Introduction

In recent years, there has been an observable increase in the level of global competition surrounding the development of AI, with the expectation that dominance in this field will result in significant economic and military advantages. The ongoing competition between the United States and China, sometimes characterized as an "AI race," exemplifies this dynamic. The two nations are engaged in intense rivalry, with the United States further tightening export controls for the most advanced AI chips in April 2025 (Mickle, 2025), before re-allowing Nvidia and AMD to export some chips to China in July 2025 (Mickle, 2025b). China is now requiring these chips to undergo additional security reviews (Huang & Whelan, 2025) and is urging local companies to train AI models with Chinese GPUs, with mixed success (Olcott & Wu, 2025). At the same time, both superpowers are trying to influence other states by exporting their respective AI stack worldwide. While the U.S.-China rivalry continues to dominate debates on artificial intelligence, a number of other countries are advancing ambitious AI strategies of their own. Through investments in R&D, digital infrastructures, and in some cases the training of regional large language models (LLMs), these states are seeking to shape, rather than simply adapt to, the emerging global AI order. Against this backdrop, we focus on the United Arab Emirates, India, and Singapore, three ambitious Asian players in the AI race that are neither firmly aligned with Washington nor Beijing. By analyzing their AI strategies and capacities, we explore under which conditions such actors can become consequential players in the global AI competition, and what lessons their strategies may hold for others.

United Arab Emirates: The Ambitious Investor

Background The United Arab Emirates (UAE), located on the shores of the Persian Gulf at the intersection of Asia, Africa and Europe, is a small but wealthy state whose prosperity has long rested on the export of hydrocarbons. As a founding member of the Gulf Cooperation Council (GCC), it collaborates closely with Saudi Arabia and other Gulf states on trade, energy and security, even as regional competition over investment, hub status and influence intensifies (Szalai, 2025). The UAE’s foreign policy is often characterized as a strategy of multi-alignment. It hosts U.S. forces and maintains a strong security partnership with Washington (Allen et al., 2025), while at the same time expanding economic and technological ties with China, cultivating relations with Europe and Russia, and signaling broader global ambitions through its accession to BRICS+ (UAE MOFA, 2023). Multi-alignment has helped the UAE position itself as a leading trade and logistics hub between Asia, Africa and Europe and one of the world’s top ten destinations for foreign direct investment (UNCTAD, 2024). Despite its success, the UAE’s economic model is still heavily dependent on the export of fossil fuels. Recognizing the limits of a hydrocarbon-based model, Emirati leaders have made diversification a strategic priority, with artificial intelligence emerging as one of the most prominent areas of investment and ambition.

Policy The country was an early mover in the field of AI, appointing the world’s first Minister of State for Artificial Intelligence in 2017 and releasing the National Artificial Intelligence Strategy 2031, which aims to make the UAE “one of the leading nations in AI by 2031” (UAE Government, 2023). To turn this vision into reality, Abu Dhabi established flagship institutions such as the Mohamed bin Zayed University of Artificial Intelligence and MGX, a state-backed fund designed to support investment in AI infrastructure, semiconductors, and startups (MGX Group, 2025). The UAE has also invested heavily in data center infrastructure, with Khazna (Khazna, 2025) and other domestic and international providers operating dozens of datacenters in the country. The UAE’s most prominent player, however, is G42, a national technology champion that has become the centerpiece of Abu Dhabi’s AI ecosystem (G42, 2025). After initial scrutiny over G42's ties with Chinese firms, the UAE aligned itself more closely with Washington in 2024 (Harris, 2024), helping G42 secure a $1.5 billion partnership with Microsoft (Microsoft, 2024; Patel et al., 2025). The deal reopened access to advanced U.S. chips and reassured American policymakers of Abu Dhabi’s alignment. Building on this rapprochement, President Trump’s state visit in May 2025 was used to announce Stargate UAE, a flagship supercomputing project led by G42 in collaboration with OpenAI and other U.S. partners (MOFA UAE, 2025). Promoted as one of the world’s largest AI compute facilities, Stargate is planned as a multi-phase project with an initial 1 GW capacity, scaling to 5 GW over time. If the project goes ahead as planned, the world's largest AI training datacenter by 2030 could be located in the gulf state. For the UAE, these initiatives guarantee access to advanced hardware, foreign expertise, and partnerships with the world’s leading firms. For Washington, they reinforce the integration of Gulf AI development into the U.S. technological sphere while narrowing the space for Chinese influence in the region (Allen et al., 2025).

