AI Solutions, from desk to data centre.
What we do.
- Consulting
We start with the value
We work out with you where AI saves your team time or money, what it would take and whether it is worth it, before you buy anything.
Talk to us - Systems and models
The right machine, ready to work
Our own AI workstations and servers, built and tested in Malaysia, and systems from other makers, with the models you need installed on them.
See the systems - Applications and support
Built around your work
AI applications around your documents and workflows, and our ATLAS app on DGX Spark. Our engineers in Malaysia support the whole system after delivery.
See solutions
We help you choose, deploy and run the right AI infrastructure.
We start as your AI consultant: what to automate, which model fits and what it is worth. Then one team in Malaysia builds, sets up and supports the system, from the first conversation to upgrades years later. Building on-premises AI since 2022.
- A Malaysian computing brand
- Since 2016
- Systems built
- 10,000+
Organisations that have bought EMARQUE systems since 2016.
All clients
How we work
Most projects follow these four steps. Buying a machine is quicker than a custom application.
- Step 1
Consult
You tell us what you want AI to do. We tell you where it adds value, where it does not, and what it would take.
- Step 2
Design
We choose the model and size the hardware to it.
- Step 3
Build and set up
We assemble our own systems, set up the systems we deliver and install the software.
- Step 4
Support
After handover we support it from Malaysia.
Our systems
Compare all systemsCompact systems for one developer or a small team.
Workstation
All workstation systemsBigger machines for a team, beside a desk or in a rack.
Server
All server systemsRackmount servers and rack-scale systems, air-cooled or liquid-cooled.
Air-cooled
EMARQUE AI Server
- GPU memory
- 576 / 768 GB
- RTX PRO 6000 GPUs
- 6 or 8
- Rackmount, air-cooled
- 6.5U
EMARQUE AI HGX B300
- GPU memory
- 2.1 TB
- Blackwell Ultra GPUs
- 8
- Max system memory
- 4 TB
NVIDIA DGX B200
- Blackwell GPUs
- 8
- HBM3e
- 1,440 GB
- FP4 with sparsity
- 144 PFLOPS
NVIDIA DGX B300
- Blackwell Ultra GPUs
- 8
- HBM3e
- 2.1 TB
- FP4 with sparsity
- 144 PFLOPS
Common questions
Something else on your mind? Contact us.
How much does an on-premises AI system cost?
It depends on the number of GPUs, the memory and storage, and the support you want. Component prices move often, so we do not publish a list price. Tell us what the system needs to do and we send a written quote in MYR.Which system should we start with?
For one person or a small team trying private AI, an NVIDIA DGX Spark. For a department where many people use it at once, a rackmount workstation such as the AI PRO 500, or the AI Server for more GPUs. Tell us what you want it to do and we will recommend one, including when a smaller machine is enough.Do we need our own IT team to run it?
No. We set the system up and install the models and software. After handover we support it from Malaysia.Where does our data go?
It stays on your premises. The models run on your own machine, so prompts and documents are not sent to an outside AI service unless you connect one. On DGX Spark, our ATLAS application counts every outside connection and shows it to you.What happens if something breaks?
Contact us and we take it from there. Systems we build carry our three-year system warranty and are repaired by us in Malaysia. NVIDIA systems carry their maker's warranty, and we handle the claim for you.Should we buy our own hardware or use a cloud AI service?
Your own hardware suits steady, everyday use, and data that must stay inside the company. A cloud service suits short experiments. We will tell you which fits your case.Do you deliver outside the Klang Valley?
Yes, across Malaysia, including Sabah and Sarawak.
Tell us about your workload.
Tell us what it needs to do and who will use it. A few lines is enough. We reply by email, and every quote comes in writing, in MYR.