Lenovo ThinkStation PGX NVIDIA Grace Blackwell - 4TB

Lenovo ThinkStation PGX 4TB NVIDIA Grace Blackwell (30KLS01700) - 3 Year Premier Support Local Warranty 

Advanced AI Workstation Powered by NVIDIA Grace Blackwell

Lenovo ThinkStation PGX 4TB NVIDIA Grace Blackwell (30KLS01700) is a next-generation AI workstation engineered for artificial intelligence, machine learning, large language models (LLMs), and accelerated computing workloads. Powered by the groundbreaking NVIDIA Grace Blackwell architecture, this workstation delivers enterprise-grade AI performance for developers, researchers, data scientists, and organizations building advanced AI applications.

Optimized for Generative AI and Large Language Models

Designed to handle demanding AI workloads, the Lenovo ThinkStation PGX supports generative AI, model inference, AI fine-tuning, machine learning, simulation, and data analytics. The powerful NVIDIA Grace Blackwell platform enables efficient processing of large AI models and complex datasets while supporting local AI development and deployment environments.

Enterprise AI Computing with High-Capacity Storage

Featuring a 4TB storage configuration, the ThinkStation PGX provides ample capacity for AI datasets, machine learning models, project files, and development environments. Its workstation-class architecture allows organizations to deploy powerful AI computing resources while maintaining flexibility, security, and operational efficiency.

Built for Future-Ready AI Innovation

The Lenovo ThinkStation PGX is designed to accelerate AI adoption across enterprises, universities, research institutions, and development teams. It provides a scalable platform for next-generation AI applications, advanced analytics, digital twin environments, and edge AI deployments while enabling on-premises AI processing and enhanced data privacy.

Technical Specifications

General
Model: ThinkStation PGX
Part Number: 30KLS01700
Product Name: Lenovo ThinkStation PGX 4TB NVIDIA Grace Blackwell
Manufacturer: Lenovo
Product Type: Personal AI Supercomputer / AI Workstation
Deployment: Generative AI, AI Development, Large Language Models (LLM), AI Inferencing, Edge AI, Data Science
Architecture: NVIDIA Grace Blackwell Platform
Target Workloads: AI Model Training, AI Fine-Tuning, LLM Deployment, Deep Learning, Generative AI Applications
Operating System: NVIDIA DGX™ OS
Platform Type: Enterprise AI Desktop Computing Platform

Processor Specifications
Processor Architecture: NVIDIA GB10 Grace Blackwell Superchip
CPU Type: Arm-Based AI Processor
CPU Core Count: 20 Cores

CPU Configuration:
• 10 x Cortex-X925 Performance Cores
• 10 x Cortex-A725 Efficiency Cores

CPU-GPU Interconnect: NVIDIA NVLink™-C2C
AI Accelerated Computing: Supported
Unified Compute Architecture: Supported

GPU Specifications
GPU Architecture: NVIDIA Blackwell Architecture
Tensor Core Generation: 5th Generation Tensor Cores
RT Core Generation: 4th Generation RT Cores
CUDA Core Architecture: NVIDIA Blackwell GPU Compute Platform
AI Performance: Up to 1 PetaFLOP AI Compute Performance

Supported AI Workloads:
• Generative AI
• Large Language Models (LLM)
• AI Inferencing
• AI Fine-Tuning
• Deep Learning
• AI Agent Development

AI Optimization: Supported
Accelerated AI Computing: Supported

Memory Specifications
Memory Type: LPDDR5x Unified Memory
Memory Capacity: 128GB
Memory Architecture: Unified CPU-GPU Shared Memory
Memory Interface: 256-bit
Memory Bandwidth: Up to 273 GB/s

Storage Specifications
Storage Capacity: 4TB
Storage Type: NVMe PCIe SSD
Storage Architecture: High-Speed AI Optimized Storage
Self Encryption Support: Supported

Networking Specifications
Network Technology: NVIDIA ConnectX®-7 Networking
High-Speed Networking: Supported
Low-Latency Networking: Supported
AI Cluster Expansion: Supported

Supported Features:
• High Bandwidth AI Networking
• Multi-System Scaling
• AI Cluster Connectivity
• Enterprise AI Deployment Support

Connectivity
USB Ports:
• USB Type-C Connectivity

Display Output Support: Supported
High-Speed Data Transfer: Supported
Peripheral Expansion: Supported

Software & AI Platform
Operating System: NVIDIA DGX™ OS

Supported AI Functions:
• AI Model Development
• AI Fine-Tuning
• AI Inferencing
• Generative AI Workflows
• AI Agent Development
• Large Language Model Deployment
• Machine Learning Development

AI Resource Management: Supported
System Monitoring: Supported
Development Environment Optimization: Supported

Performance & Features
Up to 1 PetaFLOP AI compute performance
Unified memory architecture for AI workloads
Designed for local generative AI deployment
Optimized for LLM inferencing and fine-tuning
Supports enterprise AI development environments
High-speed CPU-GPU communication via NVLink™-C2C
Compact AI supercomputing architecture
Supports large parameter AI model workloads

Cooling & Thermal Design
Cooling Type: Advanced Active Cooling System
Thermal Optimization: AI Workload Optimized
Compact Airflow Design: Supported

Security Features
Self Encrypting Storage Support
Enterprise Data Protection Support
Local AI Processing Capability
Secure AI Development Environment

Design & Construction
Form Factor: Compact Desktop AI Workstation
Deployment Environment: AI Research, Enterprise AI Development, Edge AI, Data Science
Construction: Enterprise Grade AI Computing Platform

Package Contents
Lenovo ThinkStation PGX AI Workstation
Power Adapter
Documentation
Preloaded NVIDIA DGX™ OS

Key Features
NVIDIA GB10 Grace Blackwell Superchip
128GB LPDDR5x unified memory
4TB NVMe SSD storage
Up to 1 PetaFLOP AI performance
NVIDIA DGX™ OS platform
NVIDIA ConnectX®-7 networking support
Unified CPU-GPU memory architecture
Optimized for local AI and LLM workloads
Compact desktop AI supercomputer design

Typical Use Cases
Large Language Model (LLM) deployment
Generative AI development
AI inferencing workloads
AI model fine-tuning
Machine learning development
Data science environments
Enterprise AI prototyping
Edge AI deployment
Private on-premise AI computing

Detailed Specification

ThinkStation PGX: NVIDIA Grace Blackwell - 170W
Blackwell Architecture - 1000 AI TOPS
20 core Arm, 10 Cortex-X925 + 10 Cortex-A725 Arm 
128 GB LPDDR5x, unified system memory / 4TB NVME.M2 with self-encryption 
WiFi 7 / BT 5.3/ 4x USB 4 TypeC (up to 40Gb/s) / 1x RJ-45 connector (10 GbE)
HDMI 2.1a /v HDMI multichannel audio output
NVIDIA DGX OS