AI Server Product Family
Six server product lines covering AI computing, general-purpose computing, storage, domestic (information technology application innovation), liquid cooling and customization. Backed by a large-scale customized server production base, it provides the hardware foundation and full-lifecycle service for intelligent computing centers, education and research, advanced manufacturing and government/enterprise customers.
Six AI Server Product Lines
Covering the main hardware form factors in modern data centers, from general workloads to AI training, green computing and self-controllability
AI Computing Server
GPU-intensive servers for AI training and inference, offering high-density GPU expansion and high-speed interconnects for large model training, fine-tuning and inference deployment. Representative models include RWS6250-I43, RWS4250-I60 and RWS4260-I42; availability is subject to the final solution.
General-Purpose Server
For enterprise general workloads, hosting virtualization, databases, middleware and business systems with balanced compute, memory and I/O configurations, suited to daily operations and IT infrastructure expansion.
Storage Server
For large-capacity storage needs, offering high drive density and flexible storage configurations for video surveillance retention, backup and archiving, data lakes and distributed storage foundations.
Domestic Server
For information technology application innovation requirements, supporting domestic processors and operating systems to meet procurement and compliance needs for self-controllability in government, enterprise and critical industries.
Liquid-Cooled Server
Adopts liquid cooling to improve data center energy efficiency and compute density and help lower PUE, suited to high-density intelligent computing centers and green computing scenarios.
Customized Server
Backed by a large-scale customized server production base, servers can be deeply customized for chassis structure, cooling, power supply and interfaces to fit specific workloads and data center environments.
Core Capabilities of AI Servers
Large-scale in-house production
A large-scale customized server production base supports the chain from R&D through delivery.
High-density GPUs and high-speed interconnects
AI computing servers are designed for GPU-intensive workloads with high-density expansion for large model training and inference.
Liquid cooling and green computing
Liquid cooling improves compute density and energy performance for data centers sensitive to power and space constraints.
Domestic platform support
Domestic servers support domestic processors and operating systems to meet self-controllability requirements.
Deep on-demand customization
Chassis structure, cooling, power supply and interfaces can all be customized to customer requirements for specific scenarios.
Full-lifecycle service
Expert-level service spanning pre-sales, in-sales and after-sales, with branch coverage supporting delivery and operations.
Typical Deployment Scenarios
Intelligent Computing Center
A high-density, energy-efficient hardware foundation for large model training and inference services.
Education & Research
Supports teaching labs, research simulation and data mining platforms at universities and research institutes.
Advanced Manufacturing
Compute for industrial IoT, edge computing and AI quality inspection on production lines.
Government & Self-Controllability
Meets server procurement and deployment needs tied to domestic and self-controllability requirements.
The above describes typical application scenarios and does not represent specific customer cases.
AI Server Selection Guide
| Product line | Positioning | Typical workload | Key features | Scenarios |
|---|---|---|---|---|
| AI computing server | GPU-intensive compute platform | Large model training, fine-tuning, inference | High-density GPU expansion and high-speed interconnects | Intelligent computing centers, AI R&D |
| General-purpose server | General workload hosting | Virtualization, databases, middleware | Balanced compute/memory/I/O configuration | Enterprise IT infrastructure |
| Storage server | Large-capacity data storage | Video surveillance, backup, data lakes | High drive density, flexible storage configuration | Security, archiving, distributed storage |
| Domestic server | Self-controllable compute | Government and enterprise systems | Supports domestic processors and operating systems | Government, critical industries |
| Liquid-cooled server | Energy-efficient compute | High-density deployment | Liquid cooling, higher compute density | Green data centers, intelligent computing |
| Customized server | Tailored delivery | Special workloads and data center environments | Customized structure/cooling/power/interfaces | Deeply specific scenarios |
Frequently Asked Questions
How do I choose among the six server lines?
A practical starting point is workload profile: choose AI computing servers for large model training and inference; general-purpose servers for everyday business and virtualization; storage servers for large-scale retention and archiving; domestic servers where self-controllability and localization are required; liquid-cooled servers where power and density are constrained; and customized servers when special mechanical, thermal or interface requirements exist. Final configuration is subject to an assessment of actual workloads.
Are customized servers available?
Yes. Backed by a large-scale customized server production base, servers can be deeply customized to customer requirements for chassis structure, cooling, power supply and interfaces. Feasibility and lead time are subject to actual evaluation.
What problem do liquid-cooled servers solve?
Liquid-cooled servers use liquid cooling to improve data center energy efficiency and compute density, helping to lower PUE. They suit high-density intelligent computing centers and green computing scenarios. Actual energy savings depend on data center conditions and the deployment design, and are subject to the final solution.
Which platforms do domestic servers support?
Domestic servers address information technology application innovation requirements and support domestic processors and operating systems for self-controllability programs in government and enterprise as well as critical industries. The specific list of supported platforms and compatible models is subject to the final solution.
Further Reading
AI Computing All-in-One
RoycomONE platform, DeepSeek all-in-one and the three-tier collaborative full-stack AI solution.
AI & High-Performance Computing
SuperAI deep learning, SuperHPC and SuperHPC hyper-converged infrastructure.
SuperStorage
RS8000 all-flash, RS6000 massive scale, RS4000 unified storage and Weka.
Product Center
Explore the full VISBAT enterprise network and security product portfolio.