All Content from Business Insider 07月17日 03:09
As AI surges, Schneider Electric is teaming up with Nvidia to help data centers manage their energy use. Here's how.
index_new5.html
../../../zaker_core/zaker_tpl_static/wap/tpl_guoji1.html

 

施耐德电气与英伟达合作,旨在解决人工智能驱动的数据中心能源挑战。通过合作,施耐德电气开发了新的数据中心参考设计,目标是减少约20%的冷却能耗。文章探讨了人工智能对数据中心能源需求的影响,以及施耐德电气如何通过创新的设计和技术,帮助客户构建更可持续、更节能的数据中心。合作包括开发AI导向的能源和冷却策略、AI基础设施,以及提供咨询支持,助力数据中心实现可持续发展目标。

💡 施耐德电气与英伟达合作,应对AI带来的能源挑战。由于人工智能的快速发展,数据中心的数量和能源消耗都在增加,这促使企业寻求更高效的能源管理方案。

🤝 合作的核心是开发数据中心参考架构。这些架构概述了支持数据处理、业务网络、数据存储等技术所需的能源需求和冷却系统,并提供数字化方案和实时洞察,以帮助构建节能型数据中心。

⚙️ 新的数据中心参考设计旨在支持高密度AI集群,并与英伟达的超级芯片协同工作。该设计还支持液冷技术,以更有效地管理大型计算机和处理单元产生的热量。

📉 通过虚拟建模和测试,该设计预计可减少约20%的冷却能耗,并缩短约30%的数据中心设计和建设时间,从而降低成本并提高可持续性。

Schneider Electric aims to help data centers manage their energy more efficiently.

Schneider Electric is a global technology company specializing in energy management for software and services, including data centers and industrial automation. Founded in 1836, the company is headquartered in Paris and operates in more than 100 countries. Some of its customers include Microsoft, Google, and Amazon.

Situation analysis

The use of artificial intelligence continues to expand worldwide. Pankaj Sharma, the executive vice president of data centers and networks at Schneider Electric, said as a result, more data centers will be built, the demand for energy will increase, and the environmental impact will grow.

The International Energy Agency estimates that by 2026, the AI industry will use about 10 times the electricity it consumed in 2023. Along with expanding energy, data centers also face the challenge of incorporating cooling systems that enhance energy efficiency and performance.

In December 2024, Schneider Electric announced a partnership with Nvidia to develop tools to help data center operators address their energy and sustainability challenges related to increased AI use. This includes plans for developing AI-focused energy and cooling strategies and AI infrastructure, and consulting support to help data centers reach their sustainability goals.

The partnership includes what Sharma calls "data center reference architectures," frameworks outlining the energy requirements and cooling systems needed to support data processing, business networks, data storage, and other technology. It also includes digital plans and real-time insights to help build new, energy-efficient data centers.

Sharma said the goal is to help companies improve their digital infrastructure by establishing more sustainable, AI-ready data centers.

Key staff and partners

Sharma told Business Insider that Schneider Electric collaborated with Nvidia on the project. He said Schneider's technology teams and data center engineers and architects worked closely with Nvidia's technical teams on the latest partnership to develop and test solutions. The executive leadership and business development teams were also involved.

Pankaj Sharma is the executive vice president of data centers and networks at Schneider Electric.

AI in action

Schneider Electric co-developed a new data-reference center design with Nvidia. This is a plan for building new data centers that support high-density AI clusters of up to 132 kilowatts per data-center server rack. A cluster is a network of servers that power AI, and a server rack is a structure that houses and organizes IT equipment.

Sharma said the reference plan works with Nvidia's superchips, which are designed for large-scale AI and high-performance processing. The project is also built to support liquid cooling, a process of using water or a liquid coolant to absorb and remove heat from large-scale computers and processing units.

Sharma said the new reference designs build on Schneider Electric's existing products, which help remove heat from data centers.

He said the latest designs developed with Nvidia help Schneider Electric's customers organize the development and building of data centers that meet a company's energy and cooling challenges. Sharma added that the plans can be customized for individual companies' requirements for their AI workloads in a more energy-efficient, sustainable way for the future. He said that data centers are already being built using the design created with Nvidia.

"The intent of the design is not to build a real-life, end-to-end replica of the design, but instead, to use various aspects of it to inform customers' initiatives, providing inspiration and guidance as needed," Sharma said.

Did it work, and how did leaders know?

Using virtual modeling and testing, Sharma said that the reference designs developed with Nvidia could help data centers reduce cooling energy usage by about 20%. The plan can also help shorten the time it takes to design and build new data center projects by about 30%, which also saves companies money and enhances sustainability, he added.

"It's not just speed, it's cost and sustainability," Sharma said. "All of these pieces put together are an advantage for all our customers."

Schneider Electric plans to continue working with Nvidia to develop more reference designs over the next few years, Sharma said. The company is also working to create data center designs and infrastructure that can accommodate even higher levels of energy use, as well as the cooling technology necessary to prevent overheating and manage energy consumption.

Read the original article on Business Insider

Fish AI Reader

Fish AI Reader

AI辅助创作,多种专业模板,深度分析,高质量内容生成。从观点提取到深度思考,FishAI为您提供全方位的创作支持。新版本引入自定义参数,让您的创作更加个性化和精准。

FishAI

FishAI

鱼阅,AI 时代的下一个智能信息助手,助你摆脱信息焦虑

联系邮箱 441953276@qq.com

相关标签

施耐德电气 英伟达 数据中心 人工智能 能源管理
相关文章