Can Singapore be neutral ground amid US-China AI rivalry?

The arrival of US and Chinese AI giants could strengthen Singapore’s technological capabilities and business ecosystem, while testing its ability to remain a trusted environment for both sides. SMU academic Ted Tschang gives his analysis.

17 Nov 2026

Technology

Anthropic logo and stock graphs in this illustration taken on 26 October 2026.
Anthropic logo and stock graphs in this illustration taken on 26 October 2026. (Dado Ruvic/Reuters)

In the past couple of years, Singapore has become the destination for some of the world’s most advanced and interesting AI companies, including OpenAI, Tencent, Alibaba, Manus and others. Anthropic was one of the latest US firms to announce a Singapore presence, officially opening its office last month. This has created some buzz, but what does it all mean for Singapore’s artificial intelligence (AI) sector?

First, we should clarify how this “emerging AI hub” differs from others. Singapore has already become a hub for many industries, and while hub creation has actually been policy, the notion of a hub varies from industry to industry. For many service industries including business software, this typically involves front office personnel such as marketing and sales operations. But AI is different: it is at once very technical, as well as needing a lot of situating of its “services” to customers. 

Difference of an AI hub

AI involves algorithms embedded in software and hardware, where the hardware is usually located outside the firm, such as cloud providers and their servers. Most advanced AI firms like Anthropic and OpenAI house their core development activity in their home country. They originally focused more on hiring technical personnel such as computer scientists, engineers and mathematicians, but as they grow, they have to transform and add to their capabilities. While both the Chinese and American companies appear to be hiring technical talent in Singapore, they may be customising their AI models, and will need to work with customers to create value in the customers’ downstream application areas. 

How they do so may differ, depending on the nature of the AI model. Anthropic and OpenAI want to capture the entire value through their proprietary model. Tencent and Alibaba also release pretrained models with “open weights”, allowing customers to run and customise the models independently. Customers can therefore develop more of their own AI capabilities and retain greater control over the value they create. 

Anthropic signage at the JPMorgan Chase & Co. headquarters in New York, US, on 13 November 2026.
Anthropic signage at the JPMorgan Chase & Co. headquarters in New York, US, on 13 November 2026. (Michael Nagle/Bloomberg)

Anthropic is racing towards its initial public offering (IPO), and the US AI stock market seems to be looking forward to that event. Anthropic’s hiring policies also provide insight in terms of how model development (done in the US) proceeds towards regional deployment. Some of its actions and issues are generally similar to those of other firms, while others may be specific to Anthropic. 

Anthropic’s business-to-business model

As a result of Anthropic’s technical emphasis, their product appears to have developed ahead of their business strategies, and they so far have handled some market segments at a distance. While their most diverse market is that of the consumer market, firms like Anthropic and OpenAI tend to rely on a technically sophisticated “Frontier” AI model delivered from their home location (in their case, the US). 

However, they are expecting most of their revenue to come from business itself. For this business-to-business model, in the last couple of years, Anthropic has been building more business-facing engineering capabilities. It appears that a Singapore office is another foray into the business side of their market as they face international markets. 

Augmenting Singapore’s capabilities as an AI adopter and user

They could be looking at how to secure governmental and sectoral interest in their core models. Anthropic plans to locate an initial contingent of “forward-deployed engineers” (FDEs) here. These engineers help companies deploy AI, at least in the initial stages. Think of them as technical consultants helping companies to find and prove out the uses of AI. This matches what OpenAI has done, and has implications for Singapore’s development as an AI hub. It enhances Singapore’s capabilities as an AI adopter and user. 

The same could be inferred from the Chinese firms, such as Tencent and Alibaba. Anthropic primarily commercialises its proprietary models, while Tencent and Alibaba also release pretrained models with “open weights”, allowing customers to run, customise and deploy them independently, subject to licensing conditions. This can give customers greater control over their AI applications and the value they create. 

Co-founder and CEO of Anthropic Dario Amodei (centre) looks on as US Speaker of the House Mike Johnson speaks to the press after a meeting with technology executives about artificial intelligence at the White House in Washington, DC, on 29 September 2026.
Co-founder and CEO of Anthropic Dario Amodei (centre) looks on as US Speaker of the House Mike Johnson speaks to the press after a meeting with technology executives about artificial intelligence at the White House in Washington, DC, on 29 September 2026. (Kent Nishimura/AFP)

The Anthropic/OpenAI FDEs can be particularly effective for working with multinationals that have their core operational and developmental work in Singapore, such as our own local corporations. AI customers have to have significant enough development work being done here to be able to absorb and be responsible for the AI rollouts. 

