WAICO: China's New Global AI Governance Organization

Twenty-nine countries signed an agreement in Shanghai on July 16, 2026 to establish the World Artificial Intelligence Cooperation Organization (WAICO), a China-proposed intergovernmental body that its founders describe as a permanent institutional home for AI governance built outside the frameworks of Washington, Brussels, and the G7. The signing, formalized as the 2026 World Artificial Intelligence Conference (WAIC) opened in the city, gives Beijing a standing multilateral vehicle to shape the rules governing the technology — and hands the Global South a table it helped set.

WAICO is headquartered in Shanghai and, according to its founding documents, is structured as an independent international organization guided by the principles of the United Nations Charter, with a stated mission to promote “beneficial, safe, and fair” AI development. Its arrival means the world now has two competing scaffolds for AI governance rising in parallel: one anchored in the Western capitals and the G7, and another being assembled in Shanghai.

Who signed, and who did not

The 29 founding members skew heavily toward the Global South. Reporting on the signatory list names Russia, Belarus, Serbia, Cuba, Brazil, Venezuela, Indonesia, Malaysia, South Africa, Senegal, Pakistan, Kazakhstan, and Laos, among roughly a dozen Asian and ten African states. Indonesia confirmed separately that it would participate. The common thread is countries that host no frontier AI labs of their own and have watched the technology’s rules get written by a handful of governments and companies concentrated elsewhere.

Equally telling is the absence list. Neither the United States nor the European Union nor other G7 members are among the founders. So too were major US technology firms largely missing from the Shanghai summit, a gap that analysts read as confirmation that WAICO is, for now, a bloc of developing economies rather than a universal forum. That composition is not incidental — it is the pitch. WAICO is being marketed explicitly as an alternative for nations that feel shut out of Western-led AI diplomacy, echoing the “AI for all” framing organizers have used to argue the body will help bridge the global digital divide.

What Xi said, and what the conference showed

President Xi Jinping, in a keynote tied to the conference, pressed a now-familiar line: countries should take a people-centered approach and develop AI “for the positive and for good.” The rhetoric mirrors language the West uses too — safety, fairness, benefit-sharing — which is part of what makes the split so consequential. The disagreement is less about stated values than about who administers them, whose standards become interoperable defaults, and whose companies benefit from the resulting market access.

The organization launched against a backdrop designed to project scale. WAIC 2026, running July 17 to 20, expanded its exhibition floor past 100,000 square meters for the first time, with more than 1,100 enterprises showing over 3,000 products and technologies and better than 300 global debuts — from humanoid robots to inference chips built to run without US silicon. The message the venue was built to send is that China’s AI ecosystem is broad enough, and commercially deep enough, to anchor a governance bloc of its own rather than negotiate inside someone else’s.

A rival to the UN’s inclusive pitch

WAICO does not exist in a vacuum. Weeks earlier, the United Nations convened the first session of its Global Dialogue on AI Governance in Geneva, a forum established by the General Assembly and pitched on almost identical grounds: a permanent, inclusive venue where governments without frontier labs get a real seat. The two efforts are chasing the same constituency — the countries that will deploy AI without building it — from different centers of gravity.

The overlap is the point of tension. The UN dialogue is designed to be universal but has no enforcement power; WAICO is narrower in membership but gives Beijing a body it can actually steer. For a mid-sized economy weighing where to invest its diplomatic attention, the choice between a broad forum that convenes and a focused bloc that acts is a live one. And it is being made while a separate Western track — the EU, which has moved furthest with binding law through its AI Act and general-purpose model rules — sets standards that any company selling into Europe must meet regardless of which club it joins.

The technology backdrop

WAICO’s launch lands at a moment when the assumption that the AI frontier is a US preserve is under real strain. Chinese labs have narrowed the gap fast: Moonshot’s Kimi K3 open-weight release days earlier drew comparisons to leading US systems, part of a broader pattern in which open-weight models keep closing the distance to the closed frontier. As capable models become downloadable rather than gated behind a US API, the leverage that came from controlling the frontier — and the standards around it — erodes.

That erosion is what makes a governance bloc strategically valuable. If a country can obtain a competitive open-weight model without Western permission, the remaining points of influence shift to the layers around the model: compute, chips, safety standards, certification, and the commercial rules for cross-border deployment. China has been building the physical side of that stack aggressively, pairing model releases with hundreds of billions in domestic AI infrastructure and a growing effort to run its ecosystem on non-US hardware. WAICO is the diplomatic complement — an attempt to convert that industrial base into rule-making authority. It also arrives as Chinese AI models make quiet inroads into US enterprise adoption, a sign the technical products are already crossing the very lines the governance blocs are drawing.

What it means

WAICO is, for now, more institution than instrument. Twenty-nine signatures and a Shanghai headquarters do not by themselves set a single binding standard. But institutions are how influence compounds, and the strategic logic is clear enough to take seriously.

Who it serves. The immediate beneficiaries are the Global South founders, who gain a venue where their preferences carry weight instead of being an afterthought to a G7 communiqué. Beijing benefits more durably: a standing organization it helped charter gives it a permanent seat at the head of a governance conversation it was previously trying to influence from outside. For countries choosing an AI vendor and a rulebook, a bloc that offers both models and standards in one package is a coherent proposition.

The fault line to watch. The world is not converging on one set of AI rules; it is bifurcating into at least two, with the UN’s universal dialogue caught in between. The practical question for companies is interoperability. A firm deploying a model across borders may soon have to satisfy EU law, a US-aligned framework, and WAICO standards at once — three certification regimes for one system. Fragmentation of that kind raises compliance cost and, over time, pressures the technology itself to split along geopolitical lines, the way telecom and payments infrastructure already have.

What comes next. The measure of WAICO will be whether it produces anything binding — shared certification, model-safety baselines, data-governance rules its members actually adopt — or whether it remains a signaling exercise, a forum that meets and declares. Watch the membership: if economies currently sitting inside Western frameworks begin signing on, the bloc stops being a Global South caucus and becomes a genuine rival standard-setter. And watch the hardware, because governance follows compute. As long as WAICO’s members can build and run competitive models on their own terms, the organization has something concrete to govern. The launch was the easy part. Whether Shanghai can turn 29 signatures into standards the rest of the world has to reckon with is the question the next year will answer.