Across the rolling grasslands of Inner Mongolia, wind turbines and seas of solar panels now share the landscape with potato farms and sheep herders. Then the data centres come into view: enormous structures, stretching row after row, with cranes moving overhead and sparks flying from welding work. On a recent visit, BBC reporter Laura Bicker was told that some of this land had been farmland only a few months earlier.
Workers described successive construction phases, with another phase each year and more still to come. Bicker described a $75 billion investment and buildings whose scale she compared to two airports put together. It is what people there call “China speed”.
That buildout is the physical backdrop to an increasingly prominent issue between Washington and Beijing. As President Xi Jinping prepares to meet President Donald Trump at the White House, artificial intelligence has pushed its way onto an agenda already crowded with disputes over trade and Taiwan. Recent warnings about AI safety and calls from some technology leaders to slow development have not changed Trump’s emphasis on competition: he says the United States is ahead of China and intends to keep it that way.
Inner Mongolia offers two things China wants for its data centres: vast open space and a way to power them. The turbines and solar farms are part of a renewable-energy expansion that supports the country’s ambition to have at least 80% of its data centres powered by renewables. Yet the speed and scale of construction say little on their own about what the transformation means for the people living beside it.
Bicker found construction workers pleased to have work, and residents who saw the development as evidence of China’s growing strength. One called the country an “infrastructure maniac”. Others were less convinced. “They need educated people. They don’t need that many workers,” one local said, explaining that the work he could offer was hard manual labour.
Off camera, a farmer pulled the reporting team aside and dismissed much of the enthusiasm as propaganda. In his view, the centres offered no local benefit or jobs, while taking farmers’ land without proper compensation. Another resident worried about the possibility of a village being demolished for a data centre and said he preferred living beside farmland. These were not the only views: some farmers said their lives had barely changed, while others credited the projects with creating jobs and raising living standards.
The divisions were sharper than Bicker had expected. She noted that the Stanford AI Index suggested more favourable attitudes towards AI in China, while cautioning that the transparency of surveys there can be difficult to assess. In the United States, host Asma Khalid pointed to widespread local opposition to data-centre construction, even as Trump promotes it as an economic opportunity. In both countries, leaders’ enthusiasm does not necessarily settle the argument in the communities being asked to accommodate the infrastructure.
One encounter captured the distance between China’s technological ambitions and everyday rural life. Bicker met a farmer earning roughly 3,000 yuan a month, about $450, who kept sheep and cows and drove a tractor, with his dog Huahua running alongside. Asked whether AI mattered to him, he replied: “What would I use it for?” His phone was covered in cracks. From another pocket, he produced the small device he needed to connect to Wi-Fi.
Alongside the data centres, China is trying to build familiarity with AI from an early age. Bicker described coding and AI learning entering the curriculum from kindergarten, with robotics also incorporated into schooling. At one family’s home, she met a boy called Timmy who was so absorbed in playing chess with a robot that he barely had time to speak to her. The machine talked to him as he moved the pieces, helped reset the game and congratulated him on playing well. He proudly explained that he had moved up a level.
His mother said the robot had become one of his friends. “AI is the inevitable trend. We need to coexist with it,” she said, arguing that children should encounter the technology as early as possible. She also described the strong emotional attachment her son was developing through playing with it.
For adults, AI study centres offer another route into learning about the technology. Bicker also spoke to university students studying AI, biotechnology and ways to bring artificial intelligence into the community. The ambition extends beyond training people: China wants to keep its homegrown talent rather than see its highest-achieving students leave for American universities.
Several forces are helping to keep them at home. Students told Bicker they were struggling to secure US visas. At the same time, government support for AI start-ups offers an incentive to build a career in China. This combination of education, support and infrastructure is intended to strengthen the country’s future position, even where it does not yet hold the technological lead. State media reinforce the message by celebrating advances such as those made by DeepSeek.
For Xi, AI can enhance China’s international stature and provide leverage across the negotiating table from the United States. But Bicker’s account also describes two Chinas: one pushing towards the technological frontier, the other still facing much more immediate needs. With the economy performing less well than Beijing would like, employment and local spending remain pressing concerns. The data centres are rising quickly. Their ability to answer those concerns is much less clear.