Cloud ambition,engineered together.
Integrated AI Hardware & Software Solutions
CloudPrime provides AI PCs, workstations, servers, standard motherboards and an agent platform, plus deployment and operations support.
New products & updates
XSA27V4
7U multi-GPU AI server
For model training, inference and data-center deployment.
View product
From personal devices to data centers
Choose personal, professional or enterprise products based on model size, user count and deployment method.
Enterprise AI
Business and flagship servers power training, inference and data-center-scale concurrent workloads.
View Enterprise Products
Personal AI
The MINI AI Workstation Q395 delivers 126 TOPS of combined AI performance, 128 GB unified memory and local 70B-model capability in an approximately 4L chassis.
View MINI AI Workstation
Professional AI
EPYC, Threadripper PRO and Xeon 600 workstation platforms support local LLMs, content creation, research and desk-side multi-GPU computing.
View 3 Professional Workstations
CloudPrime Agent Platform
Manage compute, models and applications
Supports on-premises and edge-cloud deployment for managing model services, agents, tools and applications.
Manage model services, agents and applications in one place
See the product page for architecture, deployment and monitoring details.
Explore Agent PlatformCompute that enters
real workflows.
Selected solutions for software engineering, document processing, archive management and financial research.
- 01Local code assistantSoftware R&D
- 02Document writing applianceGovernment office
- 03Intelligent archiveArchive management
- 04Financial research applianceFinancial research
01Local code assistantSoftware R&D
02Document writing applianceGovernment office
03Intelligent archiveArchive management
04Financial research applianceFinancial researchCompletion, generation, review and CI/CD integration — all inside your network.
- <100ms completion
- 8× AI accelerators
- Private deployment
Servers and standard motherboards
For training, inference, general compute and storage
Includes 42 servers and 9 standard motherboards for AI training and inference, general workloads, high-density storage and system development.
Large-scale Training
AI Servers
Eight-GPU B300, B200, SXM5 and OAM 2.0 systems for large-model training, HPC and data-center deployment.
Heterogeneous GPU
Compute Servers
Multi-GPU systems from 4U to 7U for model training, inference, rendering and scientific computing.
General Compute
Servers
Intel Xeon, AMD EPYC and Zhaoxin Kaisheng platforms for databases, virtualization, cloud computing, enterprise applications and lightweight AI inference.
High-density
Storage Servers
A 4U dual-socket platform with high drive density and expansion for distributed storage, data archiving and storage tiering.
Servers
Standard Motherboards
Supports Intel Xeon 6, Eagle Stream and Whitley, plus AMD EPYC Genoa, Turin, Rome and Milan, in ATX and EEB form factors.
