HS Hyosung Information Systems Targets Market With Efficient AI Inference Infrastructure Combining Arm Servers and NPU
TECHWORLD ·
✦ AI Summary
As the use of generative AI and LLMs enters full swing, the center of gravity in the AI market is shifting from model development and training to the inference stage.
HS Hyosung Information Systems said it will target the enterprise AI service operations market through collaboration between its GreenCore Arm server product and NPU partners such as Qualcomm, Furiosa AI, DeepX, and Mobilint.
The company plans to support the entire process of introducing AI inference infrastructure, from customer workload analysis and NPU selection to GreenCore server configuration, integrated verification, PoC, deployment, and operations.
As the use of generative AI and LLMs enters full swing, the center of gravity in the AI market is shifting from model development and training to the inference stage of business and service deployment and operations. Accordingly, enterprise infrastructure requirements are also being aligned with the need to efficiently process inference requests that occur repeatedly. In this process, as constraints such as cost, power, and space limitations come into focus, interest in NPU, along with GPU, is also growing.
In line with this trend, HS Hyosung Information Systems said it will move to target the enterprise AI service operations market with a high-efficiency AI inference infrastructure that combines Arm servers and NPU. The company’s Arm server product is GreenCore, and GreenCore’s defining feature is low power consumption.
HS Hyosung Information Systems is pursuing broad collaboration with major domestic and global NPU companies, including Qualcomm, Furiosa AI, DeepX, and Mobilint. Through this, it aims to provide infrastructure suited to customer environments.
The company explained that when selecting NPU, it does not rely solely on product catalog performance. Instead, it comprehensively analyzes the customer’s AI models, service type, model size, number of concurrent users, target response speed, and data center power and space conditions, then presents a suitable pool of NPU candidates based on the results.
GreenCore is a server product line based on Arm architecture CPUs, jointly developed with domestic Arm server maker XsLab, and it offers an option to maximize the advantages of NPU through Arm server GreenCore. GreenCore highlights low power consumption, low heat generation, and efficient many-core parallel processing, and its target applications are companies and institutions preparing on-premises AI services while considering data centers with power, cooling, and rack space constraints, as well as security and data sovereignty. The company said GreenCore can be an effective option for these users.
HS Hyosung Information Systems said it directly installs the selected NPU into GreenCore and conducts integrated verification of hardware and software compatibility, AI model operation, and operational stability. If needed, it also works with AI solution companies to carry out concept validation (PoC) that reflects actual service environments, with the aim of easing customers’ burden of comparing and verifying multiple NPU products, as well as their burden of directly comparing and verifying manufacturer-specific development tools.
HS Hyosung Information Systems said it plans to support the entire process of introducing AI inference infrastructure. The support scope includes customer workload analysis, NPU selection, GreenCore server configuration, integrated verification, AI solution integration, PoC, deployment, and operations. Through this, the company set out a goal of supporting customers’ AI adoption, stable operations, and expansion.
It also said it plans to organically combine the future Arm server, GPU, NPU, high-performance storage, data lake, AIOps, and AI solution partner ecosystem, and continue strengthening the portfolio of the 'HS Hyosung AI Platform.' Yang Jeong-gyu, CEO of HS Hyosung Information Systems, said the center of AI competitiveness is shifting from model development to stable and efficient operations in real business environments. He also said the company will support high-efficiency inference infrastructure that combines GreenCore with various NPU, and will help customers rapidly implement AI services suited to each workload while taking cost, power, and space constraints into account.
Source: TECHWORLD · Lee Kwang-jae
Original: https://www.epnc.co.kr/news/articleView.html?idxno=406712
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.