What Was Announced at the Gangwon Province Donghae AI Data Center?
Gangwon Province has taken the first step toward attracting large-scale investment by signing a business agreement with Donghae City and GS to establish the Donghae AI Data Center. The key plan is to develop an AI data center campus with a total capacity of 2.4 GW by 2029, aiming to create Asia's largest single infrastructure. Based solely on the announced figures, this is a significant development capable of shifting the direction of regional industrial policy, as it involves 30 trillion won in direct construction costs and the potential to attract over 100 trillion won in investment, including related industries and follow-up investments.
This plan is not merely a project to construct server buildings, but rather akin to an industrial ecosystem project where power, cooling, telecommunications, operational personnel, cloud services, and the attraction of AI companies operate as a unified whole. In particular, as the demand for AI model training and inference grows rapidly, the location of data centers capable of stably operating high-performance GPUs and large-scale power is becoming directly linked to national competitiveness.
What Asia's Largest 2.4GW Scale Means
The units frequently used when describing the scale of data centers are MW or GW. 1 GW is 1,000 MW, and 2.4 GW is 2,400 MW. Considering that typical large data centers are often referred to in the range of tens to hundreds of MW, 2.4 GW is a very large scale for a single campus.
AI data centers have a higher power density than traditional data centers for internet services. Training generative AI models or providing large-scale inference services requires the dense deployment of high-performance GPU servers, which consume significant power and generate substantial heat. Therefore, the competitiveness of AI data centers is determined not by simple building footprint, but by power availability, cooling efficiency, network latency, and operational stability.
How to easily understand 2.4GW
The figure of 2.4GW may feel somewhat abstract to the general reader. Simply put, it refers to the power capacity required to build a massive AI computing complex in stages. While actual server deployment and operation methods may vary depending on the project phase, customer demand, semiconductor supply and demand, and power grid conditions, the planned figure itself can be interpreted as a scale suitable for targeting global big tech and cloud companies.
| 구분 | Main Content | Interpretation points |
|---|---|---|
| Business location | Donghae City, Gangwon Province | Establishing an AI infrastructure hub by leveraging the East Coast location and industrial base |
| Target scale | Total 2.4GW AI data center campus | Mega-infrastructure that could be cited as the largest single-scale infrastructure in Asia |
| Promotion schedule | Phased development target by 2029 | Phased implementation is important depending on power supply, licensing, and demand attraction. |
| Direct investment | Proposed on a total scale of 30 trillion won | Linked to land, construction, power, facilities, and operational base costs |
| Expansion effect | Expectations of attracting over 100 trillion won in investment | Ripple effects including equipment, cloud, AI companies, and related industries |
Why the keyword "GS AI Data Center" is important
The reason GS is being mentioned in this announcement is that the participation of a major private conglomerate acts as a factor in enhancing business credibility. Success in a mega-data center is difficult with only the securing of land and buildings. Long-term power procurement, financing, operational know-how, attracting global customers, and energy efficiency management are all required simultaneously.
We will need to verify future detailed agreements and investment plans to determine exactly how the GS AI data center will materialize into a specific business structure. However, if large-scale private capital is combined with administrative support from local governments and regional industrial sites, there is a possibility that a new regional growth model centered around the data center could be created.
Why the East Sea: Four Criteria for Locational Competitiveness
The location of an AI data center is not determined simply by the size of the land. Important factors include the capacity for large-scale power consumption, the ability to enhance cooling efficiency, connectivity with communication networks and industrial complexes, and the ability to secure acceptance from the local community. The plan for the AI data center in Donghae, Gangwon Province, must also comprehensively evaluate these conditions.
1. Power Infrastructure and Energy Procurement Potential
One of the biggest cost items for an AI data center is power. For a 2.4GW campus, securing a stable power supply is crucial to the success or failure of the business. Since data centers operate 24 hours a day, the risk of momentary power shortages or blackouts can immediately lead to service disruptions.
- Grid connectivity: Transmission and substation facilities capable of handling large-scale loads are required.
- Electricity price competitiveness: A key factor determining long-term operating costs.
- Renewable Energy Integration: Global customers often require carbon-neutral power.
- Reserve Power and Stability: Emergency generation, redundancy facilities, and power quality management are essential.
2. Cooling Efficiency and Natural Environment
Cooling technology is critical for AI servers because they generate massive amounts of heat. Various methods, such as air cooling, liquid cooling, and immersion cooling, are used, and recently, liquid cooling or immersion cooling is gaining increasing attention for high-performance GPU servers. An East Coast location is worth considering when formulating a cooling strategy in terms of climate, marine environment, and the feasibility of securing industrial water.
