During an inspection tour of Ulaanqab, Inner Mongolia autonomous region, in October 2026, President Xi Jinping emphasized that the computing network is one of "six networks" planned and constructed by the country, and that it is in a stage of accelerated development with broad prospects.
China is accelerating a sweeping infrastructure plan centered around "six networks" — water, power, computing, telecommunications, urban pipelines and logistics — in order to stabilize investment, create jobs and lay the groundwork for the next phase of the country's technological and industrial transformation.
Highlighting the importance of "computing-electricity synergy", a critical pathway to address the explosive growth in artificial intelligence computing demand and the challenges in global energy transition, and "computing-water synergy", Xi encouraged local officials to ensure smooth construction and operation of computing center projects and play an active role in the development of green computing, thereby contributing more to the building of a digital China and promoting the development of new quality productive forces.
This directive not only outlines expectations for the industry but also identifies crucial challenges for the education sector: where resources gather, the supply of knowledge and talent must adapt accordingly. Otherwise, the computing power industry risks becoming a "computing enclave" — consuming local green electricity and water while leaving high-value-added jobs elsewhere.
This raises a critical question: when a region's allocation of resources changes due to the development of computing power, can local universities effectively turn them into educational advantages? What mechanisms are needed for such a transformation, and where are the potential gaps?
Why is computing power 'shifting westward'?
Before linking computing power to education, it is essential to address a more fundamental question. The "east data, west computing" strategy is a mega project that aims to link the computing capabilities of the country's eastern regions with the inland western regions, as well as develop green computing in Inner Mongolia.
China has initiated the construction of eight national computing power hubs, with Inner Mongolia's Horinger data center cluster covering the entire Hohhot-Ulaanqab region. The Horinger New Area has attracted over 60 large-scale computing power center projects, including those from China's three major telecom operators alongside financial institutions and leading internet enterprises. The total computing power capacity has exceeded 125,000 petaflops.
In Ulaanqab, 84 data center projects have been signed, with a total investment exceeding 500 billion yuan ($75.04 billion) and a computing power capacity reaching 132,000 petaflops. Leading companies such as Huawei, Alibaba, Apple and Kuaishou have set up operations there.
Thus, the cost of computing power depends not only on chips but also on where electricity is sourced from, how heat is dissipated and where water is obtained. And the demand is largely concentrated in the eastern region, which also has the tightest energy consumption quotas, leading to a division of labor: latency-sensitive computing tasks are handled locally in the east, while latency-insensitive tasks — such as model training, offline analysis, data storage and batch processing — have shifted to the western region. The westward movement of computing power represents more than industrial relocation; it is a reallocation of national resources. And every such reallocation is eventually reflected in the distribution of knowledge and talent.
Inner Mongolia's advantages and the overlooked talent equation
Inner Mongolia provides some key comparative advantages for hosting computing power hubs. The total installed capacity of new energy across the region has reached 174 million kilowatts, with Ulaanqab's new energy installed capacity amounting to 20.28 million kilowatts. Green electricity accounts for 67 percent of the total in Ulaanqab, ranking among the top in the country.
With an average annual temperature of 4.3 degrees Celsius, thanks to natural cooling sources available for nearly 10 months in a year, Ulaanqab can reduce energy consumption by 20-30 percent.
Inner Mongolia has thus become a low-latency computing power service circle, which allows a computing system or network to provide responses with minimal delay. This circle ensures network latency of 2 milliseconds for Hohhot-Baotou-Ordos-Ulaanqab, 5 milliseconds for the Beijing-Tianjin-Hebei region and 20 milliseconds for the Guangdong-Hong Kong-Macao Greater Bay Area.
Leveraging its advantages in green electricity, climate and geographical location, Ulaanqab has made significant progress in developing the green computing power industry. A total of 109 computing power center projects have been signed, and by August 2026, the computing power capacity reached 190,000 petaflops, with intelligent computing accounting for over 90 percent of the total.
However, advantages do not automatically translate into long-term benefits.
At present, Inner Mongolia primarily provides midstream capabilities, such as computer cabinets, electricity and computing power, to the supply chain. Talent shortages are the most easily overlooked challenge. Jobs such as those in data center operations, computing-power scheduling and industry-specific deployment require an in-location presence; whereas, the relatively limited academic foundation in frontier and ethnic minority regions often creates gaps between local universities' expertise and industry demand.
Without systematic efforts to bridge this gap, the larger the industry grows, the larger the disparity becomes, reinforcing external dependencies. Thus, empowering education through computing power is a vital factor in securing Inner Mongolia's position in this industry chain.
