Product positioning: Targeting inference and edge scenarios, with ultimate spatial efficiency, specially designed for high concurrency, low latency AI inference scenarios. It is the core computing power support for real-time AI businesses such as AIGC services, intelligent security, and autonomous driving.

core strengths

Ultimate Space Integration Design:

Integrated with 8 single width AI acceleration cards in a 2U standard chassis, it only requires 1/2 of the cabinet space of traditional servers under the same computing power, greatly saving data center resources and reducing deployment costs

Ultimate reasoning performance:

High throughput computing architecture, microsecond level low latency response, capable of supporting thousands of concurrent inference requests, meeting the ultimate performance requirements of AIGC real-time generation, video stream real-time analysis and other businesses

Adaptive reasoning for all scenarios:

Deeply adapted to scenarios such as AIGC content generation, intelligent security video analysis, multi-sensor data processing for autonomous driving, and real-time financial risk control, ensuring real-time and stable business operations

Flexible deployment in dual scenarios:

Simultaneously adapting to standard data center rooms and edge node deployments, it can serve as a central inference cluster or sink to edge rooms to meet the computing power requirements of cloud edge collaboration