EMPOWERX LLM Integrated Appliance
Built on our self-developed open AI platform, the EMPOWERX Large Language Model Integrated Appliance delivers a ready-to-use, high-performance solution for enterprise generative AI.
The system offers multiple hardware configurations to meet different enterprise requirements, ensuring performance, stability, and scalability.DeepSeek large models can be pre-installed on demand, including 32B / 70B / 671B full-series models, eliminating complex deployment processes.
With fast delivery, plug-and-play deployment, high performance, and secure operation, the appliance enables organizations to quickly deploy private large models in a controlled environment.
Flagship Edition：Extreme Compute Power · Built for the Future (671B)
Product Model
EG4812-G4 Full-Scale LLM Inference Appliance (671B)
Configuration
CPU: 2 × 4th/5th Gen Intel® Xeon® processors
Memory: ≥ 1 TB
GPU: 8 × NVIDIA GPUs
Supported Models
DeepSeek full-scale models
Typical Scenarios
Ultra-large-scale AI research, AGI exploration, cloud-based inference services
Key Advantages
Only two systems are required to deploy the full-capability DeepSeek model.
Self-developed distributed inference technology supports high-concurrency workloads with stable performance.
Professional Edition：Balanced Performance · Enterprise-Grade Inference (70B)
Product Model
EG4804-G4 High-Performance LLM Inference Appliance
Configuration
CPU: 2 × 4th/5th Gen Intel® Xeon® processors
Memory: 256 GB – 1 TB DDR5
GPU: 4 × NVIDIA GPUs
Supported Models
DeepSeek-R1-Distill-Llama-70B and smaller models
Typical Scenarios
Complex reasoning tasks, large-scale knowledge-based Q&A, enterprise AI applications
Key Advantages
Flexible GPU scalability with higher compute density.
Optimizes data center space utilization while reducing overall TCO.
Essential Edition：Reliable Performance · Flexible and Efficient (32B)
Product Model
EG4412-G4 LLM Inference Appliance
Configuration
CPU: 2 × 4th/5th Gen Intel® Xeon® processors
Memory: 64 GB – 256 GB DDR5
GPU: 2 × NVIDIA GPUs
Supported Models
DeepSeek-R1-Distill-Llama-32B and smaller models
Typical Scenarios
On-premises conversational AI, text and code generation, multi-task inference
Key Advantages
Wide applicability with low latency and high efficiency.
A cost-effective solution for enterprises seeking practical generative AI deployment.