Summary of the Analysis
Ulanqab, a small city in Inner Mongolia with a GDP of only 126 billion yuan, has seized the opportunity presented by AI computing power to attract investments totaling over 500 billion yuan from giants such as Huawei, Apple, and DeepSeek. Its current computing power capacity accounts for 8.7% of the country's total. The local authorities have even set the goal of becoming the “Token Capital.” The city’s core competitiveness does not lie in the commonly mentioned advantages of its cool climate and proximity to Beijing, but rather in a unique historical opportunity stemming from the Mongolian-Western power grid. Ulanqab enjoys the lowest industrial electricity price in the country—just 0.3 yuan per kilowatt-hour. However, it faces the same challenges as other resource-dependent cities: most of the profits from these investments do not remain locally. To avoid becoming merely a supplier of computing power, the city must expand into downstream AI applications.
---
Simplified Explanation of Key Points
1. The real reason big companies are flocking to Ulanqab: it’s because the electricity is cheaper than bottled water
Many think that Ulanqab’s low annual average temperature (4.3°C) and its proximity to Beijing (300 kilometers) are the main attractions for these companies. While these factors are certainly beneficial, they are not decisive. For example, Harbin in the northeast is even colder, and regions like Gansu and Qinghai have more scenic resources. So why don’t big companies settle there? The real key is its unique power grid situation: 20 years ago, the Mongolian-Western power grid, due to its small scale and high debt, was separated from the national grid and became an independent system. With the surge in Inner Mongolia’s power generation capacity, there was a surplus of electricity that couldn’t be sold externally, leading to its sale at a loss. The electricity price is half that of Beijing’s industrial electricity, which means a computing center hosting 10,000 servers can save 100 million yuan in annual electricity costs. Additionally, Ulanqab’s green energy usage exceeds 86%, meeting China’s policy requirement for 80% of smart computing centers to use green energy, giving it a significant competitive advantage.
2. The huge investment sounds impressive, but locals benefit little
The 500 billion yuan in investment will not significantly enrich the entire population of Ulanqab. Computing infrastructure is not a highly inclusive industry. While building a housing project can support multiple related industries, a large computing center employs at most 500 people, mostly technical personnel from other regions. Nearly half of the investment goes towards purchasing high-end components from overseas companies like NVIDIA and TSMC, with the remaining 20% used for server assembly. Only a small portion (about 750 billion yuan) stays locally in the form of electricity costs, rent, and construction fees, resulting in minimal benefits for the local community.
3. Ulanqab is following the same path as other resource-dependent cities
Ulanqab’s development is similar to that of cities like Ordos and Karamay. In the past, these cities relied on local coal for income; now, they convert cheap electricity into computing power for sale. Although Ulanqab’s GDP is high, the majority of profits go to foreign and southern companies. The local economy benefits only from electricity revenue and rent.
4. A practical solution to avoid being just a supplier of computing power
Ulanqab is focusing on downstream AI applications, where its low-cost computing power has a natural advantage. For example, AI-based content and animation production can be cost-effective here. This approach differs from the traditional model, where resources are exploited for immediate profit.
5. A warning to other northern cities: the computing power trend is not for everyone
Many northern cities are trying to follow Ulanqab’s example, but most cannot replicate its unique advantages. They need either a low electricity price or sufficient renewable energy sources. Simply claiming a cool location is not enough to attract big companies. Even if they succeed in attracting computing centers, they must plan how to develop related industries and ensure that the profits stay locally.
In summary, Ulanqab’s success lies in its access to cheap electricity and its focus on practical, downstream AI applications. Other cities must find their own unique advantages to avoid repeating the same mistakes and truly benefit from the computing power trend.