China offers free AI to developing countries; G7 limits access to US models

What happened?

On Wednesday, China's top diplomat Wang Yi stated that Beijing is "accelerating the establishment of a global AI cooperation organization" and invited all countries to join. The announcement coincided with the closing of the G7 summit in France, where leaders debated granting access to US AI models only to "trusted partners," according to Reuters. This is not an isolated event: China had already proposed the Global AI Governance Initiative in 2023 under the United Nations framework, and the G7 launched the Hiroshima AI Code of Conduct for developers in 2023. The difference now is the concreteness of the measures: China moves from statements to creating a structure, while the G7 moves from principles to export controls.

Why is it important?

This contrast reveals two opposing geopolitical visions for the future of artificial intelligence. China seeks to position itself as an inclusive, no-strings-attached provider, appealing to Global South countries that fear being left behind. According to World Bank data, over 60% of the population in sub-Saharan Africa lacks access to broadband internet, and free Chinese AI could be a technological shortcut. Meanwhile, the G7 prioritizes security and control over technologies deemed critical, but risks alienating nations that need access to AI tools for development. The Chinese proposal is part of a broader strategy: since 2020, China has invested over $50 billion in digital infrastructure along the Belt and Road, including submarine cables and data centers. The global AI organization could be the next step to standardize these investments under its leadership.

Consequences and context

The Chinese proposal is not an act of charity: it reinforces its influence in regions like Africa, Asia, and Latin America, where it has already deployed digital infrastructure. For example, in 2024, China signed AI cooperation agreements with over 40 countries, including Brazil, South Africa, and Pakistan. Moreover, the global organization it proposes could establish alternative standards to Western ones, fragmenting the AI ecosystem. This echoes the fragmentation of the internet in the 2010s with China's digital sovereignty model (Great Firewall). On the other hand, the G7 approach, led by the United States, reflects growing anxiety over technology leakage and industrial espionage, but could accelerate developing countries' dependence on Chinese solutions. A 2025 CSIS report estimates that US export controls have reduced AI chip sales to China by 40%, but have spurred Chinese companies to develop domestic alternatives, such as Huawei's Ascend 910B chip, which competes with Nvidia's H100.

For tech companies, this schism means a bifurcated market: startups in non-aligned countries will have to choose between ecosystems, while giants like OpenAI, Google, and Microsoft will see limited expansion in regions where China offers free alternatives. For example, in India, where the government has promoted national AI, companies like Reliance Jio are already testing Chinese open-source models. End users, especially in the developing world, could benefit from free tools, but at the cost of privacy and data sovereignty. The EU's AI Act, which comes into effect in 2026, could clash with Chinese standards if they do not guarantee equivalent data protection.

What readers should know

• China's global AI organization: No concrete details yet on its structure, funding, or governance. It is speculation whether it will be truly open or serve to promote Chinese standards. However, sources close to the Chinese Foreign Ministry indicate it could follow the model of the Shanghai Cooperation Organization, with consensus-based decisions and emphasis on non-interference.
• G7 export controls: No specific lists of models or countries have been published, but they are expected to include frontier systems like GPT-4 and Gemini, as well as hardware from Nvidia and AMD. The EU is pushing for a more nuanced approach, allowing exceptions for humanitarian uses, such as AI applied to natural disasters.
• Impact on startups: Emerging companies in developing countries could benefit from free Chinese AI, but risk technological dependence and potential future restrictions. For example, if China imposes licensing conditions that require data sharing, startups could lose control over their intellectual property.
• International community reaction: The UN and UNESCO have called for an inclusive global framework, but so far there is no consensus. The Chinese proposal could gain traction in forums like BRICS, which already created an AI working group in 2024. However, countries like Japan and South Korea have shown caution, fearing that the Chinese organization could undermine democratic standards.

"Artificial intelligence should not be a privilege for a few, but a global public good," Wang Yi declared, in clear contrast to the G7's restrictive stance.

In summary, we are witnessing a struggle for ethical and technological leadership in AI. While China offers access with no apparent strings attached, the G7 tries to shield its competitive advantage. History shows that technology blocs tend to consolidate: in the 5G standards war, China managed to impose its technology in over 50 countries through Huawei. Now, with AI, the battle is fought on two fronts: data control and the definition of what is ethical. The outcome will define not only who controls the next wave of innovation, but also how its benefits are distributed globally.