Artificial intelligence is increasingly perceived as a competition intended for very large states. According to Stanford University's AI Index for 2026, private US investments in AI reached nearly $286 billion last year, more than 23 times the figure in China.
Training and launching advanced models requires enormous computational infrastructure, expensive chips, electricity, and a relatively small pool of highly skilled specialists. Nevertheless, two countries with populations and domestic markets significantly smaller than these two global powers have managed to position themselves unusually close to the center of this race.
The United Arab Emirates (UAE) is channeling significant capital into developing computational infrastructure, models, and international technology partnerships. Singapore, meanwhile, has spent years building a digital government, AI governance mechanisms, a research base, and one of the most concentrated pools of AI talent.
According to Spring Fu, Managing Director of the Mohammed bin Zayed University of Artificial Intelligence Academy, attempting to fully replicate the technology ecosystem of America or China is not a realistic goal. She argues that a medium-sized country cannot expect leadership in all areas but can leverage its size as an advantage in specific fields.
You Cannot Win Everywhere
For small states, this means accelerating government operations, concentrating resources on chosen capabilities, and becoming valuable partners within the AI ecosystem, which is not controlled by any single country.
The UAE's most notable bet is on computing power. Plans include creating the Stargate UAE computing cluster, with a capacity of one gigawatt, as part of a broader UAE-US AI campus totaling five gigawatts in Abu Dhabi. The project is being implemented by G42, with OpenAI and Oracle managing it, and Nvidia supplying the advanced chips. The first phase, with a capacity of 200 megawatts, is expected to go live this year.
Singapore faces a different physical reality. It already possesses over 1.4 gigawatts of data center capacity, but the island nation is constrained by land resources and a tropical climate that makes cooling energy-intensive. Data centers consume about 7 percent of national electricity, according to its Infocomm Media Development Authority. Instead of unlimited expansion, Singapore selects new capacity based on energy efficiency and strategic value.
This comparison demonstrates why there is no single formula for becoming an AI powerhouse. Abu Dhabi can use land, energy, capital, and geopolitical partnerships to create colossal computing infrastructure, whereas Singapore must derive maximum benefit from stricter physical limitations.
Both countries are heavily dependent on access to the outside world. Fu emphasizes that for a small state, international partnership is not an addition to national capability but an integral part of its construction, concerning access to chips, joint research, and attracting major international companies.
This dependency is not necessarily a weakness, as even the world's largest AI ecosystems rely on international supply chains for chips, manufacturing equipment, energy, models, and expertise. The question for small states is what capabilities they need to control independently and what they can reliably obtain through partners.
Fu believes that people are a more serious constraint. She notes that computing power can be bought or created, and capital can be quickly directed, but without people capable of utilizing this infrastructure, launching AI systems within organizations, and making informed decisions as regulators and buyers, investments remain unused.
Her definition of AI talent goes beyond a few elite researchers. The country needs not only these researchers but also many more engineers and practitioners capable of implementing technologies, as well as another group that receives less attention: leaders who understand AI well enough to decide where and where not to apply it.
Singapore offers one such strategy. Last year, it had the second-largest concentration of AI talent in the world, with 1.82 percent of LinkedIn users classified as AI talent in the Stanford index. Currently, the country aims to train 100,000 workers to become 'AI bilingual,' combining knowledge of their current profession with the ability to use AI, while helping 10,000 enterprises deepen adoption over the next three years.
The UAE is also moving quickly in this direction. Stanford found that the concentration of AI talent in the country grew by 121 percent between 2019 and 2025, the fastest growth among measured countries. The federal government is also collaborating with MBZUAI to develop expertise in agentic AI among 80,000 civil servants, including senior management and non-technical positions.
Fu's own program is a smaller example of this third level. Twelve UAE leaders spent five days in Singapore in September, studying the country's approach to AI policy, industry, governance, and talent, meeting with government bodies, multinational corporations, and local companies. The goal was not to train them in engineering but to enhance their decision-making capacity regarding the technology.
This exchange became part of a broader relationship. A few days later, the Abu Dhabi-Singapore Joint Forum identified AI and capacity building as areas where cooperation between the two parties would expand.
Perhaps the more complex question is how to determine if all this investment volume has yielded anything long-term. Fu notes that 'activity is easy to generate: announcements, pilots, startups, and signed conference partnerships. But seeing and imitating it is harder.'
Her criteria are more demanding: Is AI working within core government and industry operations, or is it remaining in innovation labs? Can the country develop and retain the necessary personnel? Are procurement rules, data systems, regulators, and testing standards mature enough for large-scale deployment? And will these capabilities withstand changes in technology, leadership, and funding?
Singapore is increasingly focusing on this less visible institutional level. Its National AI Strategy was updated this year, and the National AI Council was established in February, chaired by Prime Minister Lawrence Wong. Its AI governance work has also shifted towards agentic systems, updating frameworks this year based on feedback from over 60 organizations and examples of agents already operating within companies and governments.
Meanwhile, Abu Dhabi is actively integrating AI directly into the government. Last month, its Executive Council used an agentic AI platform for the first time to support analysis and discussion during a session chaired by Crown Prince Sheikh Khaled bin Mohammed bin Zayed Al Nahyan.
It is here that the argument about small states becomes more interesting than just counting data centers or investment announcements. Neither Singapore nor the UAE can compare with the United States in private capital or with China in industrial scale. Their bet is that they do not need to.
For small countries, AI potential may depend on carefully selecting what to build, developing sufficient people to use these technologies, having institutions move fast enough to implement them, and maintaining relationships that ensure access to everything else. A data center without talent is infrastructure. Talent without computation is potential. Partnerships without institutions to utilize them remain mere announcements. The hardest achievement is turning all three into real opportunities.
