Data centers cluster in Ulanqab: Direct supply from the West Inner Mongolia power grid and Inner Mongolia's power surplus

[[**The real reason data centers cluster in Ulanqab is that the Western Inner Mongolia independent grid allows direct power supply from power plants, bypassing transmission and distribution tariffs, bringing the electricity price down to 0.358 yuan/kWh; combined with Inner Mongolia's severe power surplus and its rigid need for consumption, this creates an institutional cost advantage that other regions cannot replicate.**

**Key points**

**1. The low electricity price stems from the Western Inner Mongolia grid's "direct supply privilege," not natural resources**

The Western Inner Mongolia grid is the country's only provincial-level independent grid, allowing data centers to receive dedicated direct supply from power plants, exempting transmission and distribution tariffs, so that the delivered price is only 0.358 yuan/kWh, 0.442 yuan/kWh lower than Beijing; a single AI accelerator card saves about 2,700 yuan a year, and a cluster of ten thousand cards saves 27 million yuan a year.

**2. Inner Mongolia's severe power surplus is the internal driving force behind local construction**

As of the end of 2025, Inner Mongolia's wind power installed capacity exceeds 100GW and coal power is about 130GW, which local industrial load cannot absorb. Data centers have become a "power consumption tool," converting surplus electricity into GDP, tax revenue and jobs, driving 89 data centers forward simultaneously, with completion possible in 12 to 18 months.

**3. Ulanqab's low electricity price is structural and cannot be replicated**

Other computing hubs (Gansu, Ningxia, etc.) are constrained by the State Grid's transmission and distribution system and cannot eliminate the "wheeling fee"; the independence of the Western Inner Mongolia grid is a historical legacy of the power reform at the end of the last century, and no region can replicate this institutional cost advantage.

**4. Water resource constraints are a potential challenge, but in the short term do not affect the landscape**

Ulanqab's per capita water resources are less than a quarter of the national average, and large data centers consume millions of liters of water per day in summer; liquid cooling technology has entered projects to cope with this. But water constraints have not yet changed the cost advantage dominated by the power system, and in the short term will not reverse the trend of industrial clustering. 2026-10-09 07:16 Why do data centers cluster in Ulanqab? Claire Market Watch Quick Take

This article is from the WeChat public account: Claire Market Watch, author: Claire Market Watch

Why do data centers cluster in Ulanqab? The answer is not wind and solar, but the institutional flexibility of the Western Inner Mongolia grid.

A 700W AI accelerator card, running nonstop for a year, consumes about 6,100 kWh of electricity.

In Beijing, the industrial electricity price is about 0.80 yuan per kWh; in Ulanqab, Inner Mongolia, the actual delivered electricity price for data centers is between 0.32 and 0.36 yuan, and this article uses 0.358 yuan as a conservative estimate.

The difference is 0.442 yuan. Converted, a single card saves about 2,700 yuan a year; a cluster of ten thousand cards saves 27 million yuan a year—and that is just the electricity bill.

So over the past three years, Alibaba, Huawei, ByteDance, DeepSeek, Z.ai, VNET, ZData... a long list of names has poured into this small city in Inner Mongolia known for potatoes and yogurt. Ulanqab has now built or planned 89 data centers, with a confirmed computing capacity of 15GW, and is called by the industry "Asia's fastest-growing data center hub."

Almost every report gives the same explanation: because there is wind, coal, a cold climate, and cheap land here.

But is that explanation correct?

I. Why did Ulanqab emerge, and not Gansu or Ningxia?

Gansu has wind power, Ningxia has wind and solar, Qinghai has photovoltaics—if "wind, coal, and cold" were enough, why is Ulanqab running out in front, rather than Gansu or Ningxia, which have equally outstanding resource endowments?

The answer lies in a detail rarely mentioned in computing coverage: western Inner Mongolia is covered by the independently operated Western Inner Mongolia Power Grid, unlike most of the country, which is supplied by the State Grid.

By supplying power directly to data center facilities, developers can sharply cut costs and are exempt from transmission fees paid to the State Grid.

This is the real source of the 0.358 yuan electricity price—not that wind and solar are cheap in themselves, but that a unique electricity institutional arrangement lets developers bypass transmission and distribution tariffs.

II. Institutional variable one: the Western Inner Mongolia Power Grid's "direct supply privilege"

The Western Inner Mongolia Power Grid is China's only provincial-level independent grid besides the State Grid and China Southern Power Grid (Xinjiang and Tibet have partial independent grids, but their scale cannot compare). This is a historical legacy of the plant-grid separation reform at the end of the last century.

Its boost to building data centers is this: local officials allow some data center projects to establish dedicated power generation facilities directly connected to computing clusters, thereby avoiding the grid transmission restrictions that developers in other regions commonly face.

What does this mean?

In most provinces, for a data center to use electricity, it must buy from the State Grid, and the price includes generation cost, transmission and distribution tariffs, and government funds. In western Inner Mongolia, data centers can negotiate dedicated-line power supply directly with nearby power plants, and what is saved is precisely the transmission and distribution tariff.

It is not that electricity itself is cheap; it is that the "wheeling fee" is institutionally waived.

This also explains why Huawei Rotating Chairman Xu Zhijun, after inspecting many candidate sites, ultimately chose Ulanqab—for two reasons: low electricity prices, and the convenience of a two-hour high-speed rail trip to Beijing. What really drives the low electricity price is not "Inner Mongolia has lots of power," but "the Western Inner Mongolia grid can supply directly."