Analysis Although these initiatives demonstrate ambition on a scale that few others in the region can rival, structural constraints weigh heavily on the UAE’s AI strategy. AI-intensive data centers consume vast amounts of electricity and water for cooling, a challenge magnified in a desert state that is already among the most water-stressed in the world with summer temperatures up to 50 °C (Bhat, 2025). Power generation remains dominated by natural gas (~71% in 2023) while water is overwhelmingly produced through fossil fuel-based desalination (IEA, 2025). Large-scale AI infrastructure build-out remains tied to carbon-heavy inputs and risks being at odds with the UAE's sustainability and green diversification agenda (Bhat, 2025).

Beyond these physical limits, the UAE’s AI ecosystem remains heavily dependent on foreign partners. Access to advanced chips is conditioned by U.S. export controls, software and know-how are concentrated in American firms, and even flagship projects like Stargate are structured around U.S. technology providers, with reports suggesting that as much as 80 percent of its compute capacity will be reserved for U.S. partners (Patel et al., 2025). Despite high levels of digital literacy and major investment, the UAE’s AI ecosystem remains heavily dependent on expatriate expertise, a reliance that limits its ability to achieve genuine technological sovereignty (Allen et al., 2025). As Singh (2025) notes, genuine sovereign AI requires control over compute, data, talent, and infrastructure, dimensions in which the UAE remains dependent on external actors, highlighting the gap between its global ambitions and its domestic capabilities.

The UAE has positioned itself as a Gulf leader in AI, using wealth and strategic alignment to build advanced AI infrastructure, attract major partners and launch flagship projects like Stargate. These efforts give it visibility and influence well beyond its size and position it as a key player in global AI. Yet, questions remain about long-term talent retention, reliance on foreign chips and the sustainability of its resource-intensive infrastructure. Whether the UAE can progress fro`m serving primarily as a host and platform for global AI to achieving genuine technological sovereignty remains to be seen.

India: The Talent Powerhouse

Background India is often described as an awakening giant in the field of international relations. The country pursues a policy of strategic autonomy, often described as an updated evolution of its non-alignment tradition, which emphasizes independence of action while enabling tactical partnerships across rival power centers (Suri, 2025; Unger & McLean, 2025). This allows India to balance relations with competing blocs as a member of BRICS alongside China and Russia, while also deepening security and strategic cooperation with the U.S., Japan, and Australia through the Quadrilateral Security Dialogue (Quad). India remains a country of contrasts. Whilst it is a major technology hub with a rapidly expanding digital economy and one of the world's largest consumer markets, its growth remains uneven, with persistent disparities in physical and digital infrastructure and governance capacity that complicate its path toward sustained global leadership (Suri, 2025; Unger & McLean, 2025). At the same time, India is the world’s most populous country with a youthful demographic profile and one of the largest annual cohorts of STEM graduates, factors that reinforce its reputation as a global ‘talent powerhouse (MacroPolo, 2025; Suri, 2025; Unger & McLean, 2025).

Policy Recognizing both the economic opportunities and the strategic importance of technological sovereignty, the Indian government launched the IndiaAI Mission in 2024 to build a national “AI stack” spanning compute, data, talent, research and development (R&D), algorithms, applications, and capital (IndiaAI, 2024; PIB, 2025; Suri, 2025). As part of this mission, the government allocated over $1 billion (\₹10,300 crore) in 2024 for a five-year period to strengthen AI capabilities and to invest in AI compute and semiconductor infrastructure (PIB, 2025). The mission is anchored by the Ministry of Electronics and Information Technology and includes flagship initiatives such as AIRAWAT, a national AI supercomputing platform, and AIKosha, a centralized dataset and model repository to support researchers and startups (Airawat Research Foundation, 2025; PIB, 2025). It also prioritizes the development of domestic LLMs tailored to India’s linguistic diversity and economic needs, with research concentrated at the Indian Institutes of Technology and the Indian Institute of Science in Bangalore (Suri, 2025). At the industrial level, India has tied its AI ambitions to its broader push for high-tech manufacturing, offering subsidies to attract semiconductor investment. This strategy has already drawn major commitments, including Micron’s nearly $3 billion chip assembly and testing facility in Gujarat, alongside other projects backed by Foxconn (Micron, 2023; Shivakumar & Sharma, 2025). In parallel, India’s data centre sector is expanding rapidly: from just over 1.5 GW of capacity in 2024, it is projected to more than double by 2030 (Garg, 2025). Internationally, India has made AI cooperation a feature of its broader technology diplomacy, advancing work within the Quad and the Global Partnership on AI, and during its G20 presidency in 2023 it promoted “digital public infrastructure” as a model for inclusive AI governance (Suri, 2025; Unger & McLean, 2025). The 2025 U.S.–India TRUST initiative has further embedded AI and semiconductor cooperation into the bilateral relationship, described by analysts as the creation of a “trusted technology corridor” between the two countries (Shivakumar & Sharma, 2025; Suri, 2025b).

