Vertiv Liebert CRV In-Row Cooling 30kW

The Vertiv Liebert CRV In-Row Cooling system is a direct-injection precision cooling unit designed to address the thermal challenges of modern hyperscale and enterprise AI infrastructure. With a 30kW heat removal capacity, the CRV mounts directly into server racks to intercept hot exhaust air from compute nodes before it propagates into the data center plenum, eliminating hot spots and enabling higher rack density deployments. The system employs advanced refrigerant-based cooling with integrated controls and monitoring, allowing IT operators to maintain optimal inlet temperatures (typically 18–27°C) regardless of rack-level thermal output.

Launched in Q1 2024, the Liebert CRV addresses a critical supply-side gap identified in 2025 analyst shortage reports. As GPU and accelerator-based workloads drive per-rack power and thermal density to 50–100+ kW in leading-edge configurations, traditional computer room air conditioning (CRAC) and computer room air handler (CRAH) systems become insufficient. In-row cooling has emerged as the standard architecture for hyperscale and enterprise AI deployments, enabling denser compute placement without facility-level cooling infrastructure upgrades.

The CRV integrates seamlessly with existing data center management platforms via Liebert Network Management Card (NMC) interfaces, providing real-time thermal telemetry, alert routing, and remote diagnostics. Modular design allows stacking of multiple units to scale cooling capacity proportionally with rack expansion, making the system suitable for staged AI infrastructure buildouts.

Ideal for

• GPU cluster cooling for large language model (LLM) inference and fine-tuning operations running 8–16 GPUs per rack at continuous thermal load
• High-performance compute (HPC) environments executing machine learning training workloads where node-level thermal density reaches 80+ kW per rack
• Enterprise AI data centers consolidating multiple inference frameworks and vector database operations requiring sub-30°C inlet temperature stability
• Edge AI deployment sites and on-premises generative AI infrastructure where facility cooling cannot be immediately upgraded
• Hybrid cloud environments mixing CPU and GPU workloads within shared rack footprints, balancing thermal loads dynamically
• Colocation and managed service provider (MSP) facilities offering AI-optimized rack services to multiple tenants with SLA-bound uptime and thermal compliance

Technical specifications

Available from Omnixon Global. Submit an RFQ and our team will confirm configuration and availability for your order.