In a two-horse AI race, not hitching a ride is the way forward

With artificial intelligence (AI) becoming a two-horse race between the US and China, how can smaller countries survive, develop independent capabilities and stay relevant in a world driven by AI?

They can follow Singapore’s strategy: being smart, agile and playing to their own strengths.

AI is the single biggest technological leap for humankind since the invention of the steam engine in 1769, which ushered in the first industrial revolution.

Frontier AI models are the software equivalent of the steam engine, and they are the drivers of the fourth industrial revolution.

While the AI models are nothing but clever code, the plumbing that makes them work is very much physical infrastructure in the form of powerful AI chips and energy- and water-guzzling data centres.

For smaller countries, writing the code is hard but not impossible. What is harder is replicating the physical infrastructure, and this acts as a barrier to entry.

The world’s most powerful AI-enabled chips are made by the likes of Nvidia and a handful of other companies based in the US. They are in short supply, and the US has strict export controls on who can buy them.

Then, there are mammoth hyperscale data centres that host these frontier AI models.

Energy and resource guzzlers

Speaking at the Global Dialogue in AI Governance earlier in this month, UN Secretary-General António Guterres noted that by 2030, AI companies “could use more electricity than all but five nations – and enough water to meet the needs of all 1.3 billion people in sub-Saharan Africa for an entire year”.

He was citing the UN AI Environmental Transparency Initiative report.

The hard physical infrastructure, talent and money required to produce frontier AI models have ensured that the race for dominance in this field has become a race between the two largest economies in the world, China and the US.

After wowing the world earlier with the DeepSeek AI model, China, on July 17, released Kimi ​K3, a 2.8 trillion-parameter model that has been claimed to be the world’s largest open-weight AI system.

Analysts say the new frontier AI model delivers performance approaching the best models the US has on offer.

Meanwhile, even as the world has been trying to understand the implications of the dangers Mythos could pose to cybersecurity, the US government issued a directive denying all non-Americans, even those located in the US, access to both Mythos 5 and Anthropic’s even more powerful Fable 5 frontier AI model.

Worried about the off switch

Unfortunately, unlike steam engines of yore, frontier AI models come with an off switch, which the US President, Donald Trump, has a record of pressing, albeit "temporarily".

This, naturally, has sent jitters around the world.

Another fear is the potential backdoors for spyware in AI models, putting both companies and countries at risk.

This is why the call for sovereign AI models has gained traction over the past couple of years, and many countries are racing to build their own.

Germany, for example, has launched Soofi AI, an open-source AI model which has garnered good reviews. India has released its own frontier AI model, Sarvam AI and is working on a slew of others.

These are but two examples of several as countries race to protect their data and build models they can trust and have full control over.

Singapore has developed the highly regarded SEA-LION large language model (LLM), designed to understand Southeast Asia’s diverse languages, cultures, and contexts.

Despite this success, it has realised that trying to compete on raw compute power and economic muscle with larger countries would just be a distraction - and has pivoted its AI strategy.

Speaking at the International Scientific Exchange for AI Safety 2026 in May this year, Singapore’s Minister for Digital Development and Information, Josephine Teo, categorically noted that the country “will not, and cannot, try to outspend the world’s biggest spenders on AI. We do not have such deep pockets.”

Instead, the country has approached the problem in a way that can be a lesson for other middle powers who do not have the resources to compete with the big boys in developing their own AI models.

Teo noted that Singapore has been focusing on creating an environment of trust – where AI was deployed responsibly, risks were well understood, and effective protections were thoughtfully implemented.

The government has positioned Singapore as a highly efficient, agile and living laboratory for AI deployment, testing, and scaling AI solutions.

Earlier this month, Teo said, during the Global Dialogue in AI Governance, that Singapore’s position is that it believes that AI can be a significant force for good.

Realising the full potential of AI required “patient, sustained effort”, which included building public infrastructure, developing capabilities, and putting governance frameworks in place, she noted.

Moving from expertise to adaptability

Senior Minister of State for Digital Development and Information, Tan Kiat How, noted in a recent speech that AI is shifting competitive advantage from expertise to adaptability. 

As AI makes knowledge increasingly accessible, the real differentiator is no longer who has the most advanced AI, but an organisation's or country’s ability to continually rethink its business, redesign how work is done, and create new value for customers.

Singapore’s advantage lies in its trusted ecosystem, instead of competing in the global AI race on raw compute power. So, it made sense to position the country as the place where organisations can transform themselves with confidence.

Building Singapore’s AI ecosystem was ongoing work and required the collective work of the whole ecosystem, as Tan noted. 

The government, technology companies, enterprises, and professional services firms each have distinct roles to play in translating technological possibilities into lasting organisational change.

The country has identified four sectors as priorities for the AI push: connectivity and logistics, advanced manufacturing, healthcare, and finance.

Teo recently noted that Singapore already holds global standing in each of these industries, which together account for more than 40 per cent of the country’s gross domestic product.

By tightly weaving AI into public sector workflows, the government intended to serve as a model for responsible deployment, ensuring that public goods - from municipal services to border clearance - were significantly enhanced.

Global North with a Global South playbook

In an interesting paradox which works to its advantage, Singapore occupied a unique geopolitical and economic position.

In terms of capital, talent, and computational infrastructure, it easily fits into the profile of the "Global North". And yet its geographic and diplomatic roots and inclinations are closely intertwined with Southeast Asia and the broader "Global South."

This has given the country its unique position in the global AI race. Rather than adopting the monopolistic strategies of major tech powers or the rigid regulatory approach of the European Union (EU), Singapore has positioned itself as a neutral bridge.

Its strategy focuses on "collective stewardship", advocating that AI sovereignty should be accessible to all nations, not just to superpowers.

In summary, Singapore’s AI strategy has no grand design.

It rests on developing reliable infrastructure, credible data, trained manpower, particularly the civil service who lead by example, and collaboration with companies and like-minded countries.

These are easily replicable steps that can be followed by other smaller countries who do not want to be left out of the AI race but who at the same time want to have the independence to be able to do what’s good for their citizens.

Like in earlier technology eras, Singapore can and will remain a role model in this brave new world of AI.