For instance, with Anthropic’s Claude or OpenAI’s models, the customer has to define what internal processes should change, how employees have to do things differently, and so on. It may need Anthropic's or OpenAI's help, but the customer may also develop its own capabilities. Something similar may occur if the customer uses the Chinese firms’ open-weight models. In general, all of this helps Singapore develop a capacity for adjusting and using AI, and getting the most out of it.

Get the ThinkChina Weekly Newsletter

Insights on China, right in your mailbox. Sign up now.

By subscribing, I agree to SPH Media's Terms and Conditions and Privacy Policy.

The AI hub that Singapore can be

All of this leads to an interesting competition. Will business and government customers use a premier model such as Anthropic’s, or will they adopt open-weight models and heavily customise their internal use? Anthropic’s presence may be to signal “hey, we’re here, and here to help”. Chinese firms are also here to expand their reach, because they may have to service international firms on their own turf, and may not be able to reach them just operating out of China. It is not like selling manufactured goods. 

What is in it for Singapore out of all this? A regional outpost such as Anthropic’s could undertake various functions, which could draw on being in a “centre” of such multinational activity, as well as contribute to it. Staffed with the right people, a regional office could be a listening post to help the company understand the varying perceptions and reactions of different countries and regions to AI. Silicon Valley firms are often in their own bubble, and could learn from being in international business environments. Singapore may be an important experience for such companies. 

One topic has to do with local or regional knowledge. There have been critiques that AI models are trained largely on a Western-influenced or English-language internet. Could a firm use its location to improve its model by training it on regional data? While regional languages and cultures are certainly the next step for model developers, a large part of the business world has fairly uniform processes, and the extent of adaptation varies. For some, it may simply involve adapting to local languages. Advancing model development along different dimensions is another possibility.

Singapore has strong AI research capabilities in its universities and public research institutes, which could benefit from proximity to foundation model development, whether through technical collaboration or the tacit exchange of knowledge. But again, this may require the firms’ model development to take place here. Personally, I’m looking forward to seeing how each type of firm approaches technical collaboration with local entities.

Another benefit for Singapore is the tacit understanding of how AI will shape work and business. Interactions between Anthropic and Singapore’s intellectual resources could help deepen our understanding of AI’s impact on work. This is because Anthropic has done more than many other firms to study how its AI affects work. Its internal research is still rather basic, but it is now funding academic work on how AI is changing the nature of work.

Skyline of Singapore Central Business District (CBD) and Marina Bay Financial Centre as seen from Marina Bay Waterfront Promenade on 10 November 2026.
Skyline of Singapore Central Business District (CBD) and Marina Bay Financial Centre as seen from Marina Bay Waterfront Promenade on 10 November 2026. (SPH Media)

There are many ways in which AI could shape Singapore’s broader industrial ecosystem. It could help establish Singapore as a thought leader, but as discussed, there are many other developments to look out for, some of which are not yet clear. Still, it is encouraging and exciting to see Singapore becoming one of the starting points for this growing technology.

Global milieu

In general, Singapore’s capabilities as a hub for further developing and using AI will expand. There are many more ways in which AI can be introduced into businesses, and many local businesses are already doing so, some using their own internal AI models. Having diverse types of firms operating here also makes the economy as a whole more resilient.

Singapore’s emphasis on good governance gives it an advantage. As competition between the US and China intensifies, it may seem counterintuitive for firms from both countries to be located here. Yet Singapore could become a trusted environment for AI firms from both countries to develop and deploy their technologies. They can learn from each other indirectly, even without direct collaboration, through channels such as their customers. Singapore’s ability to attract firms from both sides while maintaining their trust could therefore become an important strength as an AI hub.

ThinkChina is an English language e-magazine with a China focus and powered by Singapore Press Holdings' flagship Chinese daily Lianhe Zaobao.

Get the ThinkChina Weekly Newsletter

Insights on China, right in your mailbox. Sign up now.

By subscribing, I agree to SPH Media's Terms and Conditions and Privacy Policy.

Copyright © 2026 SPH Media Limited. Co. Regn. No. 202120748H. All rights reserved.