0
SK Telecom's headquarters in Euljiro, Seoul. / Photo courtesy of SK Telecom

SK Telecom has expanded its proprietary AI foundation model project consortium by welcoming SK AX and TechnoMatrix as new partner companies.

SKT announced on July 14 that these two firms, which possess deep expertise in industrial AI transformation (AX) and manufacturing-sector AI operations, have joined the elite team to broaden its capabilities in accelerating industrial AI adoption.

The consortium expects the addition of the two companies to significantly boost its capacity to propagate AI foundation models across sectors.

SK AX is an AI transformation specialist that drives AI innovation for clients based on its deep expertise across diverse industries including manufacturing, telecommunications/media, semiconductors, finance, and the public sector. Within the elite team, SK AX will spearhead the identification of B2B AX use cases, drive proof-of-concept initiatives, and build industrial expansion frameworks. The company is currently supporting tasks such as improving business productivity, automated report generation, and agent builder development by establishing a system that allows clients to share and utilize the A.X (A.X+) model.

TechnoMatrix is an AI/ML/FM-Ops (AI, machine learning, and foundation model operations) specialist focusing on optimizing AI model operations. It has accumulated extensive experience in modeling and operational optimization across manufacturing, finance, and IT sectors. The firm is anticipated to contribute heavily to manufacturing AX and operational optimization.

The SKT elite team has also strengthened its collaborative framework, which organically links model development, data curation, domestic AI semiconductor utilization, service validation, academic research, and industrial applications. In particular, the consortium operates its internal collaboration through four dedicated tracks—Model/Infrastructure, Advanced Research, Data, and Service Expansion—to maximize synergies among participating organizations.

Building on this cooperation, the elite team developed its proprietary foundation model, A.X K2, ahead of its secondary phase evaluation. A.X K2 was developed to enhance agentic capabilities, enabling the large language model to comprehend, plan, and execute complex workflows, while simultaneously factoring in the openness and efficiency required for rapid deployment across industrial sites.

Moving forward, the SKT elite team plans to expand the applicability of its proprietary AI model into diverse domains, including industrial AX, public/defense sectors, manufacturing, gaming, mobility, search, security, and biotech.

Currently, the consortium is broadening everyday AI experiences for the public through SKT’s primary service touchpoints, such as A., A. biz, and TMAP. The long-term plan is to evolve these into lifestyle-integrated agents that handle information retrieval, schedule coordination, and public-private service linkages, allowing citizens to naturally experience proprietary AI in their daily lives.

In the industrial domain, the team is opening up expansion possibilities by developing a steel manufacturing AI agent for KG Steel and a manufacturing AI aimed at improving casting and machining process quality for Koneck. The development of a defense-specialized model is also underway. Concurrently, 42dot is advancing the models developed by the SKT elite team to expand their footprint into mobility and on-device environments.

Liner is contributing by constructing high-quality, structured training data that links queries, reference documents, and answers. The company is also gathering user feedback on the SKT elite team's models integrated into its own search and knowledge services.

"By uniting the capabilities of our elite team all the way from proprietary model development to industrial AI transformation, we will contribute to expanding the K-AI ecosystem," said Kim Tae-yoon, Vice President of Foundation Model at SK Telecom.

                                                                                                               Lee Ji-sun