Nitrogen-Liquid Dual Mode Cooling Module
specifically designed for high-load AI GPU such as the NVIDIA GB200/GB300
While liquid cooling technology offers stable operation, its slow heat dissipation under extremely high loads can easily lead to performance bottlenecks.
Nitrogen cooling, although providing extremely rapid cooling, consumes a huge amount of energy, resulting in excessively high operating costs and hindering large-scale commercial application. This invention resolves the conflict between "high-performance operation" and "cooling costs" for servers, preventing servers from throttling or becoming damaged due to overheating.
The module dynamically switches cooling modes based on server load, maintaining the server stably at 45°C. Its heat dissipation speed is 2.5 times faster than traditional single liquid cooling systems, achieving a cooling efficiency of 5°C per second.
This invention solves a major flaw in existing technologies that cannot balance heat dissipation speed and cost-effectiveness. Traditional systems often have to choose between "insufficient heat dissipation (single liquid cooling)" or "excessive cost (single nitrogen cooling)".
Hybrid Dual-Mode Architecture: Integrates nitrogen-cooling and liquid-cooling technologies within a single physical structure using specific space ratios (e.g., 4:6, 3:7, or 5:5) to balance stability and speed.Extreme Cooling Performance: Achieves a cooling rate of 5°C per second—2.5 times faster than single-mode liquid cooling—and maintains server temperatures at a stable 45°C. AI-Driven Dynamic Switching: Utilizes an artificial intelligence smart control platform to automatically switch between liquid, nitrogen, or dual-mode cooling based on real-time GPU load and temperature fluctuations.Rapid Response Control: Features precise timing where liquid cooling flow adjustments are completed within 10ms and mode switching occurs within 100–300ms with 99.8% sensitivity. High-Efficiency Resource Recovery: Includes an integrated recovery system that achieves a 60% recovery rate for gaseous nitrogen and a 70% rate for the liquid coolant, resulting in an overall system recovery efficiency of 67%. Layered Precision Design: Employs a multi-layered structure (Cooling, Liquid Cooling, Nitrogen Cooling, and Evaporation layers) within a compact 140mm140mm40mm module to maximize heat conduction and minimize space.Predictive and Feedback Logic: Analyzes historical heat load curves to start high-efficiency cooling in advance (predictive cooling) and automatically corrects parameters based on real-time results. High-Precision Monitoring: Incorporates a sensor array with minimal error margins for temperature ( ±1°C) and pressure (± 0.02 MPa) to ensure operational safety.
Industries where the invention can be useful?
AI Server,An estimate of the total addressable market?
With its 2.5 times faster cooling speed and higher recovery rate, this technology can capture 15-20% of the high-end market share (estimated value of approximately US$2-3 billion).Potential Customers/End Users. Who might benefit?
Foxconn/Ingrasys、Wiwynn、Quanta/QCT、Supermicro、NVIDIADocuments
-
1766149451931_860.pdf
Actions
Added all portfolio
| Country | Current Status | Patent Application Number | Patent Number | Applicant / Current Assignee Name | Title | Google Patent Link |
| TW | Issued | M671887 | M671887 | Dollarchip Technology Inc. | Nitrogen-liquid dual mode cooling module | Google patent link |
You may also like the following patent