DLC for High‑Density AI & HPC
Keep your AI, GPU, and CPU clusters running at peak performance with efficient, reliable, and scalable liquid cooling—delivered by Agitare Technologies in partnership with Rittal.
Why DLC, Why Now?
Modern AI/ML, microservices, and edge workloads are pushing rack power far beyond what air cooling was designed to handle. It’s now common to see 50–100 kW per rack for high‑performance clusters, while older “CPU racks” were typically 10–17 kW.
A single modern AI GPU can draw ~3.7 MWh per year—comparable to the annual usage of many households—making advanced heat removal essential for stability and efficiency.
Higher Sustained Performance
Prevents throttling in CPUs/GPUs running up to 700 W TDP, ensuring consistent compute power even under full AI/HPC workloads.
Improved System Stability
Keeps components in optimal temperature ranges, reducing unexpected slowdowns or shutdowns.
Extended Hardware Lifespan
Lower sustained heat stress means less wear and tear on high-value processors, boards, and memory modules.
Massive Energy Savings
Achieve power reductions of up to 90% compared to traditional air-cooling methods.
Designed for Extreme Density
Handles 50–100 kW+ per rack loads—ideal for AI, HPC, and edge deployments.
Proven with Leading AI Hardware
Supports cutting-edge GPUs such as NVIDIA A100 with up to 400 W TDP each.
The AgitareTech + Rittal Advantage
As a Rittal manufacturer rep, AgitareTech brings a complete DLC portfolio and solution design expertise—from racks and manifolds to room‑level heat rejection and service. Rittal’s single‑phase DLC lineup spans Liquid‑to‑Air (RPU) and Liquid‑to‑Liquid (CDU) systems with capacities at 30/60/90/500/1000 kW to match your density and growth.
Rittal CDU L2L 1000‑21
500 kW to 1 MW
An in‑row Coolant Distribution Unit engineered for today’s densest racks.
- Form factor: 600 × 2300 × 1400 mm; standard 19″ and OCP configurations available.
- Five CCUs with 3 pumps per CCU (parallel), N+1 redundancy, per‑CCU leak detection and status LEDs; front‑panel RJ‑45 comms.
- Secondary loop flow: up to 1500 lpm @ 2.5 bar with N+1 redundancy (no filter).
- Plate heat exchanger sized for >1 MW heat transfer at 4 K ATD.
- Dual tanks with auto‑fill pump, expansion vessel, and a controller module; hot‑swappable control with fail‑safe operation.
- Power architecture: Open Rack v3 with DC busbar; 18 kW PSU (N+1) and optional battery backup.
- Flexible plumbing: top/bottom connections (tri‑clamp DN100 primary); hot‑swappable temperature, pressure, conductivity, and pH sensors.
- Manifold compatibility for third‑party microchannel and cold‑plate devices; 19″ rack retrofit kit available.
Where DLC Fits Best?
- AI training and inference GPU clusters
- HPC and research compute nodes
- Edge and micro‑datacenters with constrained floorspace
- Edge and micro‑datacenters with constrained floorspace
These deployments benefit most from DLC’s ability to remove high heat flux at the source and maintain deterministic performance.
From Design to Deployment
AgitareTech helps you plan and execute an optimized solution—racks, cooling, hybrid/cloud architecture, and services—aligned with your business goals. We’ll size, integrate, and validate a DLC design tailored to your density, redundancy, and sustainability targets.
Ready to Explore DLC?
Let’s scope your environment and model the ROI.
Your message has been sent
