Inner Mongolia's "Artificial Intelligence+" initiative is now in full swing, with seven strategic innovation hubs transforming how the region's signature industries operate. From smart energy dispatch in cities to sector-specific large models for cashmere and steel, a localized push is underway to embed AI directly into the backbone of the local economy.
At the Smart Dispatch Center of Xuejiawan Power Supply Company in Ordos, a "cloud power plant" is aggregating 56 market participants, reaching a total scale of 196.08 megawatts. This virtual power plant integrates distributed photovoltaic systems, energy storage units, and charging infrastructure, focusing on four core scenarios: virtual power plant operations, electricity trading, energy-carbon management, and integrated energy services. The result is a highly intelligent and coordinated dispatch system. Li Yanchao, head of the marketing service department at Xuejiawan Power Supply Company, noted that the virtual power plant operation platform has been officially launched, with all self-operated charging stations now connected to the market and sourcing 100% green electricity. Between April and May, the average electricity purchase price dropped by 0.1189 yuan compared to the agency purchase price, saving a total of 3,538 yuan.
In Baotou, the No. 1 converter at Baotou Steel has adopted the region's first AI steelmaking project, the "Baotou Steel Smart Smelting" large model. The results are clear from the data: the double-hit rate for endpoint carbon and temperature control has surpassed 92%, smelting efficiency has increased by more than 6%, and the comprehensive cost per ton of steel has decreased by 1.5%.
Ordos is also home to the world's first multimodal large model for the cashmere industry, alongside an intelligent design agent for cashmere products. These tools target long-standing industry challenges such as homogenized pattern design, low precision in quality control, and unscientific raw material grading. Through AI-driven digital management across the entire process, the cashmere sector now has its own vertical large model to enhance precision and efficiency.
In Hulunbuir, a major project called the "Agricultural Production Multi-Agent Spatiotemporal Large Model" has been launched. Covering eight crops, including corn and soybeans, it oversees the entire cultivation cycle from plowing, seeding, and management to harvesting. From the black soil to the cloud, AI is actively participating in a new kind of digital farming revolution.
Xin Zhuoyu, an economics researcher at the Inner Mongolia Academy of Social Sciences, pointed out that rigid demand and real-world data are driving AI's integration into local industries, while supportive policies ensure these drivers remain sustainable. The "Artificial Intelligence+" action plan precisely targets key sectors with a combination of measures, including shared financial risk, open data access, platform support, and a mechanism that allows for trial and error. This fosters a positive cycle where the government provides the platform, businesses take the lead, and industries reap the benefits.
In Hohhot, the country's first green computing full-stack AI token trading platform has been launched, enabling industrial data to be efficiently converted into AI-readable token resources. This makes computing power as accessible and convenient as electricity or water. Furthermore, Inner Mongolia has introduced 18 measures to boost the token economy, promoting the transformation of vast production data from sectors like new energy, industrial manufacturing, and modern agriculture into high-quality tokens, thus feeding the training of vertical large models.
Ma Xiaoping, director of the Digital Economy Department at the Inner Mongolia University of Finance and Economics, believes the key to advancing the "AI+" initiative from single-point breakthroughs to widespread success lies in securing core elements such as computing power, tokens, business entities, talent, and the industrial ecosystem. Ma emphasized the need to first solidify the supply of computing power by leveraging a green computing infrastructure to build "token factories" and expand the supply of high-quality tokens. Next, it's crucial to activate data as a production factor, using the 18 measures to convert industrial data into usable token resources and establishing trading platforms to facilitate the flow of data value. Simultaneously, policies should support business development, innovation, talent recruitment, and the opening of application scenarios, thereby lowering the barrier for companies to engage in AI innovation.
In simple terms, the goal is to guarantee hardware computing power, produce tokens through production, foster innovation, and create application scenarios, thereby connecting the entire chain from research and development to industrial application and guiding AI technology into the real economy.