How does computing power transform education?
In April 2026, China issued a plan to make artificial intelligence a public foundational course in higher education. The plan explicitly states that provincial education administrative departments should be encouraged to leverage the national integrated computing power network to provide computing power support for AI applications. In June 2026, the central government issued a 15th Five-Year Plan for education development, listing the advancement of digital and intelligent transformation in education as one of the comprehensive reform tasks. Education will be transformed in three aspects.
First, high-quality teachers, premium courses and experimental facilities have traditionally been scarce and costly to replicate. Computing power networks consolidate such resources into callable services, drastically reducing replication costs. For sparsely populated and widely dispersed regions such as Inner Mongolia, this is the first time high-quality resources can be delivered cost-effectively.
Second, research has shifted from hypothesis-driven methodologies to data-driven approaches, transitioning from sample-based experimentation to large-scale simulations. Teaching and learning are now measurable and feedback-enabled processes. As a result, teachers are transitioning from being sole conveyors of knowledge to playing roles in value guidance and fostering innovation.
Third, because of a disconnect between local universities and industries, talent cultivation does not match market demand. Computing power industries require rapidly evolving, on-site technical roles that cannot rely on long-distance supply chains. Local educational institutions thus become indispensable here.
In essence, translating rising computing power into educational advantages involves three steps: resource accessibility, scenario accessibility and capability accessibility. Access to equipment does not ensure access to resources, having resources does not guarantee classroom integration, and having classrooms does not necessarily translate into student competency. A failure in any one of these steps will reduce computing power to a mere set of devices housed within machine rooms.
How does the transformation occur?
The realization of this transformation depends on one fundamental action: translating real-world industry challenges into actionable academic objectives. For example, Jining Normal University transforms local technical demand — such as computing-power-electricity synergy, source-grid-load-storage collaborative optimization, liquid cooling energy efficiency and green power utilization — into research topics and subsequently breaks them into coursework modules that students can engage with.
The success of such initiatives should be measured not by the number of signed partnerships but by how many students complete a full "problem-solving loop" — from formulating questions to designing and testing solutions.
Broader regional results are emerging as well. For instance, Ulanqab Vocational College established industry colleges with a company. Their courses are reverse-engineered according to job standards instead of academic disciplines, training over 650 specialized technicians for data center operations.
Across Ulaanqab city, the computing power industry has created 5,497 jobs, with 3,205 positions filled by local workers. Such numbers arguably better reflect the industry's local impact than revenue figures.
However, this transformative approach has its limitations. Four potential biases warrant caution.
First, purchasing equipment is easier, visible and measurable, but servers are not the same as talent, and platforms are not equivalent to curricula. The true metric should be whether the technology is genuinely utilized.
Second, AI as a discipline has historically attracted substantial investments, leading to redundancy as similar schools crowd the same niche, risking structural oversaturation.
Third, if local education only supplies technicians for operational maintenance without progressing to upper-level roles such as computational optimization or cluster scheduling, it risks becoming the least valuable part of the chain. Moving up requires leveraging knowledge rather than equipment.
Fourth, traditional evaluations based solely on the number of papers, projects or individual accolades undermine incentives for educators to commit long-term to industry-academic integration, which has slower yields and less conventional outcomes.
An underlying challenge remains: after intelligent tools enter classrooms, how can thought inertia be avoided? How should student privacy and academic integrity be safeguarded? Technology is neutral — it can empower or displace, depending on how it is utilized.
It has been said that the most essential element of computing power is electricity and, by extension, renewable energy. By this reasoning, the ultimate goal of education must be human development.
The transformation of local universities will not occur overnight. It requires steady direction, focused resources and enduring patience, just as is needed for achieving "resource accessibility, scenario accessibility and capability accessibility". The first two steps can be addressed through investment, but the third step can only be achieved over time.
With the 15th Five-Year Plan (2026-30) already underway and the 80th anniversary of the establishment of the Inner Mongolia autonomous region in 2027, local universities must seize the opportunity to translate regional resource advantages into academic strengths, research achievements and educational benefits.
By embedding computing power into their ecosystems and ensuring that local students grow to become problem solvers, the region can ensure that computing power does not simply operate here, it thrives here as a catalyst for sustainable growth and development.
The author is Party secretary at Jining Normal University.
The views don't necessarily reflect those of China Daily.
If you have a specific expertise, or would like to share your thought about our stories, then send us your writings at opinion@chinadaily.com.cn, and comment@chinadaily.com.cn.
