• Share:
• • Viber Share
  • Viber Download

• Share:
• • Viber Share
  • Viber Download

Rakuten Group Corporation (hereinafter referred to as "Rakuten") announced that it will strengthen its collaboration with Ampere Computing LLC, a semiconductor design company in the United States, to reduce the power consumption of data centers in order to achieve sustainable and efficient AI computing. As a result, it has decided to introduce its highly power-efficient platform for the first time in Japan.

Since 2023, Rakuten has been conducting verification experiments utilizing CPUs provided by Ampere in its own data centers. Through these experiments, Rakuten has successfully achieved a 36% reduction in power consumption per rack and an 11% reduction in space utilization efficiency in its internal container services within the Rakuten Cloud, demonstrating high power efficiency.

Rakuten joined the international initiative "RE100" in 2019 and is working towards achieving carbon neutrality by substantially reducing greenhouse gas emissions from its entire group's business activities, including its consolidated subsidiaries for the fiscal year 2023, by 2024. Rakuten is committed to realizing a sustainable society. With data centers accounting for a high proportion of Rakuten's total electricity consumption, and the increasing need for servers due to the expanding use of AI, power savings in data centers have become increasingly important. With the enhanced collaboration between both companies, Rakuten aims to further improve power efficiency in data centers after 2025 and expand the use of Ampere's platform.

Ampere is a semiconductor design company focused on the new era, leading the future of computing with its innovative CPU design approach that focuses on high performance, high power efficiency, and sustainable cloud computing. Ampere's AI inference provides high cost-performance in the cloud and delivers excellent value in AI inference computing.

Hironori Yoshida, Vice Director of the Cloud Platform Division, an executive officer at Rakuten, commented as follows: "With the increasing demand for computing resources due to the introduction of AI, the operation of cloud environments requires consideration of foundations and means for achieving more sustainable computing. The verification experiments with Ampere have yielded very positive results, and I look forward to continuing the collaboration to realize a highly power-efficient and sustainable data center."

Mr. Shawn Barley, Chief Evangelist and VP of Business Development at Ampere, commented as follows: "Through this collaboration, both companies aim to achieve the scale-out function of processing in Rakuten's in-house cloud environment, focusing on reducing power consumption and improving performance. As AI becomes a large-scale application, we will continue to explore attractive and cost-effective solutions together."

Rakuten will continue to strive for the industry and society's development through the utilization of IT technology under the mission to "empower people and society through innovation", aiming for the realization of industry and society development and a sustainable society.

(Note 1) Refers to virtualization technology that packages and builds the environment required for running applications.
(Note 2) A comparison of space and power consumption in Rakuten's private cloud container service.
(Note 3) Abbreviation for Renewable Electricity 100%. An international collaborative initiative operated under a partnership between the international environmental NGO "The Climate Group" and the NGO "CDP", which promotes climate change information disclosure.
(Note 4) The total of Scope 1 emissions (direct greenhouse gas emissions by oneself) and Scope 2 emissions (indirect emissions related to electricity, heat, and steam supplied by others) calculated and obtained third-party assurance according to the international standard "GHG Protocol" when calculating and reporting greenhouse gas (GHG) emissions.
(Note 5) In order to reduce greenhouse gas emissions, efforts are made across the group to promote energy saving and introduce renewable energy. For greenhouse gas emissions that cannot be fully reduced, carbon offsets are implemented by investing in reduction activities to offset the remaining emissions.
(Note 6) Refers to the function of increasing the number of servers and performing distributed processing to enhance the overall processing capacity of the system.

※Company names, product names, service names, etc. mentioned are registered trademarks or trademarks of their respective companies.

The end

• Year
• Month
• Category

• Year
• Month
• Category