Analysis As Singh (2025a) notes, these efforts reflect India’s pursuit of “sovereign AI”, driven by “indigenised ambitions” to develop advanced models using domestic infrastructure, datasets, and talent. Yet ambition alone will not guarantee autonomy. Despite having one of the world’s largest pools of STEM graduates and a rapidly growing base of AI professionals, India struggles with retaining top-tier talent, as many of its best researchers and engineers migrate to leading AI labs abroad (MacroPolo, 2025). Combined with limited domestic datasets and underinvestment in R&D, these dynamics continue to slow progress (Suri, 2025a). Infrastructure is another concern: while India’s data centre capacity is rising quickly, analysts highlight that this expansion brings heavy demands for electricity and water, exposing vulnerabilities in an energy system still dominated by fossil fuels (Seth & Kushwaha, 2025).

Still, India’s growing role in global semiconductor supply chains, driven in part by U.S. efforts to diversify away from China, and its influence as a middle power give it leverage that few other developing nations possess (Singh, 2025a; Unger & McLean, 2025). At the same time, new initiatives such as the U.S.–India TRUST roadmap show how strategic autonomy increasingly coexists with selective alignment with Washington, reflecting both India’s dependence on foreign compute and its willingness to integrate AI cooperation into a broader partnership (Suri, 2025b). With better alignment of investment, talent, and industrial capacity, India could evolve from an ambitious adopter to a genuine player in the global AI race, but whether it can translate ambition into sovereignty remains an open question (Singh, 2025a; Suri, 2025a).

Singapore: The Hedging City-state

Background Strategically located at the southern entrance to the Strait of Malacca, Singapore is a wealthy and highly innovative city-state that serves as a major trading and logistics hub, as well as Southeast Asia’s leading financial centre. Despite its small size, the multiethnic nation maintains deep ties with both Beijing and Washington. In this unique position, Singapore has traditionally adopted a foreign policy of “not choosing sides,” using hedging as a pragmatic tool to balance between the two superpowers while preserving its own national interests (Chong, 2023).

Policy Given the intensifying U.S.-China rivalry and the transformative potential and risks of AI, Singapore has made AI a core pillar of its national strategy. In 2017, the government launched AI Singapore to develop national capabilities, cultivate talent, and encourage widespread industry adoption (AISG, 2017). These efforts were complemented by the National AI Strategy, introduced in 2019 and upgraded to NAIS 2.0 (2023). Rather than competing with leading AI labs to develop the most advanced models, Singapore aims to integrate itself into the global AI value chain by prioritizing adoption, infrastructure, and trusted governance (Smart Nation Singapore, 2023). To support its strategic ambitions, the government has allocated substantial resources for the procurement of AI chips and collaborates closely with universities and industry partners (AWS, 2024; National Research Foundation Singapore, 2021; Tech.Gov.SG, 2023).

Analysis From an infrastructure and software standpoint, Singapore is well positioned to achieve these goals. Backed by a dense cluster of data centers with capacity well above 1 GW and the ASPIRE 2A+ supercomputer powered by 320 NVIDIA H100 GPUs, Singapore has built one of the strongest computing environments in the region (Mayer Brown LLP, 2024; National Supercomputing Centre Singapore, 2023). The city-state is also a major landing point for subsea cables that carry most of the world's internet traffic and provide the connectivity on which data centers, cloud services, and AI adoption depend on. In line with the government’s AI adoption strategy, Singapore is developing LLMs for Southeast Asian languages and use cases. Notable examples include AI Singapore’s SEA-LION and A*STAR’s MERaLiON models (He et al., 2024; SEA-LION, 2025), the latter of which was developed in collaboration with Microsoft (A*STAR, 2025; Smart Nation Singapore, 2023).

