Lenovo ThinkStation PX Workstation, Dual Intel Xeon w9-3475X

The Lenovo ThinkStation PX (30EU0010US) is a dual-socket tower workstation built around two Intel Xeon w9-3475X processors, delivering a combined 120 cores and 240 threads on the Intel W790 chipset platform. Engineered in collaboration with Aston Martin for industrial aesthetics and optimized internal airflow, the chassis supports up to 8TB of DDR5 ECC RDIMM memory across 16 DIMM slots, making it one of the most memory-dense commercial workstations available. The platform accommodates PCIe 5.0 expansion slots and supports multi-GPU configurations, enabling high-throughput parallel compute workloads that demand both sustained CPU bandwidth and discrete GPU acceleration.

The ThinkStation PX is purpose-built for professional workloads that require extreme multi-threaded performance alongside large working memory footprints. Its dual Xeon w9-3475X processors each feature 36 cores running at up to 4.8 GHz Turbo, with support for Intel Advanced Matrix Extensions (AMX) and Intel Deep Learning Boost (VNNI) instructions that accelerate AI inference and training pipelines directly on the CPU. The system supports up to four double-width professional GPUs via full-length PCIe slots, and the internal architecture provides dedicated bandwidth lanes per socket through the dual-socket UPI interconnect fabric.

Designed for enterprise data-science clusters and high-end visualization environments, the ThinkStation PX ships with Lenovo's ISV-certified driver stack and integrates with Lenovo Device Manager and Intel vPro Enterprise for remote fleet management. The chassis features a tool-less drive bay design, front-accessible storage bays supporting U.2 NVMe SSDs alongside SATA devices, and a redundant power supply option to maintain uptime in critical production environments. This model carries Lenovo's commercial-grade build quality and carries broad ISV certifications across major CAD, simulation, and data-science software ecosystems.

Ideal for

• Large-scale machine learning model training and AI inference using AMX/VNNI CPU acceleration combined with multi-GPU compute
• Computational fluid dynamics and finite element analysis requiring sustained multi-core throughput and large in-memory datasets
• High-resolution 3D rendering and real-time visualization for automotive and aerospace design pipelines
• Data-science cluster node deployment where per-node memory capacity and core density reduce job segmentation overhead
• Genomics and life-sciences workloads processing large reference datasets entirely within system RAM to minimize I/O latency
• Virtual production and real-time ray tracing workflows demanding simultaneous CPU simulation and GPU rasterization headroom

Technical specifications

Available from Omnixon Global. Submit an RFQ and our team will confirm configuration and availability for your order.