Xi Jinping’s speech at the 2026 World Artificial Intelligence Conference (WAIC) in Shanghai was widely covered as an endorsement of open-source AI and a rebuke of US chip restrictions. But a deeper reading, offered by Trivium China‘s Kendra Schaefer in the firm’s authoritative post-WAIC weekly recap, reveals something more strategically significant: Beijing is not trying to win the AI race this year. It is playing a decade-long game for who owns the AI stack that the developing world runs on, who sets the rules of global AI governance, and ultimately, whose industries make the best use of these tools for economic gain.
Balancing Development and Security: Xi’s Dual Message at WAIC
Xi’s speech was, in Trivium’s framing, “an exercise in balancing development and security.” He expressed full-throated support for AI development and reaffirmed China’s commitment to open-source AI, while simultaneously signaling that AI safety regulation is moving up Beijing’s priority list. This dual message, accelerate, but also regulate, reflects the inherent tension in China’s AI strategy. Moving too fast risks deploying systems that embarrass the government or destabilize society. Moving too slowly cedes ground to US competitors.
The open-source commitment, in particular, is not primarily about technical philosophy. It is about geopolitical positioning. The global popularity of Chinese open-source tools such as DeepSeek, Kimi, and Z.ai’s GLM series represents what Trivium describes as a “once-in-a-generation opportunity” to bring the international community, and especially the Global South, onto China’s AI stack. In the absence of cheap, open US alternatives (OpenAI and Anthropic remain largely closed-source and expensive), Chinese open-source models are filling a vacuum. As China’s open-source AI models now rank first globally in downloads, the strategy is already bearing fruit.
WAICO and China’s Global AI Governance Offensive
The formal launch of the World Artificial Intelligence Cooperation Organization (WAICO) on the eve of WAIC was not a coincidence. WAICO is Beijing’s bid to compete with the US on global AI governance and offer the Global South an alternative to Western-led bodies. As The Diplomat has analyzed, WAICO is a governance offensive, not a goodwill gesture. It is designed to give Beijing institutional leverage over the norms and rules that will govern AI deployment in dozens of developing countries. Countries that adopt Chinese AI infrastructure and Chinese governance frameworks become, in effect, part of China’s technological sphere of influence.
Xi’s WAIC speech directly reinforced this agenda. By framing open models as a counter to US restrictions, Xi positioned China as the champion of an open, accessible AI future, in contrast to what he implicitly characterized as the US’s restrictive, self-serving approach to AI governance. The message was aimed as much at the Global South as at domestic audiences: China is the country that will share its AI with the world, while the US hoards it.
The Regulatory Tightening That Contradicts the Open-Source Pitch
Trivium’s analysis flags a significant near-term risk that complicates the open-source narrative: growing speculation that AI safety concerns may force Chinese regulators to place stronger controls on both the release of new open-source models and the distribution of open model weights. Xi’s speech indicated that China will double down on AI safety, and the Ministry of Commerce is currently consulting domestic AI and chip firms on tightening export controls over advanced technologies. Regulators are also discussing ways to prevent offshore chip firms from fabricating chips designed by Huawei, Alibaba, or ByteDance, a move that would effectively extend China’s technology control perimeter beyond its borders.
This creates a potential contradiction at the heart of China’s strategy: promoting open-source AI globally while simultaneously tightening controls on what can be released and where it can be manufactured. How Beijing resolves this tension will be one of the defining policy questions of the next 12 months.
The Long Game in Numbers
The Trivium analysis is worth reading alongside the hard data on China’s open-source AI adoption. In markets across Southeast Asia, the Middle East, and Africa, where developers cannot afford expensive US frontier APIs, Chinese open-source models are often the only practical option for building AI-powered applications. This is precisely the dynamic that Trivium identifies as the core of China’s decade-long play: not winning the benchmark race today, but owning the infrastructure layer that the next generation of global AI applications is built on.
The WAICO launch reinforces this reading. By establishing an international governance body before the US has consolidated its own approach to global AI governance, China is positioning itself to shape the norms, standards, and dispute resolution mechanisms that will govern AI deployment in dozens of countries. For now, China’s open-source AI strategy remains intact, but the regulatory environment around it is tightening, and the next 12 months will determine whether Beijing can hold both sides of this strategic equation together.