[Initial Insight] A Bittersweet AI Alliance with Overseas Big Tech

Korean Companies Form Consecutive Partnerships with Global Big Tech Firms
"The Uncomfortable Truth: We Must Meet All Big Tech's Demands"
An Inevitable Choice for Korea, Now in a Subordinate Position
Urgent Need for AI Technology Independence and Talent Acquisition

These days, in South Korea's artificial intelligence (AI) industry, partnering with overseas big tech companies is a trend. Kakao collaborated with OpenAI (February 4), LG Uplus launched an AI initiative with Google (March 4), KT established an 'AI Transformation (AX) Delivery Center' with Microsoft (March 5), and Naver Cloud began targeting Southeast Asia's AI market with NVIDIA (March 21). Although the location has not yet been disclosed, SK Telecom's construction of a data center with a foreign big tech company (March 3) is another notable collaboration.

The continuous news of leading domestic companies partnering with world-renowned global firms raises two questions. Why are global big tech companies joining hands with Korean companies? When will we succeed in AI independence? First, there is an uncomfortable truth about collaborating with big tech. A senior executive in the IT industry, whom I met recently at a dinner, confessed, "When negotiating to work together on AI projects, it's like they want us to lay everything bare, down to the last detail. (Overseas big tech companies) are not cooperating to give us gifts." In other words, what sounds like an alliance is, in reality, a contract between a dominant and a subordinate party.

Another telecom executive was even more blunt: "Big tech companies abroad don't just contact one domestic company. They reach out everywhere and compare. There are widespread rumors that one overseas big tech company secured a contract to be the first to use a data center built near Seoul by one of the competitors." We are in a position where we have to reveal all our cards and wait to be chosen, and the choice is theirs. Yet, the reason we rely on global big tech is because our AI technology is lacking.

This happened a month ago at the Mobile World Congress (MWC) held in Spain. During a press conference with a domestic tech company, a question was raised: "Aren't we becoming too dependent on overseas big tech?" The company’s CEO responded, "Even the U.S. depends on other countries for what it cannot do. We were dependent on Japan in 1936 because we lacked strength. It would be better not to be dependent, but we have no choice but to improve our capabilities. We must learn quickly, use quickly, and catch up quickly."

The implication is that learning technology is more important than pride. If we can lay the foundation for AI technological independence, the current dependence is an unavoidable process. As he hopes, overseas big tech companies must also develop AI technologies unique to us that cannot be ignored. This could be a large language model (LLM) or a new AI service developed by improving open source. Such a consensus has already formed in industry, academia, and politics. Once the supplementary budget passes in the National Assembly, the government's goal to purchase 30,000 graphic processing units (GPUs) to build a national AI computing center will also receive a green light.

We also need to attract AI talent. No matter how well-equipped the computing resources are, if there is no one to utilize them to produce results, it is all for nothing. In the 1960s, NASA brought in IBM supercomputers to avoid falling behind the Soviet Union in the space race, but they were left unused because no one knew how to operate them. When scouting AI talent from abroad, the government can support companies by subsidizing labor costs or offering tax incentives. Like Japan, research funding can be provided to help overseas AI researchers settle in. These are some measures the Ministry of Science and ICT, currently developing AI talent attraction policies, should consider. To lead rather than be led, there is still a long and busy road ahead.