Grace Blackwell Ultra
NVIDIA GB300 architecture for substantial local AI workloads.
NVIDIA AI computing at IDEALTECH
NVIDIA
AI Platforms.
From your desk to your
next AI breakthrough.
Explore DGX Station, DGX Spark and RTX Spark systems for local AI development, demanding workloads and everyday creation.
DGX StationDGX SparkRTX Spark
Three platforms. Your next step.
Find the Right System
for Your AI Workloads.
Compare how you work, what you want to run and where you need your system.
Swipe to compare all three platforms
| Your platform at a glance |
 |  |  |
|---|---|---|---|
| Best suited to | Large-model development, demanding inference and advanced data science | Local agents, prototyping, supported fine-tuning and inference | Local AI, creative work, RTX gaming and everyday productivity |
| NVIDIA platform | NVIDIA GB300 | NVIDIA GB10 | NVIDIA RTX Spark N1X |
| Form factor | Deskside AI system | Compact desktop AI system | Laptops and compact desktops |
| Software environment | Confirm the supplied OS with each manufacturer | NVIDIA DGX OS / Ubuntu-based software environment | Windows 11 |
| Connectivity | ConnectX-8; system configuration varies | ConnectX networking for supported multi-system workflows | Depends on the selected system |
| Explore each platform | Explore DGX Station | Explore DGX Spark | Explore RTX Spark |
Representative devices shown. Platform overview, not a guarantee for every system. Memory, software, networking and application support vary by manufacturer and configuration. Memory capacity alone does not identify an RTX Spark CPU/GPU configuration.
For demanding AI workloads
NVIDIA DGX Station
Think bigger. Build locally.
Bring large-model development and advanced data science closer to your team with a dedicated deskside AI system.
748GBCoherent CPU–GPU memory
More about DGX Station
Room for larger models
A large coherent memory system for supported models and workflows.
Built for your team
A dedicated platform for development, experimentation and data science.
Explore the lineup
DGX Station systems
Open a product for its configurations, current availability and delivery information.
Your personal AI supercomputer
NVIDIA DGX Spark
Big ideas. Small footprint.
Develop agents, prototype applications and work with AI models on a compact NVIDIA GB10-powered desktop.
128GBUnified memory in the featured lineup
More about DGX Spark
NVIDIA GB10 Superchip
Grace Blackwell architecture on your desk.
128GB unified memory
A large shared memory pool for compatible local AI workloads.
Connect compatible systems
Extend supported workloads with the right software and networking.
Explore the lineup
DGX Spark systems
Available
NVIDIA DGX SPARK AI Founder Edition
RM25,999.00
Explore More Available
MSI EDGEXPERT AI DGX SPARK
RM25,599.00
Explore More Backorder
ASUS ASCENT GX10 AI DGX SPARK
Starting From RM34,599.00 - RM45,999.00
Explore More Backorder
GIGABYTE AI TOP ATOM DGX SPARK
RM28,999.00
Explore More Backorder
HP ZGX Nano G1N AI DGX Spark
RM25,599.00
Explore More Open a product for its configurations, current availability and delivery information.
From ideas to working AI
Your AI Workloads.
On Your Desk.
Six ways to put a DGX Spark or compatible GB10-powered system to work.
Agentic AI
Develop agents that coordinate models, tools and tasks locally.
Prototyping
Build and evaluate an application before taking it to a larger deployment.
Fine-Tuning
Adapt supported pretrained models to your own datasets and use cases.
Inference
Run compatible models and evaluate their responses on your own system.
Data Science
Explore GPU-accelerated analytics and machine-learning workflows.
Edge Applications
Develop computer vision and robotics applications with compatible frameworks.
Start small. Connect more.
Grow Your AI Setup.
Start with one 128GB DGX Spark system. Connect compatible systems as your supported workloads grow.
1 System
128GBAcross 1 system
2 Systems
256GBAcross 2 systems
4 Systems
512GBAcross 4 systems
Totals are for 128GB systems. Each system retains its own memory; supported software distributes the workload. Four-system clusters require a suitable switch, compatible networking and configuration.
Discuss a multi-system setup
AI development. Creation. Gaming.
NVIDIA RTX Spark
Your ideas, wherever you work.
Explore NVIDIA RTX Spark laptops and compact desktops that bring local AI, creative applications and RTX gaming into a Windows PC.
Up to 128GBUnified memory; check the exact model
More about RTX Spark
Local AI & Agents
Explore compatible local models and agent workflows.
AI Development
Prototype and test your AI ideas on a personal system.
3D Design & Rendering
Create with supported RTX-accelerated applications.
Video & Generative AI
Explore supported video tools and generative workflows.
RTX Gaming
Enjoy RTX features in supported games.
Streaming & Video Calls
Use supported NVIDIA Broadcast features for your setup.
Explore the lineup
RTX Spark systems
Open a product for its configurations, current availability and delivery information.
A little more clarity
Your Questions, Answered.
Which NVIDIA AI system should I choose?
Start with the applications and models you want to run. DGX Spark is a compact local AI platform; DGX Station targets more demanding deskside AI workloads; RTX Spark combines local AI with Windows-based creation and gaming. Tell us your intended workload so we can help identify a suitable configuration.
Does more memory always mean a faster system?
No. Memory affects which workloads can fit, while speed also depends on the processor, GPU, precision, software, power limits and model settings. Compare the complete configuration rather than memory capacity alone.
Is a 64GB RTX Spark a different processor from a 128GB model?
Memory capacity alone does not identify the processor configuration. Check the CPU, GPU and memory specifications of the exact product you are considering. Open the product listing for its available configurations.
Can I connect multiple DGX Spark systems?
Compatible networking and supported software can distribute workloads across DGX Spark systems. Each system keeps its own memory; this is not one automatically shared memory pool. The four-system layout shown here requires an appropriate switch and configuration. Confirm OEM networking compatibility before ordering.
Will my existing software work on these systems?
Check the application, operating system, Arm architecture, drivers and dependencies required by your workload. Windows and DGX OS environments are different, and hardware capability alone does not guarantee application compatibility.
Where can I check pricing, availability and delivery?
Open the individual product listing to review configurations and availability. If a system is marked Backorder, contact us to confirm the current delivery estimate before placing your order. You can reach our team by WhatsApp or email below.
Let's build your next step
Find the Right AI System
for Your Work.
Tell us your applications, model requirements and budget.
We'll help you explore a suitable configuration.