AI Infrastructure · Liquid Cooling · Since Singapore
Build Around Liquid Cooling.
Engineered for Compute Efficiency.
Precision cold-plate and immersion cooling for high-density AI compute — planned, designed and delivered as one system, from Singapore to ASEAN and beyond.
Company · Who We Are
A Singapore-based AI infrastructure company
QuettaFlow Technologies enables private, on-premise, edge and large-scale AI deployment. Our mission is to help enterprises deploy high-performance AI computing in a more compact, efficient, secure and high-density way.
We are two complementary operators working as one: Singapore precision-engineering and capital-markets discipline, meeting deep, China-grown immersion-cooling R&D.
We own the critical infrastructure layer behind AI.
Edge & private AI
Bring AI computing close to the customer's own data and operations.
Immersion-cooled compute
High-density GPU systems cooled directly in dielectric fluid.
Compact & secure deployment
Full performance — without building a full-scale data centre.
Our Approach
AI Compute Infrastructure Planning Study (AIPS)
Others deliver standalone liquid-cooling equipment. AIPS is our engineering-led study that plans Compute, Network, Power and Cooling as one system — designed once, globally optimised.
Deployment
Density planned to the cooling capacity — full performance released.
Interconnect
Topology co-planned around high-density racks — shorter links, lower latency.
Distribution
Co-designed with cooling and IT — a closed efficiency loop, supply to heat rejection.
Foundation
The anchor of the design — it sets the ceiling on density and PUE.
Compute · Planning & Design
Demand analysis & sizing, chip selection (NVIDIA or domestic GPUs), server configuration and compute–cooling matching — turning business need into a concrete deployment plan.
Deliverables: Compute build plan · Server configuration list · Sizing report
Network · Architecture Design & Integration
Lossless network planning (InfiniBand or RoCE), equipment selection, three-tier topology and cabling around high-density liquid-cooled racks.
Deliverables: Topology diagrams · Networking plan · Equipment list
Power · Distribution Design
End-to-end planning from grid to chip, load & capacity sizing, 2N/N+1 redundancy, and busway/PDU schemes for high-power liquid-cooled racks.
Deliverables: Electrical schematics · Power-distribution plan
Cooling · HVAC Design + In-house Equipment
In-house cold plates, immersion tanks, CDUs and manifolds; cold-source design; hydraulic-balance piping; 3D modeling & CFD before construction.
Deliverables: Floor layouts · HVAC diagrams · 3D piping models
What We Provide
Two core business platforms
Our strategic positioning: a full-stack liquid-cooled AI compute infrastructure provider, built around two complementary platforms.
Project-Based
AIDC Facility Platform
Full-Stack Liquid-Cooled AI Compute Infrastructure
- ForHyperscalers, data centers & large-scale equipment rooms
- ProvidesIntegrated compute, network, power and cooling — cooling developed in-house, compute/network/power delivered through complete solution design with partner equipment
- ProductsVasPlate & VasRack
"We build the house for AI compute."
Standard Product
Edge Compute Platform
Immersion-Cooled AI Compute Solution
- ForEnterprise on-premise and edge deployment
- ProvidesIntegrated immersion cooling hardware for AI use cases within SMEs & enterprises
- Products8U VasEdge & 16U VasEdge
"AI compute, packaged in a box and ready to use."
Two main areas of focus: Liquid cooling for AI data-centre facilities, and immersion cooling for edge computing.
Product Line
Built to remove heat where compute density is highest
Three product lines, one engineering foundation — from data-centre scale cold-plate cooling to plug-in-and-run edge appliances.
Cold Plate · AIDC Facility Platform
VasPlate
Our cold-plate liquid cooling system for AIDC deployments — a mature, reliable cooling solution for today's highest-power AI rack systems, including NVIDIA flagship racks and domestic liquid-cooled platforms.
The broadest-fit product in our AIDC facility platform — mature cold-plate technology for advanced, high-power AI platforms.
Immersion · AIDC Facility Platform · Flagship
VasRack
A pioneering blind-mate modular immersion cooling system with PowerShelf + Busbar power distribution and a built-in distributed CDU — under one meter high, serviceable by hand, ready for rapid deployment.
Immersion · Edge Compute Platform
VasEdge
On-premises liquid-cooled AI for offices, labs, hospitals, research and government — high-performance computing made accessible, with no dedicated server room required.
Operational within 4 hours — no lifting equipment, no dedicated data-centre team. Data stays on-premise for security and sovereignty.
Preconfigured AI use cases, ready on day one
Our Strategy
Differentiation & benchmarks
We understand liquid cooling better than most IT players, and we understand IT better than most cooling players.
Traditional cooling vendors
Strong in thermal management and manufacturing, but less familiar with upper-layer IT requirements such as compute, networking and server architecture.
Traditional IT integrators
Strong in compute and servers, but less familiar with the engineering details of liquid cooling.
QuettaFlow
Combines liquid-cooling engineering depth with IT and compute understanding — designing from the perspective of what compute truly requires. A scarce, hard-to-replicate capability.
Global leaders like Vertiv and Schneider operate broad infrastructure product lines, with liquid cooling as one line among many. Our approach is to anchor fully on liquid cooling and win through deep integration across cooling and IT — responding faster, fitting solutions more closely, and operating with greater flexibility. The giants are all-rounders; we are the liquid-cooling specialist that understands AI compute.
Credentials & Track Record
Proven deployments, protected IP, global partnerships
Beijing · Lanting Yunjing · Phase I
10,000-card national data hub, built on 100% green energy
A T3+ standard, Beijing-first green-power-zone AIDC scaling to 25,000 computing cards. Phase I delivered a synthetic-oil immersion cooling set, 15 modular immersion GPU servers and a full monitoring platform.
- pPUE reduced to within 1.1
- Server failure rate down >50%
- Room noise below 40 dB
Tianjin · Wuqing Telecom · Phase I Retrofit
128-card immersion retrofit on a Tier 3+ SDN data centre
A 380,000 sqm, 10-billion-RMB campus retrofitted with synthetic-oil immersion cooling using the site's existing chilled-water system, 25G RoCE networking and a liquid-cooling cloud management platform.
- pPUE reduced to under 1.2
- Failure rate down >50%, power >15%
- Noise reduced below 50 dB
Loading recent milestones…
Build Around Liquid Cooling.
Engineered for Compute Efficiency.
Let's plan your compute, network, power and cooling as one system.
Headquarters
Singapore