GPU-Accelerated Solution for Local AI Workloads
The Zstor Local AI Workstation WRX90E has been specifically designed to bring computing power for AI applications in an office-compatible workstation form factor. Be it for development of AI application, local fine-tuning of AI models or in-house interference on data sets, the WRX90E delivers the necessary performance right to your desk. This eliminates the dependency on third parties or public cloud access for processing your data, letting you keep the highest standard of privacy for the most valuable asset: your data.
Performance for AI workloads right at your desk
Thanks to integrated liquid cooling for the CPU, the system supports the latest AMD Threadripper™ PRO 9000 and 7000 processors with a TDP of up to 350 watts, high boost clocks of up to 5.4GHz, and up to 96 CPU cores.
With 8 DIMM slots, the server can accommodate up to 1.0 TB of DDR5 ECC Registered memory, thereby providing high memory bandwidth for professional applications. For low latency local caching and storage, there are 4 PCIe Gen5 x4 M.2 slots as well as 2x NVMe U.2 drives, with up to 7 additional NVMe U.2 drives or 4 additional 3.5" SATA drives being possible with optional drive kits.
A key feature of the Zstor Local AI Workstation is the ability to integrate up to four dual-slot or two triple-fan GPUs. This allows for flexible configurations tailored towards specific requirements. Those GPUs are paired with the most modern high-frequency multi-core CPUs to provide best in class Inference performance and throughput for the most demanding and versatile workflows.

Key Applications
The Zstor Local AI Workstation WRX90E as a GPU workstation platform is ideally suited for computationally intensive applications such as:
- Machine Learning
- Deep Learning
- AI model finetuning
- AI-Inference
- AI application development
- Scientific calculations
- Data analysis
- Rendering
- 3D visualisation
Zstor Local AI Workstations are designed for high-performance, low-latency inference and fine-tuning on pre-trained machine learning or deep learning Generative AI models like Stable Diffusion, Midjourney, Hugging Face, Character.AI, QuillBot, DALLE-2, etc; as well as crunching any workload based on industry standard frameworks. Unique multi-GPU cost-optimised and adjustable configurations are perfect for enabling Local AI development and Inference right at your desk.





Local AI, for enhanced IT security and compliance
Private AI systems are becoming increasingly indispensable for businesses when processing sensitive data whilst meeting the highest standards of data protection, compliance and IT security. Organisations in highly regulated sectors in particular – such as the legal sector, healthcare, financial services, public administration and education – face the challenge of deploying modern AI technologies without losing control over confidential information.
A locally operated AI infrastructure provides the necessary foundation for this. Instead of transferring data to external cloud providers or public AI services, all information remains within the organisation’s own IT environment. This allows companies to retain full data sovereignty and significantly reduce risks associated with data breaches, uncontrolled data leaks or regulatory violations.
A dedicated on-premises AI platform enables a wide range of applications – from intelligent internal assistants and AI-powered knowledge bases to automated workflows for support, document processing, analysis or decision-making processes. Companies can develop their own AI agents that are precisely tailored to internal processes, policies and specialist knowledge, without having to share sensitive information with external models.
Furthermore, a private AI environment enables significantly greater control over access rights, security policies and system integrations. Existing business systems such as CRM, ERP or document management solutions can be seamlessly integrated, allowing workflows to be streamlined and repetitive tasks to be automated. At the same time, employees benefit from faster access to information, intelligent support in their day-to-day work and a noticeable increase in productivity.
Local AI solutions are therefore ideal for all scenarios in which security, availability, customisation and compliance play a central role. Organisations gain a powerful, scalable and future-proof AI platform that enables innovation without compromising the protection of sensitive corporate data.
Specifications
| Zstor WRX90E Workstation | ||
|---|---|---|
| GPU | Up to 4x Dual-Slot GPU, e.g. RTX PRO 4000 / 5000 / 6000 (600W cards may be limited to 2 depending on configuration) Up to 2x Triple-Fan GPU |
|
| Platform | AMD Threadripper, AMD WRX90 Chipset | |
| CPU | AMD THREADRIPPER PRO 7000 and 9000 Series, 16 - 96 Cores | |
| System Power Usage | Up to 2.2 KW | |
| System Memory | 128GB up to 1TB DDR5 ECC Registered | |
| Networking | Dual-Port 10GB, 1GB IPMI, additional PCIe NICs optional | |
| Storage OS | Dual 1TB M.2 NVMe Drive | |
| Storage Data/Cache | 2x NVMe M.2 PCIe 5.0 x4 + 2x NVMe U.2 PCIe 4.0 x4 available, configuration on request, additional NVMe U.2 or SATA 3.5" drives possible | |
| PCIe Slots | 6x PCIe 5.0 x16, 1x PCIe 5.0 x16 (x8 electrical) | |
| Cooling | CPU with closed-loop liquid cooling, 360mm radiator with 3x 120mm fans GPU with air cooling Case with 4x 140mm and 2x 120mm fans |
|
| Connectivity | Rear: 2x USB 40Gbps (USB4®) with ASMedia ASM4242 controller ports USB Type-C®, 6x USB 10Gbps ports Type-A, 1x USB 2.0 port Type-A, 1x VGA port, 2x mini DisplayPort IN, 2x Audio jacks Top: 1x USB 10Gbps port Type-C, 2x USB 5Gbps ports Type-A, 1x Audio Combo-Jack |
|
| Software | Linux, Windows | |
| Size | 547 x 250 x 622 mm (D x W x H) | |
| Class | Tower Workstation | |
Tailor-made configurations for your needs
We configure our Zstor Local AI Workstations according to your specific use case.
For further information please contact the Zstor Sales Team via e-mail: