EDGE INTELLIGENCE

Put large models next to the device.

From industrial PCs to robots, edge systems must balance power, cost, latency and model evolution.

01

Industrial PCs

Run knowledge Q&A, document retrieval and voice interaction together on the device.

02

Compute boxes

Add deployable, scalable local inference to industry devices.

03

Smart terminals

Deliver natural, real-time interaction under real power and cost constraints.

04

Robots

Run vision, language, motion planning and other models in parallel.

INTEGRATION

Extend AI without replacing the host.

Add local inference to existing devices and avoid redesigning the whole platform.

Existing platform

Keep the host controller, operating system and software interfaces.

Add the accelerator

Integrate based on interface, power and space constraints.

Run models

Deploy target models and applications through the software stack.

Keep evolving

Update with model and business requirements.