Many telecom companies learned a tough lesson in the 2010s: troubled Chinese technology can carry hidden costs and risks. Today, just a handful of years since the Federal Communications Commission (FCC) formally designated Chinese 5G equipment from provider Huawei a national security threat to U.S. communications networks, America is potentially poised to relearn this lesson.
Much like China attempted to undercut Western competition with compromised-but-inexpensive 5G infrastructure, China today is attempting to dominate the AI supply chain and project the CCP’s malign influence across the globe. Only this time, instead of selling 5G telecom equipment, Chinese labs are rapidly disseminating open-weight AI models like DeepSeek.
The AI race has raised the stakes, but the contours of US-China competition are fundamentally the same. Chinese companies cannot and should not be trusted to build the digital infrastructure of the 21st century. For the United States, reclaiming our lead in open-weight intelligence is not just a matter of principle – it’s a national security imperative.
Embracing an 'open approach'
While much time has been spent in policy circles considering whether AI promotes human safety and responsible use, a perhaps more important debate is simmering behind the scenes over whether America should compel AI companies to embrace an “open” approach or favor a “closed” AI model. Upon close examination, it is clear that open platforms provide greater protection for U.S. national security and sovereignty—seemingly critical priorities for policymakers—than the alternative. Nowhere is this more evident than in the global impact and influence of U.S. AI products that directly compete with offerings from communist China.
The simplest explanation of the difference between open and closed AI is whether the infrastructure can be modified by the user. That is, an open platform allows AI developers to alter and improve the fundamental components (e.g., base code, algorithm weights) and to add functionality and capabilities on top. In this process, developers can collaborate, share, and challenge others’ work product. At the same time, users can inspect, audit, and verify the fundamental aspects of the underlying AI structure, providing some assurances needed for those leery of AI’s security and safety.
In the global AI marketplace, many countries and international companies are favoring open AI approaches. This allows them to keep much of the AI infrastructure, servers, and data within arm’s reach while enabling customization of unique features to meet evolving needs. That level of perceived control is important and is reportedly pushing the foreign marketplace away from the majority of closed U.S. AI products.
A global AI marketplace
China is trying to meet this demand and dominate the global AI marketplace. The goal is not only to reap the largest share of revenues but also to threaten the U.S. role in the international economy. China is actively trying to sell, distribute, and convince key nations and applicable companies that its products are superior to those produced in the U.S., partly because they are built using an open model. Consider the eye-opening and mind-changing moment when DeepSeek’s inexpensive, open-source R1 model was revealed to the world last year.
Part of the Chinese AI “bargain” is that the country’s values are injected into the integration process. That means the very elements of a free society are extinguished. For example, global censorship tools and embedded remote access points for Chinese spying are built into Chinese AI products. Moreover, the expansion of China’s AI products allows the CCP to exert greater influence over related elements of internet infrastructure, including standards and governance.
From a U.S. policy perspective, China’s AI growth is extremely problematic. The Trump Administration indicated as much in its AI Action Plan. Specifically, it acknowledged that open AI models offer commercial and government interests options to transfer sensitive data that “cannot be sent to closed model vendors.” Unpacking that concept highlights concern about classified and sensitive information: reliance on Chinese products may be unacceptable, but there is a real risk that China's commitment to open approaches could lead to global standards in certain critical situations. The administration even went so far as to state, “We need to ensure America has leading open models founded on American values. . . [T]he Federal government should create a supportive environment for open models.”
Given the complexities and vast capital requirements, the AI market appears to the global community as consisting of American closed AI companies and Chinese substitutes. That should prompt policymakers to ensure that U.S.-based open AI is alive and well, as ceding the open model will leave an entire lane to our adversaries. With that, the U.S. would also be ceding the foundational democratic principles that have driven technology development for decades.
Michael O’Rielly is currently President at MPORielly Consulting Inc., President and CEO at the Media Institute, Visiting Fellow at the Hudson Institute, Adjunct Senior Fellow at the Free State Foundation, Special Advisor with Crest Hill Advisors, and Member of the International Advisory Council of APCO Worldwide. He served as a Commissioner at the Federal Communications Commission from 2013 through 2020. Before joining the FCC, Mr. O’Rielly held leading staff positions during 20 years on Capitol Hill in both the U.S. Senate and House of Representatives, ending as Policy Advisor in the Office of the Senate Republican Whip.
Opinion pieces from industry experts, analysts or our editorial staff do not represent the opinions of Fierce Network. Send us your opinion pieces and we'll take a look. Email fiercenetwork@questex.com.