Large-Scale Gas Power Project in Texas, USA

This is a large-scale power plant with three units, and the total investment is estimated to reach approximately $22.3 billion. Given the characteristics of the Texas market, it is highly likely that nuclear power plants would be constructed in other regions.

Due to insufficient transmission and distribution networks in the U.S., connecting power plants takes about five years. Therefore, since last year, large-scale AI data center operators have been pursuing a strategy of building power plants next to or as close as possible to their data centers to improve the efficiency of transmission and distribution network construction. Gas power plants are advantageous for concentrated supply of large capacities.

As a result, electricity generated from natural gas can also be supplied at a low price. Wholesale electricity prices in Texas are maintained at 60 won per kilowatt-hour, and retail prices are below 100 won.

Currently, global gas turbine suppliers have already booked all their orders for the next five years. This has led to a 4-5 fold surge in gas turbine prices over the past four years. Korea is highly regarded for its strong chances in this gas power project, thanks to its package supply capabilities and relatively ample production capacity.

This indicates that more than half of the increase in electricity consumption is coming from artificial intelligence (AI) data centers. The surge in data center power demand in the Texas region highlights the urgency of expanding energy infrastructure.

Large-scale data center operators in the U.S. are demanding full power supply within a minimum of two years, and a maximum of three years. This is a very tight schedule considering the typical time required for power plant construction. Therefore, optimizing project execution to supply power in line with data center demands is considered the biggest challenge for this project.

The U.S. exports approximately 100 million tons of liquefied natural gas (LNG) annually and is projected to increase exports to 180 million tons within the next 3-4 years. Even if the U.S. increases gas production, continuous export growth could lead to domestic gas price increases. Analysis suggests that in the long term, after about 10 years, low U.S. gas prices may not be sustainable at current levels and could rise significantly. In this regard, the Trump administration's policy of expanding development on public lands and the Democratic Party's environmental regulations could act as variables.

Rare earth elements are essential for manufacturing solar panels and wind turbines, but their proportion of the total cost is not high. Some companies, such as Tesla, are developing technologies to reduce their reliance on rare earth elements through the development of high-performance magnets. Therefore, analysis suggests that an increase in rare earth prices will not significantly raise the overall cost of renewable energy.

The project cost varies greatly depending on what is included, such as financial costs, construction investment, and infrastructure costs. In particular, infrastructure construction costs for water, gas, and power grids, along with a contingency fund of 10-30%, are major factors determining the project cost. For this reason, some projects are estimated to be about 50% higher than the amounts reported in the media.

According to the memorandum of understanding (MOU) signed between Korea and the U.S., if the U.S. proposes a project, Korea has very little room to maneuver and practically cannot refuse. The U.S. wants Korean capital to invest in long-term, high-interest infrastructure projects (power generation, LNG, etc.) that private U.S. investors find difficult to finance. In particular, a peculiar phenomenon has occurred where Korea's financing costs are set lower than U.S. Treasury bonds, leading to high U.S. expectations for attracting Korean capital.

It is unlikely that Korean companies will benefit from construction using U.S. labor. Therefore, Korean companies should aim to supply as much equipment as possible, such as gas turbines, power facilities, and steel materials. Securing a track record in the U.S. through this project can make them sufficiently competitive in global markets, including Europe. In particular, achieving results in the gas turbine sector is expected to be a key stepping stone for global market entry.

Once built, a power plant needs to operate for at least 20 years, and a nuclear power plant for over 60 years, to achieve economic viability. However, AI data center servers typically have a lifespan of only 5-10 years. This mismatch between the long operational needs of power plants and the short lifespan of data center servers poses a risk to long-term power demand sustainability.

Constructing a large nuclear power plant requires a long-term guarantee of at least 40 years. However, big tech companies are reluctant to commit to long-term contracts exceeding 20 years, leading to a conflict of interest between power infrastructure investment and service period. This poses a significant challenge to the advancement of nuclear power projects, which require stable, long-term power supply.

This order is partially retroactive, with investigations into previously installed equipment already underway. This is interpreted as a preemptive measure to prepare for situations where maintenance and replacement may become impossible in the future.

Especially in data center control technology, foreign companies have already dominated the market. The single market structure centered on Korea Electric Power Corporation acts as a factor limiting the introduction of new technologies, making it an urgent task to secure competitiveness in power services and advanced control technology fields.