SK Group Chairman Chey Tae-won has issued a warning that the market's optimistic expectation for SK Hynix to achieve over $100 billion in operating profit this year could equally transform into a $100 billion loss. He emphasized that artificial intelligence is consuming everything during this transitional period, with volatility induced by technological change reaching unprecedented levels.

Speaking at the "Trans-Pacific Dialogue" event in Washington D.C. on the 20th, Chey stated that the uncertainty of the AI era has rendered traditional planning cycles meaningless. "Even a one-year plan holds little value, as the changes from the beginning to the end of the year are too vast," he noted. He pointed out that while companies with capital and resources will dominate the AI competition, even industry leaders struggle to confirm if they possess the correct solutions.

These remarks highlight the double-edged sword dilemma facing the semiconductor industry amid the AI wave. Although SK Hynix achieved record performance last year, driven by surging demand for high-bandwidth memory, and saw its market capitalization soar to 690 trillion won, Chey's warning reveals the risk that accelerated technological iteration can instantly alter the market landscape. This poses a significant challenge for the semiconductor industry, which relies on long-cycle investments.

Chey also stressed that the energy demands of AI data centers are becoming a challenge for the entire society. He indicated that AI data center infrastructure in the United States alone requires a $5 trillion investment. He stated that the manner of cooperation between South Korea, the United States, and Japan at this pivotal moment will determine the future order.

Profit expectations have doubled from $50 billion to $100 billion in just a few months. Chey revealed that market forecasts in December last year predicted this year's operating profit would exceed $50 billion, while the latest expectations have surpassed $100 billion. This figure is equivalent to approximately 144.85 trillion won.

This surge in expectation is built upon SK Hynix's strong performance last year. The company achieved sales of about 79 trillion won and an operating profit of 47 trillion won in 2024, setting a historical record. Growth in high-bandwidth memory demand and strong sales of general DRAM were the primary drivers, with the company's market capitalization, alongside Samsung Electronics, helping to push South Korea's KOSPI index into the 5,000-point era.

However, Chey remains highly cautious about this seemingly positive outlook. He stated, "New technology can be a solution, but it can also make everything vanish," emphasizing that in the industrial upheaval triggered by AI, even the most advanced technological advantages can be lost rapidly.

Chey pointed out that AI infrastructure construction is causing structural distortions in the semiconductor market. He said companies need to produce "monster chips" using the most advanced technology, like HBM, which can generate truly massive profits.

Yet, an anomalous phenomenon is appearing in the market. Because AI infrastructure is consuming a vast supply of memory, the profit margins for general memory are now exceeding those of the more technologically advanced HBM products. Chey described this phenomenon as a "distortion," reflecting how the impact of AI demand on the entire semiconductor supply chain has exceeded normal market patterns.

This distortion underscores the strategic dilemma facing semiconductor firms: on one hand, they need continuous investment in R&D for cutting-edge technology; on the other, rapid changes in market structure can make the value of technological advantages difficult to predict.

Chey identified energy supply as one of the most severe challenges of the AI era. He stated, "AI is practically consuming almost all the energy and electricity we need," and warned that failing to properly address power demand would become a massive challenge for the entire society.

Specific data illustrates the severity of the issue. Chey revealed that building a single data center requires about $50 billion, and the capacity demand for US AI data centers is approximately 100 gigawatts, requiring a $5 trillion investment in infrastructure alone, excluding energy costs.

He admitted, "I'm not entirely sure if we have a solution," adding that the situation has evolved to the point where a 1-gigawatt power station must be matched to each data center. This tight coupling of energy and computing power requirements far exceeds the planning capabilities of traditional industries.

In response to the energy challenge, Chey revealed that SK is preparing a new integrated solution involving the simultaneous construction of AI data centers and power stations. He said that even after building data centers, many future technologies could make them more efficient, and the company will concentrate resources on R&D to enhance competitiveness.

This strategy aligns with the direction of energy infrastructure construction being prioritized by the current US administration. When asked about investment plans in the US energy sector, Chey stated, "We have been doing it previously," but emphasized that AI changes occur over months, while energy projects typically require five-year cycles, creating a serious mismatch in pace.

Regarding the US Supreme Court's ruling on a specific tariff issue, Chey indicated he would need to review the judgment before commenting further.

Chey emphasized that the current transformation is not merely a challenge but a structural reality determining the survival of all players. He stated that at this turning point, how South Korea, the United States, and Japan cooperate will determine the future order. This statement reflects the rising importance of technological alliances and supply chain collaboration amid increasingly fierce competition in the AI and semiconductor industries.

Chey's warning provides a sobering perspective for the market: despite the enormous opportunities presented by AI, the speed of technological iteration, market volatility, and infrastructure bottlenecks mean that even industry leaders face unprecedented uncertainty.