최태원 SK 회장이 ‘SK AI 서밋(SUMMIT) 2024’에서 기조연설을 통해 인공지능(AI)이 성공하기 위해서는 AI를 이용한 성공사례와 AI 가속기 및 반도체 공급부족 문제, AI 인프라에 필요한 에너지 문제를 극복해야 한다고 밝혔다.

▲SK Chairman Choi Tae-won is giving a keynote speech at the 'SK AI Summit 2024.'
Need a use case to create a business model with AI
AI Semiconductor Shortage, Jensen Huang Demands 6-Month Advance Payment for HBM Delivery
SK Chairman Chey Tae-won said that for artificial intelligence (AI) to succeed, there must be successful cases of AI use, and overcoming the shortage of AI accelerators and semiconductors, as well as the energy problem required for AI infrastructure.
SK Chairman Chey Tae-won gave a keynote speech at the 'SK AI Summit 2024' held at COEX on the 4th.
In his keynote speech, Chairman Choi Tae-won said that AI poses many difficult problems that leaders from various fields must think about and solve together, and that everyone’s cooperation is needed for AI to positively lead to changes in our lives.
He then said that there are bottlenecks that need to be resolved for the future development of AI, and that first, a use case for what kind of business model to create with AI is needed.
To this end, SK said it is working with global telecommunications companies to create a B2B AI solution called Telco LLM and a personalized AI agent called Adot to provide it to users.
The next bottleneck, he said, is the shortage of AI accelerators and semiconductors.
Chairman Choi Tae-won shared a scene of a conversation with NVIDIA CEO Jensen Huang.Nvidia CEO Jensen Huang also revealed that he had asked for HBM supply to be brought forward by six months.
He also covered the TSMC story, saying that related companies such as SK, Nvidia, and TSMC are working closely together for AI, and that they should join forces to supply AI semiconductors.
He then argued that energy is needed to drive massive computing power based on AI semiconductors, and that building a typical LLM AI data center requires 10 gigawatts of power, which costs a lot of money, and that the power problem must be solved.
Chairman Choi Tae-won said that generating electricity somewhere and using it in the power grid is not suitable for AI data centers, and that an independent form of electricity is needed.
In particular, SK said that as an expert in the energy field, it is researching as much as possible ways to reduce the heat generated in AI data centers, creating low-power semiconductors, and considering ways to reduce energy consumption by applying glass substrate materials.
He also said that it is important to find an independent and sufficient energy source, and that the country is actively developing hydrogen and nuclear energy sources.
He also said that lastly, support at the national level is also necessary, and that in order for our country to lead the AI era, investment in AI infrastructure and training of talent fit for the AI era are most important.