Hyundai Motor Group Unveils AX Experience for the First Time Externally, Shares AI Field Applications and GPU Strategy
TECHWORLD ·
✦ AI Summary
Hyundai Motor Group held 'Softeer Tech Talk Off Day 2026' and publicly unveiled, for the first time, the AX experience and on-site application cases it had been pursuing internally.
The event consisted of a keynote speech and three sessions, and introduced plans for companywide expansion of generative AI platforms and data-based adoption, as well as GPU infrastructure expansion in preparation for physical AI.
Hyundai Motor Group also disclosed achievements involving H Chat, H-Launch, its AI agent management system, VIN recognition AI, sales and maintenance AI, and integrated data management.
Hyundai Motor Group said it held the event as an opportunity to publicly unveil, for the first time, the AI transformation (AX) experience and on-site application cases it has been pursuing internally. The group said it recently held 'Softeer Tech Talk Off Day 2026' for domestic corporate IT leaders and media.
'Softeer Tech Talk' is an in-house technology event where Hyundai Motor Group IT employees share projects and achievements. The event has been held every year since it launched in 2023, and this year's fourth event was opened to the public for the first time.
At this event, Hyundai Motor Group released success cases. It also introduced trial-and-error from the AX pursuit process, along with remaining challenges related to AX.
The event opened with a keynote speech by Jin Eun-sook, president in charge of ICT at Hyundai Motor and Kia. It then consisted of three sessions: 'Preparing for the AI Journey,' 'Business Results Achieved Through AI,' and 'The Future of AI.'
Hyundai Motor Group said it is pushing ahead with companywide expansion of its generative AI platform and data-based adoption across the organization. The event also explained the direction of companywide AI expansion.
Hyundai Motor Group said it plans to gradually expand GPU infrastructure in preparation for future physical AI development. At the event, the group linked companywide AI expansion with its plan to expand GPU infrastructure for future physical AI.
Before introducing AI, Hyundai Motor Group first reorganized its collaboration environment and data management system. To that end, it built a work management and knowledge-sharing platform and established a cloud-based collaboration environment. It also centralized management responsibility for data and IT systems that had been dispersed across departments.
On that foundation, Hyundai Motor Group secured about 140 PB of vehicle sensor data from about 12 million connected cars worldwide. It also integrated and managed operational data across multiple affiliates, including customer, quality, and sales data. For integrated management, it operated more than 7,000 data pipelines.
Based on this, Hyundai Motor Group introduced its in-house generative AI platform, 'H Chat,' in 2023. More than 70% of the group's employees are now using H Chat in their work.
Key challenges in putting AI into practice were identified as a common platform, trust from frontline businesses, and a data management system. The common platform includes search, chatbot, and agent functions. Accordingly, new AI services can be applied by combining the needed functions without developing them from scratch.
The company broadened its applications on production lines by integrating AI models that had been operated separately by each site. It has operated a vehicle identification number (VIN) recognition AI to detect vehicle assembly-sequence errors, and it converted this VIN recognition AI from separate models at each plant to a single model. The single VIN recognition AI has been applied to 73 processes in Korea and overseas.
The company also said it applied AI to sales and maintenance work. It enabled AI to propose draft vehicle sales demand plans, and the maintenance AI was also designed to provide information based on official manuals. The company said this allows frontline staff to verify the results produced by the maintenance AI.
The company also revamped its data management system. It consolidated work information by department in one place and allowed each department to keep the information updated directly. The company said this revision shortened internal information search time from about 10 minutes to about 20 seconds based on integrated search.
Park Geun-han, managing director of Hyundai Motor and Kia Machine Learning Lab, said the deciding factor in whether AI tasks succeed was not AI technology itself, but a clear problem definition reflecting on-the-ground issues, a common platform, and a sustainable data management system led by frontline departments.
Hyundai Motor Group will unveil 'H-Launch,' an in-house distribution platform this month that enables AI tools created by employees to be used at the organizational level rather than remaining at the individual level. 'H-Launch' supports shared organizational use of AI work tools created by nondevelopers.
Hyundai Motor Group will also establish an AI agent management system. The system assigns each agent a unique ID and records its activity history. It will also apply a human approval process for tasks that are difficult to reverse, such as data deletion.
Hyundai Motor Group will also expand GPU infrastructure for physical AI development. Starting in the fourth quarter this year, the group will pursue securing operational capabilities for GPUaaS based on large-scale GPU and its own AI platform. The group set a goal of operating services using more than 10,000 GPUs at the Saemangeum AI data center starting in the second half of 2029.
Jin said AX is an ongoing process of exploration that is not completed simply by adopting a particular technology. She also said an organizational culture that quickly applies new technologies and connects them to business results is important.
Source: TECHWORLD · Kim Seung-gi
Original: https://www.epnc.co.kr/news/articleView.html?idxno=408624
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.