III. Institutional variable two: Inner Mongolia does not lack power; what it lacks is "people who use power"

If the Western Inner Mongolia Power Grid solves the "electricity price structure" problem, then the supply-demand relationship of Inner Mongolia's power itself solves the "why local governments are willing to do this" problem.

Lantau Group energy analyst David Fishman points out that Inner Mongolia may be the administrative region with the most pronounced "local electricity oversupply" in the world.

The numbers are straightforward: by the end of 2025, Inner Mongolia's installed wind power capacity exceeded 100GW, the highest in the country and several times the UK's total wind capacity; at the same time, it also has about 130GW of fossil fuel generation capacity, mainly coal power.

But with such enormous power generation capacity, local industrial load simply cannot absorb it. A large amount of this power capacity here is operating at low load.

In other words, Inner Mongolia does not lack electricity; what it lacks is "people who use electricity."

Against this backdrop, what does the arrival of data centers mean for the local government?

It is not as simple as attracting a technology industrial park. It is converting power generation capacity into GDP, tax revenue, and jobs. A data center is essentially a "power absorption tool"—computing power is merely the form of absorption; revitalizing power assets is the real imperative.

This is the key to understanding why Ulanqab's construction speed far exceeds market expectations. Goldman Sachs data shows that local data center construction costs are about 20% lower than in first-tier major cities, and many projects can be completed in 12 to 18 months, whereas comparable projects in the United States usually take 18 to 24 months. Thousands of workers have entered the site, with 89 data centers advancing simultaneously—this pace of execution is not merely about "attracting technology companies," but even more about "revitalizing local power assets."

IV. First principles: Ulanqab is not a technology park, but an extension of energy infrastructure

Putting these two institutional variables together, Ulanqab's first principle becomes clear:

It is not a technology park in the policy story of "channeling computing resources from the east to the west." It is an outlet for Inner Mongolia's surplus power, plus the institutional arrangement under which the independent Western Inner Mongolia grid allows direct supply—an extension of energy infrastructure created by the combination of the two.

SemiAnalysis calls China's prefabricated modular data centers "Lego data centres," and in a July 2026 report tracked more than 61GW of modular data center capacity worldwide. These standardized container-like units have built-in server racks and can be assembled extremely quickly, and are now gradually being adopted by large US cloud service providers. But more noteworthy than the "Lego" aspect is: why can China assemble these Legos to a scale of 15GW in such a short time?

The answer is not engineering capability, but the local driving force behind it: "the power has to have somewhere to go."

According to a SemiAnalysis report, ByteDance accounts for about one-fifth of China's delivered data center capacity, making it the largest data center tenant in China, and the vast majority of that capacity comes from leasing rather than self-building; DeepSeek and Z.ai have also set up dedicated facilities in Ulanqab. These demand-side players chose Ulanqab not because "there's a tech vibe here," but because "the electricity here is genuinely cheap, and that cheapness is institutional and sustainable."

V. Why can't others copy it?

A natural question is: if direct supply from the Western Inner Mongolia Power Grid is so good, why don't other computing hubs copy it?

The answer is that they can't.

The independence of the Western Inner Mongolia Power Grid is a historical result of the power system reform at the end of the last century. Even if Gansu, Ningxia, and Qinghai have cheaper wind and solar resources, they must still go through State Grid's transmission and distribution tariffs, and the "wheeling fee" in between cannot be avoided.

This means that Ulanqab's low electricity price is not a temporary investment incentive, but a structural cost advantage—it does not depend on the investment efforts of any particular government, but on a power system that others cannot replicate.

Of course, this model also has its scope of application.

A local data center executive also pointed out that Inner Mongolia is relatively short of water resources, and as the industry scale continues to expand, this may become a challenge that needs to be addressed in the future. Specifically for Ulanqab, the city's per capita water resources are only 484 cubic meters, less than a quarter of the national average; a large data center using evaporative cooling can consume millions of liters of water per day in summer. Vertiv's liquid cooling equipment has already entered the projects of ZData and China International Capital Corporation (CICC) Data—in a semi-arid region, liquid cooling here is more of an engineering response to water resource constraints than a purely environmental upgrade.

But in the short term, the water constraint has not yet reached the point of changing this landscape. What truly determines Ulanqab's long-term value is still that power system that others cannot copy.

VI. Conclusion

When the outside world is still debating whether China's computing power is in surplus or shortage, and whether chips can fill all the racks, if you return to Ulanqab itself, you will find a more plain fact:

The core asset here is not servers, not racks, and not even that 15GW computing power plan.

It is the power system that makes the 0.358 yuan electricity price possible—the independence of the Western Inner Mongolia Power Grid, plus the rigid demand in Inner Mongolia for an outlet to absorb surplus electricity.

Chips will be affected by the external environment, and racks can be built in a year, but a power system formed by history and impossible for others to replicate is the city's true moat.

Only by understanding this can one understand why Ulanqab's expansion speed far exceeds market expectations; and only then can one predict who will remain when the industry shifts from "competing on scale" to "competing on efficiency." Channel: Business & Consumer. This content is published with the author's authorization. The views represent only the author and do not represent Huxiu's position. If you have any objections or complaints about this article, please contact tougao@huxiu.com. Everyone who is changing and wants to change the world is on the Huxiu APP[ ]]]