Few other locations host American and Chinese technology companies on a comparable scale. Singapore has also succeeded in attracting substantial foreign investment. For instance, Google has pledged $5 billion for cloud and data center expansions, and AWS has committed $12 billion through 2028 (AWS, 2024; Reuters, 2024). Meanwhile, Chinese firms such as Alibaba Cloud and Huawei have significantly expanded their regional AI and cloud operations in the city-state. This unusual concentration of competing ecosystems reinforces Singapore’s position as the leading digital and AI hub in Southeast Asia and strengthens its credibility in playing a more active role in regional and global AI governance (Wu, 2025).

Alongside domestic initiatives aimed at fortifying its AI infrastructure and accelerating AI adoption, Singapore has emerged as a prominent actor in the realm of AI governance, operating at the regional, plurilateral, and multilateral levels. Within the framework of ASEAN, it spearheaded the development of the bloc's Guide on AI Governance and Ethics (ASEAN, 2024). Furthermore, the country has augmented its collaborative endeavors with other blocs such as the EU, notably through the Singapore-EU Digital Partnership (European Commission, 2023). Singapore's international engagement also encompasses participation in the OECD and contribution to the Global Partnership on AI (OECD, 2025a, 2025b). The country has also promoted frameworks such as the Model AI Governance Framework and the AI Verify toolkit to encourage international uptake and interoperability (AI Verify Foundation, 2022). At the same time, Singapore engages in diplomatic discourse with both Washington and Beijing, positioning itself as an impartial forum where contending powers can engage (Cuyegkeng, 2023). This strategy reflects the logic of a small state: unable to compete on scale, Singapore embeds itself in governance networks to amplify its influence, shape emerging norms, and reduce the politicization of AI in an era defined by U.S.-China rivalry (Wu, 2025).

Despite its advantageous position in the global competition for AI, Singapore is confronted with substantial structural impediments. As a strategic hedger, Singapore faces constrained margins of maneuver as the technological rivalry between the United States and China intensifies. This is most visible in hardware, where U.S. export controls and allegations of GPU diversions from Singapore to China have raised the risk of restricted access to advanced chips (Mandell et al., 2025). Land scarcity is another factor contributing to these pressures, as the city-state's compact geography leaves limited industrial space for data-center expansion (Ng et al., 2023). The issue is further exacerbated by energy and electricity constraints. Data centers already account for approximately 7% of the nation's electricity demand (Marzouk, 2022) and given that more than 90% of Singapore's power is generated from imported natural gas, each increase in demand results in higher costs, emissions, and supply risks (Malik, 2024). In an effort to address these pressures, the government implemented a moratorium on the construction of new data centers in 2019, which was partially relaxed again in 2022. Finally, talent remains a bottleneck: surveys indicate that nearly half of firms in Singapore encounter difficulties in hiring AI professionals (Goode et al., 2023).

Conclusion

The experiences of the UAE, India, and Singapore show that smaller states and middle powers can carve out meaningful roles in the global AI race, even as the field remains dominated by the United States and China. Each has pursued a distinct strategy shaped by national strengths and constraints.

The UAE has positioned itself as a Gulf hub for AI through large-scale investment and high-profile partnerships. Yet its reliance on U.S. expertise and chips limits autonomy, while massive infrastructure projects raise questions about electricity demand, water use, and long-term sustainability.

India, by contrast, has the scale, market size, and talent to pursue a more independent path. It is building models tailored to its linguistic diversity and vast consumer base. However, persistent brain drain, underinvestment in R&D, and limited access to advanced semiconductor technology continue to slow its momentum.

Singapore takes a different approach and continues to hedge between Washington and Beijing. It focuses on adoption and governance, leveraging world-class infrastructure to position itself as a trusted hub where both U.S. and Chinese firms operate. Yet land scarcity, rising energy needs, and dependence on U.S. chips expose its vulnerabilities, prompting tighter export controls.

Taken together, these cases underscore several lessons. Infrastructure and capital alone do not ensure technological autonomy. Retaining and developing domestic talent is just as essential. Governance and adoption can provide middle powers with meaningful influence even without control over cutting-edge hardware. At the same time, access to advanced chips and the constraints of land, energy, water, and sustainability remain a decisive bottleneck.

Middle powers cannot escape these structural and geopolitical limits. By aligning strategies with their national strengths, they can still play significant roles in shaping the global AI order.

References