However, for seawater cooling or large-scale water usage to be actually implemented, it must undergo rigorous environmental impact assessments, water quality management, public opinion consultations, and licensing procedures. An image of an eco-friendly data center alone is not sufficient; trust can only be gained by publicly managing cooling water usage, emissions, and energy efficiency indicators.
3. Site Size and Expandability
Mega-scale campus-type data centers are often expanded in stages to meet demand rather than being completed all at once. Therefore, not only the initial site but also long-term expansion potential, surrounding road networks, equipment transport routes, and space for power facilities are important. The fact that Donghae City possesses an industrial base and can be connected to logistics hubs on the East Coast can be viewed positively.
4. Communication Networks and Global Connectivity
For data centers, communication networks are just as important as power supply. High-speed networks and low latency are required for the movement of AI training data, the provision of cloud services, and the operation of real-time services for enterprise customers. In particular, to attract global cloud customers, connection strategies involving major domestic internet switching networks, corporate demand in the metropolitan area, and overseas submarine cables must be considered together.
Ripple effects that could lead to the local AI infrastructure industry
The reason the AI Data Center in Donghae, Gangwon Province, is attracting attention is not solely due to the large scale of the investment. Once the data center becomes fully operational, it could have various ripple effects on the local industrial ecosystem. The key is to design the project so that it does not stop at boosting the construction industry but connects to a long-term AI infrastructure industry.
Direct effects of the construction, power, and equipment industries
In the initial phase, various processes are carried out, including civil engineering, construction, electrical work, mechanical installations, cooling systems, security systems, and firefighting systems. During this process, opportunities for local companies to participate arise, and the demand for related materials and equipment increases. However, because mega-scale data centers require high technical standards, local companies must possess capabilities in quality, on-time delivery, and safety management to participate.
- Increased demand for companies related to electrical construction and substation facilities
- Expansion of technical cooperation in the fields of cooling, HVAC, piping, and water treatment
- Opportunities to Establish Security, Disaster Prevention, and Control Systems
- Increased demand for local service industries such as accommodation, food, and transportation during the construction period
Specialized personnel required during the operation phase
Data centers require a steady stream of operational personnel even after completion. Roles are diverse, including server operation, power management, cooling system inspection, physical security, network management, disaster response, and facility management. In particular, because AI data centers handle high-performance computing equipment, there may be a demand for not only simple facility management staff but also cloud, network, and system engineers.
If local universities, specialized high schools, and vocational training institutions establish educational courses on data center operations and AI infrastructure, it will provide a reason for young talent to remain in the region. This could also help alleviate the issue of youth outflow that Gangwon Province has long struggled with.
Attracting AI companies and cloud ecosystems
While data centers are important in themselves, their greater value comes from the services operating on top of them. If AI model developers, cloud service providers, security firms, data processing companies, and manufacturing AI solution providers gather in the vicinity, the fundamental nature of local industries can be transformed. To foster the regional AI infrastructure industry, Gangwon Province must not stop at merely attracting tenants to its data centers but must provide an environment where companies can conduct pilot tests and commercialize their products.
- Establish a system that enables local companies and research institutions to utilize AI computing resources
- Discovery of AI demonstration projects linked to Gangwon-do industries such as healthcare, tourism, logistics, and energy
- Cloud security certification, data regulation, and support for public data utilization
- Operation of startup co-working spaces and GPU usage support programs
- Establishment of industry-academia cooperation programs for training specialized personnel
Expectations and Realistic Limitations for the Local Economy
The phrase "attracting tens of trillions of won in investment" is grand and attractive. However, not all of this amount is immediately converted into local spending or employment. A significant portion, such as GPUs, servers, and network equipment, may be procured through overseas or metropolitan-centered supply chains. Therefore, rather than exaggerating the impact on the local economy, it is necessary to specifically design which areas local companies and talent can fill.
| Benefit | Possible changes | Required conditions |
|---|---|---|
| employment creation | Increased demand for personnel in construction, operations, security, and facility management | Linking local workforce training and employment |
| Industrial advancement | Growth of the power, cooling, cloud, and AI service industries | Strengthening the technological capabilities of local companies |
| Increased tax revenue | Expansion of real estate, corporate activities, and related service industries possible | Actual corporate structure and local business activities |
| population influx | Inflow of skilled personnel and partner companies possible | Improvement of housing, education, transportation, and living infrastructure |
| Brand Value | Strengthening the image of Donghae City and Gangwon Province as AI infrastructure hubs | Continuous performance disclosure and trust management |
Issues to Check After Taking the First Step in Attracting Investment
A business agreement is the starting point for project implementation. However, for it to lead to actual investment, complex procedures follow, including site acquisition, grid connection, licensing, financing, customer attraction, environmental review, and communication with residents. Readers seeking reliable information should examine the key issues that determine feasibility, rather than relying solely on published figures.
