OCI Holdings’ Solar Crisis Overcome and AI Infrastructure Challenge – How U.S. Tariff Policy Changed the Game

By examining OCI Holdings’ third-quarter earnings report, we can simultaneously glimpse the current state and future of the Korean solar industry. It’s shocking that a single U.S. tariff policy resulted in a 65 billion KRW deficit, yet the company’s strategic moves to turn this crisis into an opportunity are even more noteworthy. Particularly, their ambition to reduce reliance on solar energy by venturing into AI infrastructure offers significant implications for the renewable energy sector both domestically and internationally.

According to Chairman Lee Woo-hyun, after the U.S. announced “unprecedented tariffs” in May, nearly all solar companies temporarily halted factory operations. It wasn’t until August that specific measures were finalized, allowing factories to resume operations in September. The fact that a mere two-month shutdown resulted in a 65 billion KRW deficit starkly illustrates the vulnerabilities of the global supply chain. OCI Terasys, for example, has a production capacity of 2,500 tons per month, highlighting the substantial impact of halting such a large-scale facility.

However, the saying “crisis is an opportunity” seems particularly apt in this situation. The U.S. OBBBA (One Big Beautiful Bipartisan Act) law, set to take effect in January next year, is expected to be a game-changer. According to this law, only non-Chinese solar products will receive tax benefits and subsidies, presenting a golden opportunity for companies like OCI Holdings that can produce non-Chinese polysilicon and wafers. In a market where China dominates over 80% of the global solar market, securing a “non-Chinese premium” signifies a considerable competitive advantage.

In this context, OCI Holdings’ acquisition of a 65% stake in the Vietnamese wafer company Neosilicon Technology is a strategically significant move. They plan to produce non-PFE solar wafers with an annual capacity of 2.7GW, and they can double this capacity within six months if necessary, allowing them to flexibly respond to market demand changes. However, the fact that the joint investors are Chinese companies raises some concerns. Although Chairman Lee Woo-hyun explained that it is a joint venture to receive technology transfer and support, it remains to be seen how this structure will be affected in the long term amid escalating U.S.-China tensions.

Seismic Shifts in the Domestic and International Solar Industry

Observing OCI Holdings’ situation provides insights into the broader changes within the domestic solar industry. Major Korean solar companies like Hanwha Q CELLS and Shinsung E&G are facing similar challenges. The solution Korean companies have found between China’s low-cost offensive and U.S. protectionism is the “non-Chinese premium” strategy. Hanwha Q CELLS operates a 1.7GW solar cell plant in Georgia, USA, and Shinsung E&G is expanding its production bases in Southeast Asia.

Globally, the situation becomes even clearer. While Chinese companies like LONGi Green Energy, JinkoSolar, and Trina Solar dominate the global solar market, diversification policies in the U.S. and Europe are creating cracks. The European Union has also reinstated anti-dumping duties on Chinese solar panels starting in 2024, and India is strengthening policies to protect its domestic solar industry. In this trend, the intermediary role of Korean companies is becoming increasingly important.

Particularly noteworthy is the shift in positioning within the solar value chain. OCI Holdings’ announcement to shift its business direction from solar cells to wafers can be understood in this context. Wafers are a relatively technology-intensive area within the solar value chain, where China’s monopoly is high. If they can supply non-Chinese wafers on a gigawatt scale in a market where Chinese companies account for over 95% of global wafer production, they could create significant market value.

The New Challenge of AI Infrastructure

However, the most intriguing part of this announcement is undoubtedly the plan to enter the AI infrastructure business. The goal to make AI infrastructure account for 30% of total sales and profits by 2030 is quite ambitious. Considering OCI Holdings’ current annual revenue is around 3 trillion KRW, this translates to nearly 1 trillion KRW in revenue from the AI infrastructure business alone.

The core of this strategy is leveraging the capabilities of OCI Energy, which develops solar and ESS projects in North America. The plan is to develop AI data centers using OCI’s idle sites, which are equipped with infrastructure like power and water. Their strength lies in having developer capabilities that cover the entire process from site acquisition to permits, design, financing, construction, and operation. Given that two-thirds of the world’s AI data centers are being built in the U.S., with 40% concentrated in Texas, OCI Energy’s Texas base is expected to be a significant advantage.

The growth trend of the AI data center market shows why this strategy is attractive. The global data center market size is expected to grow from $250 billion in 2023 to $550 billion by 2030, with an average annual growth rate of 12%. The explosive increase in demand for high-performance computing due to the spread of generative AI is a key driver. With a single NVIDIA H100 GPU costing over $30,000, the value of data centers that can accommodate such equipment is also rising.

However, the AI infrastructure business is not an easy field. Major cloud companies like Amazon Web Services (AWS), Microsoft Azure, and Google Cloud have already established a strong market presence, and domestically, companies like Naver Cloud Platform and Kakao Enterprise are expanding their AI infrastructure investments. Given the enormous power consumption of AI data centers, stable power supply and cooling systems are crucial, and how OCI Holdings can offer differentiated solutions in this area will be key.

Personally, I find the integrated solution connecting OCI Holdings’ solar-ESS-AI infrastructure intriguing. Considering that AI data centers consume 3-5 times more power than regular data centers, an eco-friendly AI infrastructure combining renewable energy and energy storage systems could have significant competitiveness. In fact, big tech companies like Google, Microsoft, and Meta are actively pursuing the construction of renewable energy-based data centers to achieve carbon neutrality.

Changes are also being detected in the domestic market. SK Telecom has announced a 1 trillion KRW investment to build AI data centers, and LG Uplus is also expanding its AI infrastructure. The government is supporting the establishment of a domestic AI infrastructure ecosystem through the K-Cloud policy. In this trend, it will be interesting to see how OCI Holdings’ entry into AI infrastructure will impact the domestic market.

However, there are also concerns. The AI infrastructure business is much more complex and technically challenging than the solar business. It requires comprehensive capabilities, including hardware infrastructure, software platforms, security, and operational know-how. Considering that OCI Holdings has been primarily focused on material-centric businesses until now, significant organizational capacity expansion and the acquisition of specialized personnel will be necessary for success in the AI infrastructure business.

Additionally, the rapid pace of change in AI technology presents substantial investment risks. While large-scale language models based on transformers are currently mainstream, entirely different architectures could emerge in a few years. Making long-term infrastructure investment decisions amid such uncertainty requires considerable courage. However, on the flip side, this uncertainty also presents an opportunity to differentiate from existing leaders.

Ultimately, the success of OCI Holdings will depend on how effectively they can apply the project development capabilities accumulated in the solar business to the AI infrastructure field. Experience in site acquisition, permits, and financing will undoubtedly be helpful, but AI-specific technology and operational know-how are areas that need to be newly developed. It will be crucial to identify specific demands through consultations with clients and develop tailored solutions.

Personally, I believe OCI Holdings’ strategy could serve as a model for Korean companies adapting to global new technology trends. They are demonstrating strategic flexibility in responding to geopolitical risks while discovering new growth engines based on the strengths of their existing business. While there will undoubtedly be many challenges during the execution process, if they achieve the goal of 30% AI infrastructure revenue by 2030, it could serve as a good benchmarking example for other domestic companies.

This article was written after reading a ZDNet Korea article and adding personal opinions and analysis.

Disclaimer: This blog is not a news outlet, and the content is the author’s personal opinion. The responsibility for investment decisions lies with the investor, and no responsibility is taken for investment losses based on the content of this article.