The Lenovo ThinkStation PGX [30KL000GMY] is a compact AI development workstation powered by the NVIDIA GB10 Grace Blackwell Superchip. Designed for local AI development, prototyping, fine-tuning and inference, it combines up to 1 PFLOP of FP4 AI performance, 128GB of unified LPDDR5x memory and a high-capacity 4TB NVMe SSD in a compact 1.13-liter chassis.

NVIDIA GB10 Grace Blackwell Superchip - The 20-core Arm processor integrates 10 Cortex-X925 and 10 Cortex-A725 cores with NVIDIA Blackwell GPU architecture, 5th-generation Tensor Cores and 4th-generation RT Cores. It delivers up to 1 PFLOP of FP4 AI performance and supports AI models with up to 200 billion parameters.

128GB Unified LPDDR5x Memory - Featuring 128GB of LPDDR5x unified system memory on a 256-bit interface with up to 273GB/s bandwidth, the PGX provides a large shared memory pool for demanding AI models, datasets and development workloads.

4TB NVMe SSD & NVIDIA DGX OS - The 4TB NVMe SSD provides ample high-speed local storage for AI models, datasets, applications and development projects. NVIDIA DGX OS and the NVIDIA AI software stack provide a ready-to-use environment for AI development, fine-tuning and inference.

ConnectX-7 & Compact 1.13L Design - NVIDIA ConnectX-7 provides high-speed networking and enables PGX systems to be linked for larger AI workloads. Connectivity includes 10GbE, Wi-Fi 7, Bluetooth 5.3, HDMI and USB-C, while the compact 150 × 150 × 50.5mm chassis weighs approximately 1.2kg.

Customer Reviews on Lenovo TS PGX/GB10/128GB/4TB SSD/NVIDIA DGX OS

Q&A on Lenovo TS PGX/GB10/128GB/4TB SSD/NVIDIA DGX OS

Customer questions & answers