Author: Revaz Topuria, Research Fellow
In February 2025, world leaders gathered in Paris for the third global AI summit. Sixty-two countries, the African Union, and the European Union signed a joint declaration embracing an open, inclusive, and ethical approach to artificial intelligence. The United States and the United Kingdom did not join the declaration. Speaking at the summit’s closing ceremony, US Vice President JD Vance made the administration’s position unambiguous: “American AI will not be co-opted into a tool for authoritarian censorship.” In counterpoint, the Chinese delegation signed the declaration, actively seeking to expand its own global role in the AI domain.
That moment made clear something that had been taking shape for years: the competition between the United States and China in artificial intelligence is not merely a race to build the most powerful models or the fastest chips. It is a deeper struggle over who will write the rules, the norms, standards, and institutional frameworks, that will determine how AI is developed, deployed, and regulated globally. In the time since that summit, it has become even clearer that both states are actively working to build AI institutions around their own rules for this contest.
The Price of Standard-Setting
Most of the attention in the US-China AI race tends to go to new model releases and the chip war, but to understand what this competition actually entails, it is worth examining the significance of technical and governance standards. Standards are not neutral administrative documents. They determine which technologies are interoperable, which platforms become dominant, which companies gain market access, and which governance values become embedded in the infrastructure of the digital economy. A country or bloc that shapes international AI standards gains structural advantages for decades, since other actors will automatically adapt their own systems to its specifications.
History already offers an example of this. In its time, the United States achieved dominance in the internet era’s foundational standards, in protocols, platforms, and technical architectures. As a result, every other actor either fully adopted the standards the US had established, or adapted them with minor modifications. However, at the time this dominance was not directly the result of deliberate state policy; it came about through the success of American companies whose products became the global standard. China absorbed this lesson and drew a specific conclusion: allowing another power’s de facto standards to become the global standard is a form of strategic dependency that it cannot allow to be repeated in the AI era. The question now is whether Beijing’s more active and coordinated approach to standard-setting can overcome the structural advantages created by the dynamism of America’s private sector.
China’s Effort
China’s strategy for shaping global AI governance has developed rapidly and is now operating simultaneously along several parallel tracks. The clearest evidence of this appeared on July 16, 2026, when representatives of 29 countries gathered in Shanghai to formally establish the World Artificial Intelligence Cooperation Organization (WAICO), as an independent intergovernmental body headquartered in Shanghai. The proposal, first put forward by Premier Li Qiang in July 2025 and then personally championed by Xi Jinping at the APEC summit in Busan that November, is no longer merely a proposal. It is now an established institution, and its opening ceremony was attended even by UN Secretary-General António Guterres, which represents a significant achievement for Beijing.
Alongside WAICO’s founding, China also held the World AI Conference, at which Xi Jinping was the keynote speaker. In his address, Xi stated clearly that AI governance should be “an important driver of shared prosperity and shared security,” rather than something dictated by a single state. His speech also included commitments: he pledged 5,000 AI training and seminar opportunities for developing countries over five years, and announced the creation of cooperation centers with ASEAN, the African Union, and BRICS. Among WAICO’s founding countries are Russia, Brazil, Indonesia, Pakistan, Kazakhstan, South Africa, and other countries of the Global South. Not a single major Western democracy appears among the founders. This step makes clear that Beijing is not trying to win over Washington’s allies. Instead, it is building a parallel institutional order for those states that feel they are not adequately represented in the existing order.
At the same time, China is also working actively on the technical track. In 2025, China held chair and manager positions in ISO/IEC JTC 1/SC 42, the primary international committee for AI standardization, where it submitted 8 international proposals and had more than 400 registered experts involved. This is technical work that is largely invisible to the general public, but its effect grows over time: once a standard is adopted internationally, technological architecture begins to be built specifically around that standard.
China’s actions in this regard carry a clear ideological dimension as well. Beijing’s narrative, that AI should be an “international public good,” that governance should be inclusive, that the digital divide between developed and developing countries must be closed, is attractive to many countries that feel excluded from the G7’s Hiroshima Process or the Bletchley Summit series, where the agenda is set largely by developed Western countries. This kind of rhetoric from China implies a long-term effort at building dependency and normative alignment: countries that build their own AI infrastructure on Chinese technology, train their own engineers through Chinese programs, and adopt governance frameworks modeled on Chinese templates will, over time, naturally come to support Chinese positions in international forums.
Washington’s Counter-Approach
America’s approach to setting global AI norms has generally been more market-oriented and, until now, has not included an institutional framework. However, in response to China’s actions, the US is also shifting its outlook and building its own AI coalition alongside WAICO.
The State Department established Pax Silica in December 2025 as its flagship initiative on AI and supply-chain security, on the argument that the AI economy requires its own dedicated economic-security framework, separate from the G7 and G20. Unlike WAICO, Pax Silica is not conceived as an ethics or governance body. Its central idea is securing the AI economy’s physical infrastructure, critical minerals, energy, semiconductors, manufacturing, and AI infrastructure, and reducing dependence on supply chains under Chinese control. Nevertheless, the institutional logic is identical to WAICO’s: assembling a coalition of states, coordinating policy and investment. As of now, Pax Silica’s declaration has 25 signatories, including the EU, the UK, Japan, South Korea, India, the Philippines, the UAE, and Qatar.
It is worth noting that the US is not limiting itself to Pax Silica alone. On July 23, 2025, the Trump administration unveiled the US AI Action Plan, titled “Winning the Race,” which states directly that “whoever has the largest AI ecosystem will set global AI standards and reap broad economic and military benefits.” At the same time, the Action Plan provides for “exporting the full AI technology stack, hardware, models, software applications, and standards, to every country that joins America’s AI alliance.”
So Washington, too, is building its own coalition in the AI domain, though with approaches and institutional frameworks different from Beijing’s.
The Future of AI Standards
What makes the US-China competition over setting AI norms analytically interesting is that both sides are offering the rest of the world their own, distinct model of AI governance.
China’s model places its main emphasis on multilateralism, development, and sovereignty. WAICO, the Global AI Governance Action Plan, and the BRICS cooperation centers are all framed in terms of collective decision-making and the fair distribution of benefits, presented openly in opposition to what Beijing calls Western-dominated governance initiatives. The American model, by contrast, rests on the claim that open markets and private-sector leadership produce better AI, with its emphasis placed primarily on innovation and development.
Which country manages to establish its own vision and standards for AI globally will also determine which values, assumptions, and political preferences become embedded in the infrastructure of the AI-enabled world. Where, as recently as a year ago, this competition was still substantially rhetorical, consisting largely of declarations signed or refused, and proposals put forward without institutional form, the situation today has changed. WAICO is already a functioning organization, with its headquarters in Shanghai and 29 member states. Pax Silica is a functioning coalition, with more than 25 signatories and physical infrastructure projects underway from the Philippines to Panama. Both states have moved from arguing over the rules to building the institutions that will write them. Which of the two will prove more successful in establishing these rules globally, only time will tell.
This article was translated from the original language with the assistance of AI tools and revised by the author.