While speaking at the India AI Impact Summit 2026 in New Delhi last month, White House science and technology adviser Michael Kratsios announced the launch of Tech Corps. Positioned as “bringing America’s historic Peace Corps into the 21st century,” the program promises to offer thousands of American volunteers and advisers with technical and scientific backgrounds the opportunity to help partner nations adopt American AI technology.
Kratsios suggested the initiative will help to provide “last-mile support in deploying powerful AI applications for enhanced public services” in sectors including health, education, and agriculture.
But the Tech Corps also serves another purpose: steering countries toward American AI hardware and software, and away from the Chinese equivalents as the Trump administration sizes up an ongoing U.S.-China AI race. Given the administration’s belief that AI will “define the future of economic growth, national security, and global competitiveness for decades to come,” the deployment of U.S. AI technology may be vital to support the promotion of American geopolitical, economic, and security interests.
“The United States has recognized that it has to compete on the unit economics of AI,” American Enterprise Fellow Ryan Fedasiuk told The Dispatch. “American AI can’t just be the best in the world in terms of quality, it has to be the most accessible.”
In his speech at the summit in India, Kratsios also mentioned financing initiatives to help support partners building on the U.S. tech stack (in this case the components needed to run AI systems), as well as a “National Champion Initiative” that would work to support companies in partner nations with American technology. (It’s unclear which countries will participate, but India will likely be among them, given it recently joined the Pax Silica—a U.S.-led international initiative aiming to secure trusted supply chains for advanced technologies.)
The introduction of the Tech Corps is the latest of a number of measures and proposals under Trump’s administration aimed at diffusing the full American AI stack abroad. In July, the administration released its flagship AI strategy document, “Winning the AI Race: America’s AI Action Plan,” which argued that to succeed in global AI competition, the U.S. must “drive adoption of American AI systems, computing hardware, and standards throughout the world.” Two weeks later, Trump signed an executive order to coordinate a national effort for “promoting the export of full-stack American AI technology packages,” which was formalized in October under the Commerce Department’s “American AI Exports Program.”
The Tech Corps can thereby be seen as the latest push to support the promotion of the American AI stack abroad.
Where China sits relative to the U.S. when it comes to deploying its own technology stacks around the world varies. While the American AI stack remains the default infrastructure for most of the world’s AI development, Chinese open source models, and to a certain extent cloud service providers, are “increasingly competitive particularly in the Global South, due to a combination of cheaper pricing and [architectural] efficiency gains” according to Center for A New American Security research associate Ruby Scanlon.
When it comes to exporting AI chips, data centers, and deep learning frameworks (the software toolkits that developers use to build and train AI models), America remains far ahead of China. Yet following the release of DeepSeek R1 in January 2025, a wave of Chinese open-weight models began to appear on the market—models in which the underlying code and trained parameters (the “weights”) are made publicly available for anyone to download, modify, and deploy. Dominant American models such as ChatGPT, Claude, and Gemini, are closed models, functioning more like private property, made available for public use through licensing agreements. While better performing, they are also generally more expensive, making cheaper Chinese open-source models more alluring.
“Chinese models are very strong, especially on benchmarks, but still behind the best U.S. models. They are likely serving the models in ways that make it cheaper and more accessible at a lower price point, because if they are competing at the same price point they’d get killed,” Allen Institute for AI researcher Nathan Lambert told The Dispatch.
Chinese models are reportedly being used in Africa, Southeast Asia, and the Middle East, yet to what extent this is the case, and how it compares to the diffusion of U.S. AI models remain unclear.
Yet analysts warn of the importance of prioritising the U.S. AI stack, fearing Chinese infrastructural lock-in, and security risks.
“We are at risk of repeating the experience of Huawei’s dominance in 5G equipment, in AI models today in 2026,” Fedasiuk said. His latest report discusses potential Chinese security risks, data exfiltration, intelligence collection and censorship, when relying on the Chinese AI stack.
The U.S. closed infrastructure is certainly strong. According to a recent RAND report, the U.S. accounted for approximately 93 percent of global large language model site visits, and as of the second quarter of 2025, 90 percent of Fortune 500 companies reportedly utilized OpenAI technology.
Yet Chinese open-weight models such as Qwen, Kimi, and Minimax remain popular. According to a recent MIT study, such models have surpassed U.S. open-weight models in total downloads, and on Hugging Face—the main platform where AI researchers and developers share, download, and build on each other’s models—Alibaba’s Qwen model has reportedly been downloaded more than 700 million times, making it the most downloaded model on the site. A reporter assessing the statistics from the developer platform OpenRouter found that Chinese open-weight models grew from just 1.2 percent of global AI usage in late 2024 to highs of nearly 30 percent by late 2025.
There are caveats, though. The geographic locations of users remain unclear. Closed models continue to dominate, averaging 80 percent of total usage despite higher prices.
“There could be a world in which the U.S. provides a lot of the low-margin things around the infrastructure, and then China comes along and builds profitable applications on top of that, that actually reaches the end user and manages the surplus value,” Anton Leicht, a visiting scholar at the Carnegie Technology and International Affairs Program, told The Dispatch. A way to prevent that could be to send technical experts abroad through initiatives such as the Tech Corps.
Unlike the U.S., China sends individuals abroad to help support and monitor global diffusion of Chinese technology. A 2021 report from the Center for Security and Emerging Technology found that Chinese science and technology diplomats operate in 52 countries, stationed in Chinese embassies, and work to monitor technological breakthroughs and identify investment opportunities for Chinese firms.
But the U.S. only has a “small sprinkling in select embassies of technology-focused foreign service officers,” Scanlon said. In 2023, Ambassador at Large for Cyberspace and Digital Policy (CDP) Nathaniel Fick described the State Department’s goal of having just a single trained cyber and digital officer in every embassy around the world by the end of next year. But those roles were cut in 2025.
While he supports the program and initiatives aimed at exporting the American stack, Council on Foreign Relations Senior Fellow Chris McGuire sees such programs as “on the margin moves, when we are facing a very significant strategic challenge” of getting global companies to build on the full U.S. stack. He sees export controls on advanced semiconductor chips—which the Trump administration is eschewing— as more effective in curtailing Chinese AI stack deployment. “We need not only positive inducements to push countries to use the U.S. AI stack, such as the Tech Corps, or financing program, but also sticks that will prevent them from using U.S. tech to make or support untrusted infrastructure.”
International development efforts may also face some challenges, including balancing commercial incentives with global development goals. Scanlon noted that the introduction of the Tech Corps is a “pretty striking divergence from the original intent of the Peace Corps program,” given its clear commercial and geopolitical incentives.
Others also fear that the Tech Corps program may not catch on abroad. Meicen Sun, an assistant professor at the University of Illinois Urbana-Champaign and an MIT affiliate noted that China has accumulated a significant customer base in the emerging markets for more than a decade. Over time, this has helped them build up “local connections, business ties, and consumer trust.” Sun notes that these advantages are hard to quantify, but offer important advantages.“Much of it is social and political capital that doesn’t show up in the books,” she told The Dispatch.
