Open Source at the UN, Export Curbs at Home: Decoding China’s Two-Faced AI Strategy in 2026

In July 2026, Beijing presented two sharply different faces to the world. At the United Nations in Geneva, Chinese diplomats championed open-source AI as a tool for bridging global digital divides, capacity-building, and inclusive governance . On the same day the UN dialogue opened, Reuters reported that Chinese authorities had held meetings with top tech firms to consider restricting overseas access to the country’s most advanced AI models — including open-source ones .

This is not a contradiction. It is a calibrated dual strategy that reveals how Beijing views open-source AI as a tool of diplomatic influence and frontier AI as a tool of national power.

The Open-Source Push in Geneva

China’s public-facing AI diplomacy in 2026 has been heavily focused on UN-led governance events in Geneva:

• UN Global Dialogue on AI Governance (6–7 July 2026): Convened at Palexpo in Geneva, the dialogue was established by the UN General Assembly as a platform for all governments, private sector, academia, and civil society to discuss international cooperation on AI governance . Proposed thematic clusters included bridging AI divides, capacity-building, access, and digital foundations .
• UN Open Source Week (June 2026): A dedicated session on “Open Source for AI and Emerging Technologies” assessed open source as a gateway to a sustainable AI future, aiming to bridge digital divides and promote responsible AI development .
• Global Network of Centers for Exchange and Cooperation on AI Capacity Building (7 July 2026): Established in Geneva with UN support, this network brought together leading AI institutions from Africa, Asia, Europe, Latin America, and the Caribbean to enhance international cooperation .

Chinese officials have strongly aligned with these efforts. At a November 2025 “Geneva Dialogue: Solidarity for AI Governance” co-convened by China, ambassador Chen Xu stated that “a number of open-source, low-cost, and high-efficiency innovative achievements have provided a new paradigm for global AI development” .

In 2025, China also launched the International Open Source AI Cooperation Initiative at the World AI Conference in Shanghai .

The Reported Export-Control Discussions

On 7 July 2026 — the opening day of the UN Global Dialogue — Reuters exclusively reported that Chinese authorities had held meetings with top tech firms over the previous month about potentially restricting overseas access to China’s most advanced AI models, including those yet to be released .

Key details from the Reuters reporting:

• The meetings were led by China’s Ministry of Commerce .
• At least three companies were involved: Alibaba, ByteDance, and Zhipu AI (referred to as Z.ai in one report) .
• The discussions covered both closed-source and open-source models, suggesting even currently available open-weight models could face new restrictions .
• Officials discussed making the leak or theft of proprietary AI technology an offence under China’s national security laws .

This follows a November 2025 directive requiring state-funded data-centre projects in China to use only domestically produced AI chips, signaling a broader push toward technological self-reliance .

What the Contradiction Reveals

The dual strategy can be summarised in a simple table:

Track	Message	Audience	Goal
UN Geneva	Open-source AI, bridging AI divides, capacity-building, global governance cooperation	Developing countries, global public, UN member states	Diplomatic soft power, norm-shaping, market access for Chinese AI in the Global South
Domestic policy meetings	Restrict overseas access to frontier models, reduce reliance on foreign AI chips	Chinese tech firms, domestic ecosystem	National security, technological sovereignty, protecting a strategic advantage

Beijing appears to want to shape the rules of global AI governance and position itself as a leader on AI capacity-building, especially for countries concerned about digital divides . Promoting open-source AI in UN settings serves that diplomatic goal, builds goodwill, and supports wider adoption of Chinese AI tools abroad .

Simultaneously, China’s leadership is considering how to limit overseas access to its most advanced AI capabilities, mirroring the logic of US export controls on advanced chips and AI models . Washington has restricted access to advanced AI-relevant chips and systems on national-security grounds, and US policymakers are debating restricting foreign access to advanced AI models themselves .

As Reuters noted in a follow-up newsletter: “China appears intent on pulling a silicon curtain around its own AI technology — including its open-source models — just as Washington has been ramping up restrictions on foreign access to top-shelf AI” .

A Mirror, Not a Contradiction

Both major powers increasingly treat AI governance as an arena of geopolitical competition, not just technical cooperation.

The key insight: open-source AI is a tool of influence; frontier AI is a tool of power. Beijing can promote openness and capacity-building where it helps win friends and shape norms, while guarding the most advanced capabilities to preserve its competitive and national-security position. This mirrors the broader US approach of engaging internationally on AI governance while tightening controls around sensitive AI technologies at home.