wireless Stereo) earbuds, including features like active noise cancellation (ANC) and low-latency audio transmission.

Other Audio Devices: The chip's powerful audio processing capacities are also used in AI speakers and other portable audio equipment. 

Building on their partnership, Qualcomm and Alibaba collaborated on automotive technology, using the Snapdragon 8397 for an AI-powered vehicle. At the World Artificial Intelligence Conference (WAIC) 2025, Alibaba announced a new AI cockpit for vehicles, developed with Qualcomm and AI cockpit solution provider Banma. For Alibaba's customized AI vehicles, the in-house Qwen large multimodal model was used alongside Qualcomm's Snapdragon 8397 platform. This provides a more natural, intuitive, and proactive in-car AI experience, enabling drivers to use voice commands for tasks like navigation and restaurant reservations.

Additionally, while Alibaba had previously launched open-source AI models, the company announced a significant expansion on Thursday, September 19, 2024. According to CNBC, Alibaba unveiled over 100 new open-source AI models, known as Qwen 2.5, to strengthen its competitive position against rivals. The new models, which utilize the company's enhanced proprietary technology, offer improved capabilities in mathematics and coding and are intended for various applications and industries, such as automobiles, gaming, and scientific research. The company has committed to a multibillion-dollar, multi-year investment plan for AI development and cloud infrastructure. Details about the Qwen models include the following:

Qwen AI Models: Features, Capabilities, and Resources

Qwen is a collection of large language and cross-modal models created by Alibaba Cloud. The Qwen models are pre-trained on a large dataset that consists of multimodal and multilingual data. The models are capable of comprehending and generating images, texts, and audio, as well as performing logical reasoning and acting as an intelligent agent. The main features include multilingual support, cross-modal capabilities, digital image editing and creation, and the ability to combine different thinking methods to solve problems by dynamically switching between methodical and analytical thinking. 

Key Features of Qwen Models

Multilingual and Cross-modal: Qwen models are trained on large datasets in many languages and can process and comprehend different types of data, including images and videos.

Digital Image Creation and Editing: A model like Qwen-Image can create photorealistic images and perform multiple edits with great accuracy, including image rendition with text and neural style transfer.

Hybrid Thinking: Different versions of the model can automatically switch between a detailed, sequence-thinking mode and a faster, unthinking mode for various tasks, such as complex reasoning and general conversation.

AI Agent Abilities: Qwen models are designed to act as agents, with the capability to use tools and prepare complex APIs for tasks like marketing automation.

Reasoning and Understanding: The models have the ability to process large amounts of information (measured in tokens) and possess advanced reasoning capabilities, enabling them to be used for tasks involving deep analysis, such as code construction and managing long-form text. 

You can also read: Groundbreaking 10Gbps Direct Link Was Established Between Lagos And Cape Town

Access and Resources

Interactive Chat: Qwen models can engage in interactive chat with users at chat.qwen.ai.

Code and Checkpoints: Most of the Qwen models are available on online repository platforms like Hugging Face and GitHub. This means you can contribute to the improvement of Qwen models on these platforms. Hugging Face is an open-source platform for Machine Learning.

Cloud Platforms: The models are also available on cloud services such as Amazon Bedrock, which manages them in a fully serverless way. Additionally, there are non-open-source commercial versions available on Alibaba Cloud. Alibaba is an international technology corporation based in China that operates a large network of businesses comprising e-commerce, cloud computing, logistics, and digital media. 

Editorial Thought

The majority of technology companies worldwide are now dedicating significant resources to artificial intelligence (AI) and the Internet of Things (IoT). Additionally, businesses outside the tech sector are integrating AI to optimize their daily operations. As a result of this widespread adoption, AI's influence is becoming increasingly pervasive. From intelligent systems to AI-powered IoT, the technology is impacting a wide range of industries and aspects of daily life.