Lenovo ThinkStation PGX 1TB NVIDIA Grace Blackwell

Lenovo ThinkStation PGX 1TB NVIDIA Grace Blackwell (30KLS00L00) - 3 Years Local Warranty 

Overview

The Lenovo ThinkStation PGX 1TB NVIDIA Grace Blackwell (30KLS00L00) is a next-generation AI workstation designed for heavy compute workloads across machine learning, simulation, model training, and high-performance computing. Built with Lenovo’s professional-grade reliability, it supports developers, research teams, and enterprise organisations running continuous AI pipelines. Its architecture is engineered for long, intensive compute sessions with consistent stability and predictable performance.

AI Performance & Processing Power

Powered by NVIDIA’s latest Grace Blackwell architecture, the ThinkStation PGX delivers exceptional acceleration for AI model development and large-scale compute tasks. The unified Grace CPU and Blackwell GPU platform enables massive parallel processing, reducing training cycles and improving model optimisation. This configuration excels in LLM fine-tuning, deep learning research, multi-agent simulations, and enterprise AI workloads where sustained performance and efficiency are critical.

Storage & Workflow Efficiency

Equipped with a fast 1TB NVMe SSD, the workstation provides smooth handling of large datasets, multi-stage pipelines, and concurrent model development. The high I/O throughput supports rapid file access, efficient caching, and stable data flow for complex machine learning tasks. It is well-suited for LLM projects, simulation assets, large research datasets, and other compute-intensive workloads that demand fast, reliable storage.

Design & Reliability

Engineered with Lenovo’s workstation-grade thermal and structural design, the ThinkStation PGX maintains optimal temperature control during full GPU utilisation. Its cooling system delivers quiet, consistent operation even under sustained load, ensuring long-term stability for demanding workflows. The chassis is serviceable, upgrade-friendly, and scalable—making it ideal for AI labs, enterprise R&D teams, and production-level deployment environments.

ThinkStation PGX – NVIDIA Grace Blackwell (170W) Technical Specification

Processor & Architecture

• NVIDIA Grace Blackwell Platform (170W TDP)

• Built on Blackwell Architecture, delivering up to 1000 AI TOPS

• 20-core ARM CPU configuration:

  • 10 × Cortex-X925 high-performance cores

  • 10 × Cortex-A725 efficiency cores

• Unified CPU–GPU architecture optimized for LLM inference, generative AI, RAG, and high-density edge AI workloads

Memory

• 128GB LPDDR5X Unified System Memory

• High-bandwidth and low-latency design

• Shared CPU–GPU memory supporting large transformer models, embeddings, and multimodal workflows

Storage

• 1TB NVMe M.2 SSD

• Self-encrypting drive (SED) for enterprise-grade data security

• Fast sequential performance for model loading, dataset management, and AI development

Connectivity

• Wi-Fi 7 next-generation wireless connectivity

• Bluetooth 5.3 for peripherals and low-latency wireless accessories

• 4 × USB4 Type-C ports (up to 40Gbps each) for high-speed storage, displays, and expansion

• 1 × RJ-45 (10GbE) for high-bandwidth wired networking

Display & Multimedia

• HDMI 2.1a video output

• Supports high-resolution displays and advanced multimedia workflows

• HDMI multichannel audio output for professional AV environments

Software Platform

• Runs NVIDIA DGX™ OS (Ubuntu LTS–based AI-optimized operating system)

• Supports NVIDIA AI Enterprise, CUDA, TensorRT, and full NGC container ecosystem

• Ready for enterprise AI deployment, containerized workflows, and scalable inference