Summary of Key Points
SK Hynix has been very active recently: on one hand, it is rushing to restart its Dalian NAND factory, which has been shut down for four years, before the expiration of the US equipment license at the end of this year; on the other hand, it plans to build a memory chip factory in Miyagi Prefecture, Japan; it also has long-term expansion projects in South Korea and the United States. Behind these moves lies a complex game of interests among various parties. The US aims to lock in SK Hynix's advanced production capacity within its allies (the US or Japan) through equipment restrictions. China, with its existing factories, workers, and supply chain, is an irreplaceable production base for SK Hynix in the short term. Japan, on the other hand, is offering substantial subsidies and advantages in materials and equipment to attract investment. SK Hynix's "three-pronged strategy" reflects the difficult balance between political pressure (from the US) and industrial realities (China's production efficiency and Japan's material advantages), highlighting the reality that the global semiconductor supply chain is not easily disrupted by geopolitics.
Detailed Analysis
1. The Urgent Restart of the Dalian Factory: Seizing the Opportunity Before the License Expires
Why is SK Hynix accelerating the restart of its Dalian factory now? There are two key reasons:
- Capitalizing on Rising Memory Chip Prices to Lock in Production Capacity: In the past two years, demand for AI and mobile devices has surged, driving up memory chip prices significantly. SK Hynix wants to expand production to earn more money. With the expansion of the Dalian factory, combined with the existing factory, the total production capacity will reach 150,000 chips per month, nearly half of its global NAND production capacity. Dalian is also its only overseas NAND front-end manufacturing base (the front-end is the most critical part of chip production).
- Racing Against the US Equipment License Deadline: Last year, the US revoked the equipment import license for SK Hynix's Chinese factory, but later granted a temporary license until the end of this year. After this period, it is uncertain whether advanced equipment can be imported. Therefore, SK Hynix is moving the equipment in November to avoid missing this window.
More importantly, SK Hynix's new projects in South Korea and the US will not be operational until 2028, so the production gap in the coming years will have to be filled by the Dalian factory—after all, it cannot let its customers wait.
2. Building a Factory in Japan: Putting aside Past Rivalries for Money and Materials
South Korea and Japan have been competing in the chip industry for many years (for example, Japan once restricted the export of photoresists to South Korea). Why does SK Hynix now dare to build a factory in Japan?
- Japan's Subsidies Are Very Attractive: Japan has invested heavily to attract chip companies. For instance, TSMC received a subsidy of 476 billion yen for its project in Kumamoto, and Micron received 500 billion yen. SK Hynix will likely get a share of these subsidies, which can significantly reduce construction costs.
- Japan's Material and Equipment Advantages: SK Hynix relies on Japanese companies (such as Shin-Etsu Chemical and Tokyo Institute of Technology) for silicon wafers, photoresists, and packaging materials. Building a factory in Japan ensures a stable supply chain and reduces transportation costs.
However, there are challenges: South Korean public opinion may be dissatisfied ("Why build a factory in a competitor's country?"), and the US may be unhappy, as it had hoped SK Hynix would build in the US.
3. The US's "Double Standard": Trying to Keep Production Capacity within Allies
The US's attitude towards SK Hynix is contradictory:
- Using Equipment Licenses as a Lever: The US's license for SK Hynix's Chinese factory is only valid until the end of the year, aiming to force the company to shift its future advanced production capacity to the US or Japan (both are allies), preventing China from acquiring advanced technology.
- Pressuring SK Hynix to Build in the US: The US has explicitly demanded that South Korean semiconductor companies build storage factories in the US and is dissatisfied with South Korea keeping a 800 trillion won semiconductor cluster within its borders, believing South Korea is not giving priority to the US in terms of "scarce production capacity."
However, the US cannot completely cut off equipment supplies to SK Hynix's Chinese factory, as AI companies (such as those producing HBM and NAND) rely heavily on memory chips. If Chinese production capacity were to stop, the global supply chain would be disrupted, affecting US companies as well. This reflects the US's internal contradiction: it wants to restrict China but cannot do without China's current production capacity.
4. SK Hynix's Balancing Act: Neither Side to Offend
SK Hynix's strategy seems to have multiple advantages, but it is actually forced by circumstances:
- It Cannot Abandon China: China provides ready-made factories, skilled workers, and a complete supply chain, allowing for rapid production (as evidenced by the restart of the Dalian factory). This is something South Korea, the US, and Japan cannot match.
- It Cannot Ignore Japan: Japan's subsidies and material advantages can significantly reduce costs and stabilize the supply chain.
- It Cannot Ignore the US: The US is one of the largest chip markets and holds significant influence over equipment and technology. Offending the US would mean cutting off its own supply.
This dilemma shows that the decades of globalization in the semiconductor industry are not in vain—the supply chain is a complex network where one part affects the whole. Geopolitics can delay factory construction, but it cannot change the industrial reality that companies will go wherever they can produce chips the fastest and most cost-effectively.
In Summary
SK Hynix's global strategy is a result of being forced to make difficult choices between political and market considerations. It also highlights that the globalization of the semiconductor industry cannot be easily reversed by administrative orders. The company that can meet the actual needs of the industry will be the ultimate winner.