Late last year, President Trump announced that Washington would loosen controls on chip exports to China, allowing Nvidia to sell its H200 AI chip to authorized Chinese companies. The president’s decision raised concerns among lawmakers that he was sacrificing US security and the country’s edge in AI for a trade agreement with China’s leader, Xi Jinping. But any concerns that Chinese companies would buy Nvidia’s prized chips were soon quelled by Beijing’s ban on imports of the H200. Officials have recently allowed other companies to use these chips, but have struggled to limit their distribution. Beijing would prefer Chinese companies to use domestic chips, even if they are not good.
To expand the country’s global power and protect its economy from external threats, China’s leaders have worked to isolate its technology sector from the rest of the world. They want Chinese companies to sell Chinese technology designed in Chinese laboratories and built from Chinese supply chains. The strategy seems to be working, thanks to the many Chinese electric cars on the streets of Europe and AI designs from Chinese companies like DeepSeek and Silicon Valley’s fearsome Z.ai. Yet China’s growing isolation threatens to limit innovation at home by crowding out the world’s best technology. In global technology industries, such as AI, China’s defensive posture may hold the country back.
China’s rise and leadership in various sectors can actually be attributed to its previous openness and foreign ingenuity. In the 1980s, when the country began to modernize its economy by engaging in international trade, China’s outdated industries absorbed foreign technology. Beijing’s policymakers introduced many policies to help promote and protect new industries, which helped China become a leader in industries such as telecommunications and high-speed rail.
The rise of digital media threatened the value of transparency, prompting Beijing to erect the Great Firewall to protect citizens from the flow of unwanted ideas and information. International networks and services such as Facebook, Google, and X were shut down, and Chinese alternatives emerged. Companies that followed Beijing’s strict regulatory rules, such as search engine provider Baidu and WeChat operator Tencent, rose to prominence; they had 1.1 billion customers largely to themselves. But as latecomers to the Internet revolution, targeted locally and controlled by Beijing’s media censors, these companies have never really competed with Meta and Alphabet in global appeal. With the best exception, TikTok, founded by Chinese tech company ByteDance, cannot operate in China.
Anyone traveling to China will soon discover the everyday wonders of this digital landscape. The Chinese send messages on WeChat, not WhatsApp or Signal. Instead of scrolling Instagram or TikTok, locals post on Weibo, RedNote, and Douyin.
Xi stepped up China’s inward turn in technology nearly a decade ago, as geopolitical tensions flared between the United States and China. Sensing that China’s reliance on foreign technology is a national security liability, and that developing this technology is essential for the country’s economic growth, he took immediate action to expand domestic research and industry and encourage greater self-sufficiency in semiconductor production and supply. Xi’s government has spent hundreds of billions of dollars to support domestic chips, electric cars, AI, and other emerging areas to compete with the West’s most advanced industries.
In some sectors, Xi’s plan is succeeding. China has built an almost entirely Chinese EV supply chain, including the batteries that run it and the software that runs it. But no single country, not even China, can create the best of everything. China’s efforts to build a complete supply chain for Chinese computer chips, for example—spurred by President Biden’s 2022 export controls on high-end American AI chips and chip-making equipment to China—have proven difficult. The process of manufacturing semiconductors is highly complex and dauntingly global, involving supply chains that span many countries. Rebuilding the chain in one place is “absolutely impossible,” Scott Kennedy, a senior consultant at the Center for Strategic and International Studies who has studied the Chinese technology industry, told me. Even trying it “is expensive and self-defeating.”
China will continue to pursue its own chip industry however. Yet this emphasis on self-sufficiency has certainly come at a cost. China is using huge resources to create its own versions of what the rest of the world already has. And while China’s chip industry is making strides, it’s still playing catch-up with the West. Chinese AI companies have had to operate without the best chips, and this disadvantage will persist if Beijing insists on using Chinese alternatives. Sharing technology across borders “is how a lot of innovation happens,” especially in the semiconductor industry, Paul Triolo, a Chinese technology specialist at the consulting firm Albright Stonebridge Group, told me. Since there are now “two fragmented piles” – in China and the United States – “some innovation will not happen,” he added.
Like their internet platform predecessors, Chinese AI companies are emerging in a highly regulated and protected domestic market. The mandate that Chinese companies and customers use AI models developed by DeepSeek, Alibaba, Moonshot, and other Chinese companies, not OpenAI’s ChatGPT or Anthropic’s Claude, means that domestic users are being denied the world’s most advanced AI while the technology affects economies and societies around the world. In April, Beijing regulators took the unusual step of ordering Meta to rescind its $2 billion purchase of Singaporean AI firm Manus, a Chinese startup known for creating deeply researched reports, on national security grounds. The message to Chinese tech entrepreneurs was clear: Partnering with foreigners is unacceptable.
Beijing is threatening to push back on Chinese technology in other ways as well. Chinese AI designs are becoming increasingly popular around the world, in part because they are open source, which means that users anywhere can adopt and modify them at no cost. Yet Chinese managers, wary of foreign competitors using Chinese expertise, are is reported to discuss new restrictions on foreign access to the most advanced models of Chinese AI, which promises to prevent their increase and appeal.
In a sign that Chinese companies realize they are still lagging behind the progress of their American rivals, American AI firms are accusing their Chinese competitors of “downgrading,” or harvesting advanced American models to train them. In a letter sent to two US senators last month and reviewed by AtlanticAnthropic alleged that Chinese e-commerce company Alibaba, which operates a model called Qwen, “performed the largest spraying attack” on Claude, using approximately 25,000 fraudulent accounts to illegally obtain some of its “most valuable assets.” (A spokeswoman for Alibaba declined to comment.) Anthropic has responded to these attacks by beefing up Claude’s defenses and hunting down Chinese accounts that find the model inappropriate, reinforcing China’s technological isolation and potentially jeopardizing China’s efforts to keep pace with resourceful US rivals.
Perhaps China’s technology industry will move further in its direction, creating products that are somehow different but similar to those in the United States. Triolo believes that the world of AI could be divided into two parts, with China and the United States at the center of the two worlds with little connection between them.
What seems fair is that China, in building an impenetrable wall around its technology sector and limiting the country’s access to better technology from elsewhere, is placing itself in a distant second place.
