Intense Competition Among Industries for 'Core Partners' in National AI Projects

As the government accelerates efforts to secure proprietary AI technology, competition for alliances and mergers among industry players is intensifying. To gain an early lead in sovereign AI initiatives, a diverse range of technology companies are leveraging their unique strengths to form strategic partnerships.

Notably, government-led national AI projects such as the "AI Foundation Model for Cybersecurity (Security Foundation Model)," "Proprietary AI Foundation Model (Proprietary Foundation Model)," and "Everyone's AI," all overseen by the Ministry of Science and ICT, are characterized by the active participation of companies from different industries. Since the scope ranges from AI model development to security and public-facing services, a wide spectrum of technical expertise is required.

The Security Foundation Model project sees a consortium of SK Telecom and Naver Cloud vying for the position of the national security AI provider, with a final selection expected as early as September. For the Proprietary Foundation Model, Upstage, SK Telecom, and LG AI Research are in the final phase of competing to have their proprietary AI models recognized for global competitiveness, while "Everyone's AI" sees multiple companies aiming to expand public services powered by Korean AI models.

As participation in national projects itself is regarded as proof of technological prowess, there is heightened industry interest in the composition of each consortium. The alliances companies form and the specific technology firms joining each camp have become another standard for assessing the competitiveness of these consortiums.

In the IT industry, there has reportedly been behind-the-scenes competition among lead companies to recruit firms with outstanding expertise in specific fields. Some companies are even said to have been approached with offers to join more than one group at the same time.

A review of company rosters reveals some have joined multiple projects. For instance, security company RaonSecure is participating in both the Security Foundation Model and the Proprietary Foundation Model. Of note, RaonSecure has maintained cooperation in both projects with Upstage, with whom it signed a business agreement in April.

RaonSecure is a company that possesses digital authentication technology implemented in the national mobile ID system. Based on this, the company is set to unveil a solution that gives unique identities to agentic AI, controls authentication and permissions, and utilizes its white-hat hacker groups for blind penetration testing and vulnerability assessments. This ability to cover defense technologies for controlling AI identity and permissions, as well as offensive, real-world security capabilities, is considered a key reason why RaonSecure is sought after by multiple consortiums.

This trend has also been affected by the recent "Mythos Incident." With Anthropic's security AI model Mythos demonstrating abilities in system vulnerability detection and attack scenario generation, there is a rapidly growing consensus that AI must not only be applied for security but that its identity, permissions, and behavioral boundaries must also be tightly controlled. Such changes are believed to have increased demand for security technologies in the government's recent AI projects.

It is also notable that companies from everyday industries such as retail, housing, education, and finance are joining in large numbers. In "Everyone's AI," firms like Musinsa, Zigbang, EBS, and BC Card, which operate services closely tied to people's daily lives, are participating. As the ability to apply AI models to real industries and services becomes as important as model performance itself, leaders in each sector are regaining attention.

The "AI coalition" competition is happening overseas as well. In Europe, efforts to develop indigenous frontier AI models and industrial applications are giving rise to consortium-based projects involving technology companies, research institutes, and public agencies. Similarly, Japan is connecting industry-specific specialists and data holders through the government-led GENIAC (Generative AI Accelerator Challenge) project.

Looking forward, competition for government AI projects is expected to further intensify efforts to secure "core partners" with strengths in specific areas. Rather than relying solely on the technological capabilities of individual companies, the key competitive advantage will be how robust an ecosystem can be established through alliances.

An industry official commented, "Although government AI projects may appear outwardly to be a contest among major AI and telecom companies, in reality, the roles of specialized firms in each field are substantial. These projects can expand the base of Korea's AI industry and provide new business opportunities for technology companies."

This content was produced with the assistance of AI translation services.