By clicking “Accept All Cookies”, you agree to the storing of cookies on your device to enhance site navigation, analyze site usage, and assist in our marketing efforts. View our Privacy Policy for more information.
PreferencesDenyAccept

Solution

Smart Thermal Control
From Chip to Data Center
Modular, Efficient, and Built for Future of AI

GemaTEG Thermal Management Ecosystem

40% More Energy Efficiency

Unlock the next wave of computing with GemaTEG. Our active direct-to-chip cold plates are engineered to manage on-chip hotspots as workloads change.

Each plate is orchestrated by a smart, integrated platform that enables adaptive control, seamless monitoring, and predictive responsiveness to real-time computing demands.

The GemaTEG Thermal Management Ecosystem allows data centers to operate with unprecedented warm-fluid temperatures, even in the most performant and dense installations.

Achieve 40% higher efficiency than conventional Direct Liquid Cooling, boosting performance only where and when it matters most —independently of chip generation or power. Avoid fluid pre-cooling with inefficient and costly chillers, even in extreme environmental conditions.

The Challenge

AI’s Next Thermal Bottleneck

Chip architecture and increasing power consumption are becoming the next major bottleneck in AI infrastructure. Conventional cooling technologies are struggling to provide efficient and flexible responses to these new demands, leading to higher capital investments and reduced data‑center efficiency.

DaTEG: Active Cold Plate

DaTEG is engineered to actively control each chip’s operating temperature, adapting in real time to the demands of HBMs and Logical Cores. This eliminates the need for pre‑cooling on fluid loops from chillers.

Seamless integration with DLC infrastructure for enhanced compatibility.

GemaTEG Supervisor

The Supervisor is a new way to interact with the cooling infrastructure: It monitors, controls, and reacts, from granular DaTEG orchestration to policy-driven control.

Seamless integration with infrastructure monitoring platforms.

Features

A New Standard for Thermal Performance

Warm‑Fluid Operation (>43°C)

Guaranteed across chip generations and environmental conditions, with no chiller pre‑cooling required.

Higher Efficiency

Achieve 40% higher efficiency than conventional direct liquid cooling systems.

Integrated Monitoring

Enhance data‑center efficiency with real‑time data collection and reduce downtime using the Supervisor’s predictive features.

Future‑Proof Architecture

Reduce infrastructure investments as technology evolves and improve overall TCO.

Improved Compatibility

Seamless integration with single‑phase direct liquid cooling infrastructures.

Case Studies

Mission-critical Performance With Less Energy
Challenge
  • Servers were overheating and throttling under intense compute loads.
  • Traditional cooling methods could not keep up with performance demands.
  • As a government lab, Sandia needed a long-term, energy- efficient solution that aligned with strict budget and sustainability goals.
Solution
  • GemaTEG installed DaTEGs on Intel Xeon 8480+ Platinum chips.
  • The system included Supervisor software for remote monitoring, on-site setup, and staff training.
  • The approach was collaborative, allowing Sandia to customize and test the solution on-site.
Expected Results
  • The deployment is expected to help Sandia maintain stable, high-performance compute under extreme thermal stress.
  • Server uptime should improve, with less heat-related wear and tear.
  • The partnership reflects Sandia’s pioneering spirit and strengthens collaboration with GemaTEG.
Commercial Potential

Ideal for national labs and research institutions facing similar high-intensity workloads. GemaTEG offers a testable, adaptive solution where failure is not an option.

A Strategic Leap Beyond Liquid Cooling
Challenge
  • Heat-related server failures were common, with traditional cooling unable to keep up.
  • Leonardo needed a cost-effective, high-performance solution to support growing compute demands.
  • Energy limits, sustainability targets, and tight public-sector budgets made this a long-standing, high-stakes problem.
Solution
  • Leonardo deployed GemaTEG’s Supervisor and DaTEGs on a NvidiaGPU platform to test against their existing liquid cooling.
  • The goal: higher performance, better efficiency, and longer hardware life.
  • Software integration added predictive maintenance and advanced energy controls.
Expected Results
  • The project is expected to achieve an average efficiency increase of 30%.
  • Anticipated benefits include longer hardware lifespan, reduced downtime, and lower emissions.
  • Success would position Leonardo as a forward-looking leader in sustainable high-performance computing.
Commercial Potential

Ideal for companies with systems hitting their limits, and under pressure to future-proof their infrastructure with sustainable solutions. It fits existing infrastructure and scales easily.

Contact

GemaTEG introduces the adaptive and intelligent approach to thermal management.

Contact Us