Hong Kong Aspires to Be a Global AI Development Hub
Strong R&D capabilities and robust infrastructure of computing power
Home to five universities ranked among the world’s top 100, Hong Kong boasts a solid foundation in scientific research. Cheong revealed that through the AIR@InnoHK research cluster, the Government has cumulatively funded the establishment of 16 global laboratories which focused on AI and robotics, bringing together more than 1,000 research talents. The Hong Kong Science Park and Cyberport have also gathered nearly 1,000 outstanding AI-related enterprises and start-ups in recent years. To further strengthen the frontier basic research, the Government launched the HKD3 billion Frontier Technology Research Support Scheme, drawing in leading international research talents including those engaged in AI to Hong Kong, In addition, the Hong Kong Park of the Hetao Shenzhen-Hong Kong Science and Technology Innovation Cooperation Zone was officially opened in December 2025. The Park focuses on developing cutting-edge technologies including AI and data science. By improving infrastructure and cross-boundary data mechanisms, the park promotes “AI+” demonstration applications which is conducive to the attraction of I&T resources and talent from home and abroad.
Cheong believes that computing power is the cornerstone of AI development. The computing capacity of the AI Supercomputing Centre in Cyberport was ramped up to 3 000 PFLOPS in 2025. In addition, the Government launched a HKD3 billion AI Subsidy Scheme, which has already supported over 20 university and enterprise projects to drive more scientific research breakthroughs so far. To meet long-term computing power demand, the Government has planned a site of 10 hectares in Sandy Ridge in the Northern Metropolis for the development of a data facility cluster. The site was put up for tender with a “two-envelop approach” earlier. Now the tender evaluation process is underway and the Government is striving to announce the result as soon as possible. This data facility cluster will help promote the development of data and AI-related industries in Hong Kong and even the GBA area.
Government drives AI application
Application is a key to AI development. Cheong said that to actively expand the application of AI in public services, the AI Efficacy Enhancement Team was established to co-ordinate and steer government departments in the effective use of AI technology. The Digital Policy Office (DPO) will provide AI Toolbox and roll out “AI flagship projects”, using AI to assist departments in expediting the process of handling, vetting and approval of various types of licences and permits as well as applications. Besides, the DPO will provide an “AI Assistant” to users of “iAM Smart” and “Digital Corporate Identity” to answer enquiries and offer personalised services and to recommend suitable subsidy scheme and public support services to enterprises in particular the small and medium-sized enterprises (SMEs). Meanwhile, the application of AI tools - HKPilot, HKChat and HKMeeting developed by the Hong Kong Generative AI Research & development Center (HKGAI) has been piloted internally within the Government since mid-2024 while the mobile version of HKChat has also been rolled out for a public trial in December 2025.
To promote a wider adoption of AI, the Government proposed in “The 2025 Policy Address” to optimize the Digital Transformation Support Pilot Scheme and provide subsidies for SMEs to adopt AI and cybersecurity related solutions. The Hong Kong AI Research Institute will be established this year to facilitate upstream R&D, the transformation of midstream and downstream research outcomes, and the exploration of application scenarios, as well as empower various industries to adopt AI. At present, various industries are making solid progress in rolling out more AI applications, for example in financial services, the Government has issued a policy statement on the application of AI in the financial market and rolled out the Generative AI Sandbox initiative to support the financial industry in developing and testing innovative AI solutions.
Facilitating data flow
Governance and regulation, talent cultivation
The Government is committed to striking a balance between AI development and regulation. Cheong said that the Innovation, Technology and Industry Bureau has published the Policy Statement on Facilitating Data Flow and Safeguarding Data Security in Hong Kong, to optimise data governance. The DPO has issued Ethical Artificial Intelligence Framework and other guidelines to provide recommendations for the development, design, processing and secure use of AI services.
Cheong pointed out that strengthening talent cultivation is the key to driving the development of the AI industry. She added that the Government has been fully advancing various initiatives to nurture local AI talents and attract overseas ones. For primary and secondary education, the Quality Education Fund has earmarked HKD2 billion to promote digital education. The Government has also encouraged universities funded by the University Grants Committee to add new programmes relevant to STEAM and the “eight centres”.
Call for the business sector to accelerate digital transformation
Establishment of Supercomputing Centre Helps Address Research Challenges
The HKSAR Government is committed to positioning Hong Kong as an international hub for AI industry exchange and collaboration, and supports Cyberport in setting up Hong Kong’s first and largest Artificial Intelligence Supercomputing Centre (AISC). Leveraging its super computing power, the centre supports the implementation and application of scientific research, thereby boosting the development of the digital economy and innovation and technology sectors.
Officially launched in December 2024, the AISC supports and enhances the R&D capabilities and business transformation of local industries such as health technology, AI, manufacturing and new energy technology. Coupled with the three-year HKD3 billion Artificial Intelligence Funding Scheme, the initiative encourages local universities, startups and other enterprises and organizations to make full use of the AISC’s computing resources.
Instant resolution of pain points in time, manpower and costs
As of October last year, 15 applications from local universities, government policy bureaux and departments have been approved, covering diverse areas including medical large models, generative AI and new materials. Crystal Fok, Head of Business Development at Cyberport, revealed that the AISC boasts a computing speed equivalent to processing nearly 10 billion images per hour, demonstrating its powerful and high-efficiency computing capabilities. She firmly believes that the AISC helps tackle the challenges of insufficient supercomputing capacity and fragmented resources that have constrained local research development, building an ideal innovation ecosystem and a solid foundation for the long-term growth of various industries.
“The chip configuration of the AISC represents the highest performance currently available in Hong Kong, enabling it to process massive volumes of data and conduct model training. For instance, some local universities aim to develop research projects related to cancer screening, but were previously unable to process large-scale imaging data due to hardware limitations. Building an in-house computing infrastructure would entail substantial operational costs in terms of procurement, time, and manpower. In contrast, leveraging the AISC’s computing power can instantly resolve these pain points.”
AI sandbox creates a safety net to advance industrial R&D and application
Citing the Generative AI Sandbox jointly launched by Cyberport and the Hong Kong Monetary Authority as an example, Fok explained that this “safe space” provides a collaborative platform for enterprises and banks. It enables the financial sector to safely and effectively explore the application of generative AI technologies, identify potential risks in the process, and obtain early feedback on financial regulation and compliance, thereby facilitating the formulation of practical guidelines and regulatory frameworks.
Fok emphasized that the implementation of a scientific research project or a tech innovation product involves long-term development and investment, which demands high-end computing infrastructure but incurs substantial costs with uncertain returns. Additionally, there may be potential gaps in supporting policies, which could impede the overall progress of technology application. The launch of the sandbox, combined with the AISC’s computing power, comprehensively addresses the concerns and potential risks of users including research teams, investors, and enterprises.
Intelligent transformation solutions empower latecomer enterprises to catch up
Amid the rapid advancement of innovation and technology, numerous enterprises are focusing on the development window opened by AI and technological transformation, accelerating the upgrading of industrial development. Fok pointed out that many enterprises embarked on industrial digitalization in the early stages, gradually progressing towards industrial intelligence. However, some enterprises that started their transformation journey relatively late are still striving to catch up.
“Today, the market is well-supplied with solution providers offering industrial upgrading and transformation services, and leveraging AI for intelligent identification and processing of massive handwritten document data is no longer a challenge.”
Fok opined that late-moving enterprises can leverage these solutions to skip the digital transformation phase entirely and leapfrog into intelligent management. This allows them to integrate the cumbersome information and data accumulated from traditional operations in one go, achieving an accelerated transition to intelligent transformation.
