To address the needs of local AI applications, Qualcomm has launched local AI device solutions and inference kits to meet enterprises' needs for local AI application deployment and inference computing for AI applications in vertical fields, while ensuring data privacy and security, deployment flexibility, and further reducing construction costs.
Qualcomm proposed this solution primarily in response to the fact that most current AI application deployments rely on the cloud or require high-cost procurement of related hardware equipment. It emphasizes that its solution will maintain a balance between data privacy and computing power requirements while also reducing costs.
This solution primarily utilizes Qualcomm Cloud AI Pro or Ultra-class accelerator cards to support AI applications with parameter scales of 100 billion or up to 700 billion groups. At the same time, the device weighs only around 10 kilograms, or less than 15 kilograms, making it easy to deploy in any computing environment.
This solution can be used for currently common application services such as computer vision or automatically generated artificial intelligence, as well as local artificial intelligence agent services. It also includes the tool resources and related database content required for service construction, allowing enterprises or developers to quickly create various artificial intelligence application services and connect application needs from local to cloud deployment.
Qualcomm's AIOT edge computing solution is currently being adopted by leading AIOT edge computing providers Aetina, Honeywell, and IBM. This solution can be applied to a wide range of automation scenarios, including retail stores, fast food restaurants, shopping malls, dealerships, hospitals, factories, and production sites.
