ASUS Combines AI Infrastructure and Governance at 'AI Tech 2026'
TECHWORLD ·
✦ AI Summary
ASUS Korea held 'AI Tech 2026' on the 3rd and shared strategies and insights on AI factory construction and operation, agentic AI, and AI governance.
ASUS established partnerships with NVIDIA, Aleria, AMD, Crusoe, Foxlink, IBM, Intel, Samsung Electronics, Schneider Electric, and WD, and presented a strategy to integrate accelerated computing, networking, storage, deployment software, and AI operations governance into a single platform.
At the event, next-generation AI infrastructure products including 'ASUS AI POD' and 'XA VR721-E3' were introduced, along with ways to pre-simulate and optimize AI factories and monitor and manage them after deployment.
ASUS unveiled an integrated infrastructure strategy for building and operating large-scale AI factories that goes beyond the scope of AI systems. The company said it is taking a strategic approach that extends from AI system units to the construction and operation of AI factories.
ASUS Korea held 'AI Tech 2026' on the 3rd. At 'AI Tech 2026,' strategies and insights related to AI factory construction and operation, agentic AI, and AI governance were shared.
ASUS said it has established partnerships with NVIDIA, Aleria, AMD, Crusoe, Foxlink, IBM, Intel, Samsung Electronics, Schneider Electric, and Western Digital (WD). Based on this network of partnerships, it presented a strategy to integrate accelerated computing, networking, storage, deployment software, and AI operations governance into a single platform.
ASUS explained that the integrated platform can shorten time to first token generation, optimize infrastructure efficiency, and accelerate revenue generation from AI-based services. It highlighted the impact on improving enterprise service performance.
CEOs and officials from ASUS, NVIDIA, Samsung Electronics, IBM, Aleria, and Foxlink took the stage as speakers at the event. Participants shared insights on the overall direction of AI technologies, including large-scale AI factories, agentic AI, and AI governance, as well as practical implementation methods.
At the event, Paul Ju, Executive Vice President and Head of ASUS Infrastructure Solutions Business Group, said ASUS is fully embracing the strategic initiative of 'All-in on AI.' ASUS said it is applying this as the embedding of AI across infrastructure, commercial solutions, and innovative physical devices.
Paul Ju said that as AI factories evolve into always-on production environments, infrastructure performance and AI governance cannot be separated. He added that infrastructure is the foundation of AI production.
ASUS also presented a policy of applying AI governance from the AI factory construction stage. To that end, it said it uses advanced systems engineering and NVIDIA DSX blueprints before model deployment to pre-simulate and optimize the entire AI factory before physical infrastructure is built.
Through collaboration with partners including Schneider Electric, AVEVA, and IBM, ASUS provides an integrated digital twin environment. This allows users to review the configuration and operating environment of an AI factory before actual infrastructure is deployed, and to optimize the configuration and operating environment of an AI factory before actual infrastructure is deployed.
After deployment, ASUS monitors and manages the entire system based on ACC and AIDC. It also provides an AI software stack that includes resource management and an MLOps portal, using the stack to improve data center operational efficiency. The software stack plays a role in directly connecting enterprise governance with AI services.
Based on this framework, ASUS plans to expand its support scope from AI infrastructure design and construction to the actual service operations stage. It also unveiled next-generation AI infrastructure products developed in collaboration with NVIDIA.
The products unveiled were 'ASUS AI POD' and 'XA VR721-E3.' 'ASUS AI POD' and 'XA VR721-E3' are large rack-scale platforms based on NVIDIA Vera Rubin NVL72, used for large-scale intelligence and low-cost token generation. ASUS said the products deliver a 10-times performance improvement per watt compared with the previous generation.
The company unveiled enterprise products and introduced its lineup by use case. For enterprise use, it presented 'XA NR1I-E12LR' and 'XA NR1I-E12L' based on NVIDIA HGX Rubin NVL8, saying the two products support high-density AI computing infrastructure for demanding enterprise training, inference, and post-training workloads.
In the agentic AI segment, it newly unveiled the 'XA P2N-E2' server, which applies the NVIDIA MGX platform. In the edge and visual computing segment, it unveiled the 'ESC8000-E12P' server and the ultra-compact 'PE3000N' platform, and it presented the 'UF920-E3-RS24,' based on the NVIDIA STX module, as a product for data foundation purposes in AI computing environments.
It also held technology demonstrations with partners. Foxlink presented a blueprint for advanced industrial automation using autonomous robots, while Aleria emphasized the importance of sovereign AI, which supports localized data control and secure operations.
ASUS said its integrated approach shortens the time it takes customers to move from AI adoption to actual monetization. The company explained that the ultimate goal of this direction is to prevent enterprise AI use from remaining at the investment stage and to help embed intelligence into operations across the business.
To support this, ASUS plans to strengthen cooperation with the global technology ecosystem. It also said it will continue to enhance its capabilities in building and operating AI factories for scalable growth.
Source: TECHWORLD · Park Gyu-chan
Original: https://www.epnc.co.kr/news/articleView.html?idxno=406531
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.