Protecting your AI Investments with industry best practices for high density, liquid based operations

Salute AI terminology tool

Have a question about our AI capabilities?
Try our AI assistant.

For AI/HPC, heat rejection must begin at the chip.
Direct-to-Chip (DTC) Liquid Cooling is the solution.

Awaiting content from Rich

For decades, blowing air on servers was good enough for thermal management of servers. But air cooling is only effective for 30-40kW/rack, and power densities are magnitudes higher for AI/HPC computing equipment. “Blowing cold air on it” simply is no longer enough. For AI/HPC, heat rejection needs to start at the chip with Direct-to-Chip Liquid Cooling.

120-150kW+

NVIDIA’s Grace / Blackwell

250kW+

NVIDIA’s Vera Ruben

400-600kW+

NVIDIA’s Feynman

1MW+

Future NVIDIA

Direct-to-Chip Liquid Cooling Requires an Entirely New Approach to Data Center Operations

Utilizing liquid cooling requires a fundamental rethinking of data center operations to successfully operate complex DTC liquid cooling systems and to mitigate the new risks that liquid cooling creates.

Even brief interruptions to liquid cooling can lead to temperature spikes that cause catastrophic damage to expensive AI/HPC servers. Likewise, leaks in the liquid cooling systems can cause expensive equipment failures and create potentially fatal safety risks for workers. To mitigate these operational and financial risks, your organization must establish an entirely new set of operational processes based on proven best practices.

Whitepaper: Maximize your investment and minimize downtime with an operating model based on industry-leading best practices

Industry-Leading Operational Model for AI/HPC, 
Customized to Your Design

Salute has created best practices, e-learning and hands-on training for our operators, ensuring your liquid-cooled infrastructure is managed with best-in-class EOPs, MOPs, SOPs, emergency drills, chemistry management and leak management. We safeguard your AI investments with the expertise and discipline needed to prevent leaks, manage safety, and maximize uptime.

We safeguard your AI investments with the expertise and discipline needed to prevent leaks, manage safety risks, and maximize uptime.

Design and operational assessments to create an operational model for HPC environments

Industry best practice procedures –
EOPs, MOPs and SOPs

Comprehensive on-site classroom training,
e-learning, and lab-based certification programs
for operations staff specializing in direct-to-chip
liquid cooling systems

Commissioning
Validation of AI Design
Facilities Management Operations
Operational Acceptance
Operational
Requirements
Chemistry
Management
Leak
Management
Management Systems

Let's enable world-class DTC operations for your AI/HPC data center

Salute’s Best-in-Class Operational Model will ensure your data center operations can evolve to support this new Direct-to-Chip Liquid Cooling model. Contact us today to begin the process of assessing your design, analyzing your operational requirements and creating an operational model that meets your business objectives.

Commissioning of DTC Equipment
and Operational Protocols
Continuous Improvements
to Operational Best Practices
Assessment of DTC Operational
Requirements and Data Center Design
Customized Operational Processes for your specific equipment and customer requirements based on our best practice-based library of DTC operational processes
(EOPs, MOPs, and SOPs)
A Proven, Best Practice-Driven Approach to Education and Training for Chemistry Management, Leak Detection, Safety Protocols and Other Operational Processes

AI Insights & News

Download our white paper

Let’s optimize your DTC Liquid Cooling for what’s next

Salute’s Best-in-Class Operational Model will ensure your data center operations can evolve to support this new Direct-to-Chip Liquid Cooling model. Contact us today to begin the process of assessing your design, analyzing your operational requirements and creating an operational model that meets your business objectives.

Subject *
Any other info
Consent

John Shultz

Chief Product Officer, AI & Chief Learning Officer