Developing a comprehensive value chain in artificial intelligence, cementing an ecosystem, and strengthening collaboration in the Greater Bay Area are crucial in beefing up Hong Kong’s competitiveness in the sector. More important, perhaps, is a roadmap that would encompass all these aspects to generate a coordinated development plan. Oswald Chan reports from Hong Kong.

Editor’s note: Hong Kong is reinforcing its value proposition for innovation and technology with a new focus on developing artificial intelligence as a core industry. The first part of the series looks at how the city can make itself more competitive in the global AI race.

Artificial intelligence — dubbed as the next new engine for propelling technological revolution and industry transformation — took center stage in the Hong Kong Special Administrative Region’s 2025-26 Budget.

Financial Secretary Paul Chan Mo-po unveiled a new focus on making AI a core sector by concentrating on building supercomputing capabilities, sophisticated analytical models and robust data.

He proposed a HK$1-billion ($128-million) AI Research and Development Institute to ensure that innovation elements flow through the entire AI ecosystem. Last year, the SAR government set up a HK$3-billion Frontier Technology Research Support Scheme to solidify the city’s position as a leader in frontier AI technology research.

READ MORE: HK to drive AI as an emerging I&T industry

But, despite all the initiatives, there is still plenty of room to propel the SAR’s fundamental strength in the industry.

The local computing infrastructure has been relatively underdeveloped compared to neighboring cities. Even with the start of the first-phase facility of Cyberport’s Artificial Intelligence Supercomputing Center — one of the Asia-Pacific region’s large-scale supercomputing facilities — it only generates a computing power of 3,000 petaflops, or equivalent to processing nearly 10 billion images an hour — still far below the requirement of at least 15,000 petaflops of computing capacity by 2030.

Hong Kong is caught in a technology war between the Chinese mainland and the United States that may severely hamper its potential role in AI fundamental research capability in indigenous chip-making. The mainland has made significant strides in developing domestic AI chips through companies like Huawei with its Ascend chips, and Cambricon with its Siyuan chips.

The SAR’s AI research has, primarily, concentrated on AI applications rather than fundamental work on algorithms, models and computing infrastructure. This critical gap must be addressed to enable the city to develop truly novel AI technologies that can make it stand out in the international AI competition landscape.

Hong Kong’s high-cost structure in land and supercomputing resources is deterring global AI companies from setting up vital facilities in the city. Coupled with high recruitment costs and a dearth of AI talent, these problems have hampered the building of a comprehensive AI ecosystem.

However, despite the fundamental defects in its AI industry, Hong Kong can still catch up. It would depend on how it can turn its weaknesses and threats into strengths and opportunities.

AI industry practitioners, lawmakers, scholars and think tank researchers have laid down recipes to turn it around.

Cementing the upstream, midstream and downstream segments of an entire AI industry value chain, building a comprehensive AI ecosystem comprising capital, enterprises and talent; and forging stronger collaboration and integration between Hong Kong and other cities in the Guangdong-Hong Kong-Macao Greater Bay Area are the key ingredients.

Structuring the AI value chain

The upstream, midstream and downstream segments of the entire AI value chain refers to each of the elements required. The upstream sector concerns computing infrastructure, data center construction and chip development.

Eunice Yung Hoi-yan, a lawmaker and a barrister, says Hong Kong needs preferential policies to get private enterprises to invest in its supercomputing and data center infrastructure.

“Preferential policies should cover tax incentives for computing infrastructure investments, accelerated depreciation schedules for AI hardware, subsidized electricity rates for data centers (power consumption is a major operational cost for data centers), and streamlined approval processes for facility construction,” she suggests.

The government can create a special zone for data center development by providing reliable power supply, cooling infrastructure and high-bandwidth connectivity. It should consider building a coordinated computing resource network that integrates public and private facilities to boost AI competitiveness.

In chip development, local universities can leverage their significant expertise in materials science, electronic engineering, and computer architecture for semiconductor development. Research and development in specialized AI chips should be prioritized for optimizing specific applications rather than directly with established players in general-purpose chips. By providing specialized chips for targeted applications in financial services, healthcare diagnostics, and logistics optimization, Hong Kong can create a competitive edge in specific market segments.

The two pilot lines at the HK$2.8-billion Hong Kong Microelectronics Research and Development Institute will begin operating at the Microelectronics Centre in Yuen Long InnoPark next year, vying to sharpen Hong Kong’s edge in proprietary AI technology by advancing pilot production of silicon carbide and gallium nitride related to third-generation semiconductor.  

The midstream of the AI industry value chain refers to the foundational research in algorithms, models and computing infrastructure. “The Frontier Technology Research Support Scheme and the Hong Kong Generative AI Research and Development Center can address this gap by allowing researchers to develop novel technologies without the pressure of short-term commercialization,” says Yung.

The lawmaker is calling for the adoption of a vertical sector-specific model to foster the downstream segment of the AI industry value chain. “By targeting Hong Kong’s traditional strengths in financial services, healthcare, logistics, education and legal services to deliver more immediate value, rather than just building a general-purpose system, it can incorporate domain-specific knowledge and address well-defined business problems.”

