Schneider Electric: In AI Semiconductor Fabs, 'Power and Cooling' Determine Production Competitiveness
TECHWORLD ·
✦ AI Summary
As demand for AI semiconductors and high-performance computing expands, semiconductor manufacturing competitiveness is shifting to a combination of power, cooling, automation, and data-driven operating capabilities.
Schneider Electric said it will push ahead with an integrated infrastructure covering power, cooling, automation, and data-driven operations, and support energy savings and improved operating efficiency at domestic semiconductor fabs.
It also plans to pursue stable operation and greater efficiency at semiconductor fabs and AI factories through Galaxy VXL, utility optimization, and digital twin collaboration with NVIDIA.
As demand for AI semiconductors and high-performance computing expands, the factors that determine competitiveness in semiconductor manufacturing are shifting. As higher power demand resulting from more advanced production equipment and process technologies continues, along with higher cooling loads, the need to secure a stable operating environment is growing, and the importance of improving energy efficiency across the factory and of infrastructure strategy is also increasing.
Against this backdrop, semiconductor production competitiveness is changing in a direction where it is determined not only by process technology but also by management capabilities that combine stable power supply, cooling, automation, and data-driven operations. On the 8th, Schneider Electric said it will push ahead with an integrated infrastructure covering power, cooling, automation, and data-driven operations, and strengthen semiconductor manufacturing competitiveness in the AI era.
Schneider Electric said it will secure stable power supply based on high-efficiency power protection technology and pursue utility optimization, while setting energy-use reduction and lower downtime risk through the use of digital twins as goals. Through this, it plans to strengthen both the stable operating environment of semiconductor fabs and energy efficiency management.
Seong Dong-jun said that in the AI era, semiconductor equipment and process technology must improve productivity while power grids, cooling, automation, data, and AI-based operating capabilities are also advanced in tandem. He said Schneider Electric has experience applying solutions for global semiconductor companies and expressed its willingness to support energy savings and improved operating efficiency at domestic semiconductor fabs.
Because semiconductor manufacturing plants require precise equipment control and continuous process operations, stable power supply is essential. Accordingly, when expanding power infrastructure, companies must consider not only increased supply capacity but also power quality, energy efficiency, and operational continuity that allows production to continue during maintenance.
Schneider Electric presented Galaxy VXL as a product that responds to these needs. Galaxy VXL is a 3-phase UPS that supports 500~1250kW (400V) and is based on a high-density design and fault tolerance. The product also supports core IT infrastructure for large-scale data centers, cloud services, and service providers.
Galaxy VXL delivers up to 97.5% efficiency in double-conversion mode, and up to 99% efficiency in its patented high-efficiency mode, eConversion. It also uses Live Swap technology, allowing power modules to be replaced without stopping the device. Schneider Electric said this enables Galaxy VXL to secure operational continuity during maintenance and capacity expansion.
Schneider Electric plans to apply the power protection technology it has used in data centers to domestic semiconductor fabs. The goal is to reduce energy loss in the power supply process and support stable operation of key equipment. In addition, it plans to expand the application of utility optimization solutions to domestic semiconductor fabs.
To improve energy efficiency in semiconductor manufacturing plants, operational efficiency in utility systems such as cooling, HVAC, and compressed air must also be improved alongside production equipment. For this reason, Schneider Electric proposed Chiller Energy Management System, HVAC Ventilation Simulation, AI-based autonomous operations, and wastewater treatment solutions. The company plans to improve energy efficiency across the plant by analyzing energy use and operating status by equipment, identifying opportunities for improvement, and reflecting actual operating conditions.
This approach was presented based on Schneider Electric's experience of providing related solutions after conducting long-term PoC with a U.S. semiconductor company. Through this, the company expects about 10% power savings in the utility sector and about 2.5% power savings across the fab.
The view was presented that it is necessary to expand efficiency improvements from individual equipment to optimization of the entire plant. To this end, the company said it needs technology to understand interactions between power and cooling systems, as well as technology to pre-review changes in various operating conditions.
To address these challenges, Schneider Electric and NVIDIA are collaborating. The two companies have set their focus on implementing digital infrastructure across the design and operation of AI factories.
The two companies will use the NVIDIA Omniverse AI factory digital twin blueprint and combine it with ETAP power digital twin technology. Through this, they plan to implement an integrated simulation of power, mechanical, thermal, and network elements.
The application approach is Grid to Chip. The scope of this approach runs from the grid to the chip, and it has the capability to precisely model dynamic loads by AI workload.
This structure also makes it possible to support real-time monitoring and predictive maintenance. Through this, the strategic goals are to improve AI factory infrastructure energy efficiency and enhance operational stability.
In semiconductor manufacturing, the combination of NVIDIA Omniverse and ETAP power digital twins is supporting the buildout of TSMC AI factory. This approach can be linked with existing semiconductor fab systems and can also use the operating environment of manufacturing facilities that are already in operation, allowing digital management capabilities to be expanded.
Schneider Electric plans to apply these global implementation experiences to domestic semiconductor manufacturing environments. To that end, it aims to combine power, cooling, automation infrastructure, and data- and AI-based operating technologies to support the performance of production equipment while increasing productivity and energy efficiency at the same time, and to contribute to strengthening semiconductor manufacturing competitiveness in the AI era.
Source: TECHWORLD · Park Gyu-chan
Original: https://www.epnc.co.kr/news/articleView.html?idxno=407926
References
This article was produced with the help of an automated content generation algorithm.
Source: TECHWORLD
View originalThis article was summarized and organized by BizCrush based on the original article from TECHWORLD. For exact quotations and full details, please refer to the original article.