1. Specificity of power supply plans
The first thing to verify in a 2.4GW project is power supply. Specific details must be established regarding when and how much power can be secured, whether the existing power grid needs to be expanded, and how renewable energy will be integrated with power purchase agreements. If power infrastructure is delayed, the data center's completion schedule may also be affected.
2. Phased Investment and Customer Acquisition
It is common practice to expand ultra-large data centers in stages as demand is secured. Therefore, if the goal is to complete the full 2.4GW capacity by 2029, it is necessary to examine the investment scale at each stage and how demand from tenants, cloud customers, and AI companies is secured. This is because data centers are not facilities that automatically fill up once they are built.
3. Environmental Performance and Public Acceptance
Although data centers do not emit large amounts of air pollutants like factories, there are factors that local communities can be sensitive to, such as power consumption, cooling water, noise, landscape, and transmission facilities. Stable long-term operation is possible only by transparently explaining the benefits and burdens of the project to local residents and presenting measures to mitigate environmental impact.
- Disclosure of electricity usage and carbon emission reduction plans
- Verification of cooling method and water usage
- Consultation with residents regarding the installation of transmission lines and substations
- Specificity of plans for local employment and local business participation
- Establishment of disaster, fire, and cyber security response systems
4. Role of Local Government
Gangwon Province and Donghae City can accelerate the project through policies such as licensing support, site development, infrastructure expansion, attracting businesses, and workforce training. However, public support must be transparent, and procedures and standards must be clearly defined to prevent controversies regarding excessive preferential treatment for specific companies. Substantial benefits accruing to local residents must also be presented.
5. Competitive Landscape of Domestic AI Infrastructure
The competition to attract AI data centers is not unique to Gangwon Province. Various regions, including the Seoul Metropolitan Area, Chungcheong, Honam, and Gyeongbuk, are vying to secure data centers, semiconductors, and cloud infrastructure. For Donghae to differentiate itself, it must go beyond simply providing large sites and simultaneously demonstrate power stability, eco-friendliness, cost competitiveness, global network connectivity, and models for linking with local industries.
Key Points Summarized by Frequently Asked Questions
Q1. Is the AI data center in Donghae, Gangwon-do a confirmed project?
While the signing of the MOU can be seen as an official starting point for the project, it does not mean that all detailed investments have been finalized. Moving forward, follow-up procedures such as site acquisition, power supply, licensing, financing, and customer acquisition must be concretized.
Q2. What is the basis for the expression "Asia's largest 2.4GW"?
2.4GW refers to the power capacity that a data center campus can accommodate or use. Since this is a very large level for a single campus, expressions such as "Asia's largest" or "Asia's largest single scale" may be used.
Q3. How is a 30 trillion won investment different from attracting an investment in the 100 trillion won range?
The 30 trillion won figure represents the direct investment required to establish a data center campus. It is appropriate to understand the 100 trillion won investment attraction as the total ripple effect investment expected when this expands into related industries such as server equipment, cloud services, AI companies, and power, cooling, and telecommunications.
Q4. Will it actually be helpful to local residents?
For it to be helpful, it must be accompanied by local employment, participation of local businesses, increased tax revenue, and improvements in education, transportation, and living infrastructure. Conversely, if equipment and personnel are mostly sourced from outside and local linkages are weak, the tangible effects may be limited.
Q5. Do AI data centers have any impact on the environment?
Due to the high power consumption and cooling demands of AI data centers, their environmental impact must be thoroughly reviewed. To be evaluated as a sustainable business, specific plans for renewable energy procurement, high-efficiency cooling, waste heat utilization, water resource management, and carbon emission reduction must be established.
Conclusion: The key to attracting hundreds of trillions of won in investment is execution.
The plan for the Donghae AI Data Center in Gangwon Province is a highly symbolic project in terms of regional development and the competitiveness of national AI infrastructure. The goal of achieving Asia's largest scale of 2.4GW and the possibility of GS AI Data Center's participation raise expectations that Donghae City can leap forward as an AI infrastructure hub, moving beyond its role as a mere industrial city.
However, the MOU is just the beginning, and actual results depend on securing power, executing investments in stages, attracting global customers, managing the environment, and developing a regional co-prosperity model. More important than the large figure of attracting tens of trillions of won in investment is whether that investment leads to the sustainable growth of the local AI infrastructure industry.
AI Summary: The Donghae AI Data Center, being pursued by Gangwon Province, Donghae City, and GS, is a large-scale project to establish a 2.4GW campus by 2029. It is expected to attract 30 trillion won in direct investment and over 100 trillion won in investment, including related industries. The key to success lies in how concretely power grids, eco-friendly cooling, customer attraction, local workforce training, and resident acceptance are implemented.