Expanding AI applications

Earlier this year, Hong Kong launched its first ever generative AI large language model — the HKGAI V1 — heralding a new chapter in AI advancement. Developed by the Hong Kong Generative AI Research and Development Center under the government’s InnoHK innovation program, the HKGAI V1 model will be harnessed to provide a chatbox service “HKChat” — an AI chatbot that integrates local data and knowledge bases, and supports the use of Cantonese, English and Mandarin — for Hong Kong residents.    

Besides “HKChat”, the HKGAI is developing a series of open-source foundational large language models to create technologies for other applications, such as in medical diagnosis, legal affairs and environmental protection.

“The HKGAI V1 model can enable Hong Kong to create a ripple effect across industries, and foster widespread integration and innovation,” says Kenny Shui Chi-wai, vice-president of Our Hong Kong Foundation and executive director of the think tank’s Public Policy Institute.

“Hong Kong should accelerate AI applications by integrating AI capabilities into public services, drawing inspiration from Shenzhen, by integrating cloud with the DeepSeek-R1 model. Tax credits and public-private partnerships can drive AI adoption in key sectors like finance, logistics, and healthcare in which Hong Kong has a competitive edge,” Shui tells China Daily.

Neil Tan, chairman of the Artificial Intelligence Association of Hong Kong, says he believes Hong Kong should focus on AI technology applications in finance, logistics and property, given their significance in the local economy.

“AI’s advancement relies heavily on commercialization and use cases, with many corporations and institutions seeking opportunities for development and adoption of AI in Hong Kong,” he says.

An AI research report conducted by the Hong Kong Productivity Council and HKU Business School emphasized that the government can leverage the smart city objective, the economy’s reindustrialization and the upgrading and transformation of traditional industries to promote more large-scale use scenarios for AI technology.

Building an industry ecosystem

Another recipe is to cement a comprehensive AI industry ecosystem comprising capital, enterprise and talent.

“Hong Kong needs to attract more AI companies to set up shop here, nurture more home-grown AI talent by investing in AI research and development and develop a better incubation program,” says Jack Jiang Zhenhui, a professor of Innovation and Information Management and a Padma and Hari Harilela professor in Strategic Information Management at HKU Business School.        

An efficient AI industry ecosystem, he explains, can lure more AI companies to settle down in Hong Kong, leveraging the financial hub’s geographical advantage in bridging the mainland and international markets.

Tan also calls for efforts to bridge the AI talent gap. “As a new and evolving field, many candidates lack the necessary qualifications. The high demand for AI experts may lead to unqualified individuals holding prestigious titles, while qualified practitioners often do not have specific degrees in AI, despite their extensive knowledge.

“We can bridge the gap by clarifying what constitutes AI expertise while safeguarding local talent amid an influx of overseas professionals,” he adds.

Shui says the government could introduce tax incentives, streamline visa processes for global AI talent, host global AI events, and create a vibrant AI startup culture that can help raise venture capital funding to support early-stage AI startups.

Enhancing GBA collaboration

Forging stronger collaboration in the Greater Bay Area is vital for Hong Kong to stay ahead in the AI field.

“Developing a cohesive integration strategy with the mainland’s AI sector is critical, necessitating greater collaboration between Hong Kong and other Greater Bay Area cities in data sharing, algorithm/model development, talent acquisition, and computing resources,” Tan tells China Daily.

To achieve its objectives, Hong Kong should strengthen integration in the Greater Bay Area in several domains — cross-boundary data sharing, establishing data centers on the mainland, integrating generative AI models, leveraging the city-cluster area’s niches in use cases and supply-chain capabilities, utilizing the cross-border regulatory sandbox, and playing the standard-setting role.

According to Tan, Hong Kong’s advantages include its access to mainland and international data, while setting up data centers north of the boundary allows AI industry players to benefit from lower computing costs in the Greater Bay Area.

“As Hong Kong develops its generative AI models, integration with those from the mainland is essential, otherwise, it would risk creating barriers to collective AI advancement,” he says.

Jiang tells China Daily that the mainland’s edge in AI development lies in its abundance of application scenarios. “Although many AI technologies originate from the United States, it is the Chinese mainland that provides many diversified use cases of the technology.”

Since the Greater Bay Area has a comprehensive supply chain, the 11-city cluster can help Hong Kong enhance its manufacturing capabilities, enabling AI entrepreneurs to create more use cases of AI technology in the city.

ALSO READ: Chan: Hong Kong well-equipped to develop AI

Shui says the SAR can contribute to three fronts in AI collaboration in the region. The robust legal and financial frameworks of the city’s technology sandbox can provide a secure and compliant setting for testing cutting-edge AI innovation technologies.

Hong Kong can also take a leadership role in setting and harmonizing technical standards for AI across the Greater Bay Area by actively participating in global standard-setting bodies, such as the International Organization for Standardization.

Hong Kong can also contribute to cross-boundary data collaboration and capital linkage, with the latter attracting substantial investments from global venture capital firms to invest in AI startups.

However, although Hong Kong could implement these ideas, it lacks a coordinated vision to promote the AI industry. “Hong Kong does not have a comprehensive AI development blueprint, and this has led to the sector’s fragmented development,” Yung tells China Daily.

The blueprint should outline specific sector priorities, resource allocations, regulatory frameworks, and development timelines, creating a roadmap for enterprises to align their strategies with Hong Kong’s.

Thus, it can establish a coherent vision that can offer clarity and make companies more confident in investing in Hong Kong’s AI sector, she reckons.

Contact the writer at oswald@chinadailyhk.com