Hong Kong Cyberport is rapidly transforming into the preferred launchpad for Chinese artificial intelligence companies aiming for international expansion. It has evolved from a local technology park into a comprehensive globalization platform, connecting innovations from mainland China with markets in Southeast Asia, the Middle East, and global markets.
The first aspect of this transformation is computing power. The Cyberport AI supercomputing center, launched in 2024, provides 3000 PFLOPS of capacity. This center is supported by a three-year AI funding fund totaling 3 billion Hong Kong dollars, which covers between 70% and 90% of AI development costs. The CAS Innovation Institute in Hong Kong utilized this computing base to train the CARES Copilot model, which became the most downloaded model in its category worldwide.
The second key element is standards and regulatory compliance. Chinese AI companies often encounter unforeseen obstacles in foreign markets, even if their products have successfully passed tests on the mainland. These barriers relate to data governance, contract law, liability frameworks, and cultural expectations. Cyberport offers a controlled international environment that allows companies to adapt their developments to global standards before full-scale entry into the international market.
A major Chinese cloud provider, WaveCloud, implemented a three-stage strategy: first, it established joint laboratories with Hong Kong universities to incubate technologies; then, it officially settled in Cyberport in 2024, where 26 international technologies were incubated; and finally, it launched the Hong Kong Smart Cloud Platform in 2025. As a result, the company not only gained market access but also secured a leadership position in setting international AI standards using Hong Kong as its base.
The third aspect is market access. Yunji Technology, which deals in service robots and was listed on HKEX in October 2024 as the first service robot stock, demonstrates how Cyberport connects product technical capabilities with market readiness. Yunji's robots already serve over 40,000 hotels globally, performing 750 million service operations annually across 20 countries. However, Hong Kong taught the company a different lesson: market acceptance depends not only on technical specifications but also on subtle issues such as the readiness of hotel staff and guests to accept robots, the applicability of fire safety and property regulations, defining responsibility for failures, and adapting workflows developed for mainland Chinese hotels to the management styles of overseas hotels.
The CEO of Cyberport, Dr. Zheng Sunyang, described this process as a transition from isolated company registration to the expansion of an industrial cluster. Meanwhile, many cities in mainland China, including Zhongguancun, Suzhou, and Guangxi, are opening specialized overseas representative offices in Cyberport. With over 200 connections to venture capital, IPO assistance programs, and organized trips to Southeast Asian and Middle Eastern markets, Cyberport has evolved from a mere landing point in Hong Kong into a comprehensive globalization platform where Chinese AI companies learn not just to export technology, but to become multinational technology enterprises.
The conference on artificial intelligence and high-level global AI governance will take place on July 17, 2026, at the Shanghai Exhibition Center in China. This meeting, along with the World AI Conference 2026 (WAIC), is dedicated to balancing innovation development with ensuring safety amid rapid AI progress.
During the event, several international specialists emphasized that China's early steps in AI regulation, as well as the country's active participation in initiatives to form international standards and build capacity, provide valuable experience for the global community in developing responsible and comprehensive AI. Furthermore, a roadmap for international ethical AI governance was presented, aimed at stimulating more coordinated and practical international cooperation in this field.
Mark Nisberg, Executive Director of the Human-Compatible AI Center at the University of California, Berkeley, stated China's significant contribution to global AI governance. He noted that several regulatory systems were introduced in China for the first time, adding that international researchers regularly exchange views with developers of Chinese norms, viewing this practice as a valuable benchmark.
Experts also highlighted China's role in establishing international AI standards. Sebastian Hallensleben, Chairman of JTC 21 CEN-CENELEC, where European standards supporting EU regulation are developed, noted that AI standardization is inherently an international process, despite regional differences. He stressed that China is one of the key participants in this process, leading several working groups and serving as an editor for numerous standards projects.
Another aspect covered at the conference was AI capacity building. As part of WAIC, China announced plans to provide 5000 opportunities for AI training and workshops to developing countries over the next five years.
Shahbaz Khan, Director and Representative of the UNESCO Regional Office for East Asia, pointed out that the international community already possesses principles of ethical AI, including the UNESCO Recommendation on the Ethics of AI. Nevertheless, he particularly highlighted the need for countries to have sufficient capacity to understand these principles and integrate them into their legal systems. In this context, he noted that China's capacity-building efforts will provide significant support.
