Pacific Money

Singapore Is in Perfect Position to Court AI Companies From China

Recent Features

Pacific Money | Economy | Southeast Asia

Singapore Is in Perfect Position to Court AI Companies From China

With China facing advanced chip and equipment bans from the United States, Singapore is an increasingly attractive place for Asia-based talent to set up AI companies.

Singapore Is in Perfect Position to Court AI Companies From China
Credit: Depositphotos

Singapore is once again attracting talent and capital from the wider Asian region, especially from China, as artificial intelligence companies race to develop the next super app. The Lion City has in the last few years been a top destination for big tech firms to set up their Asia-Pacific headquarters as geopolitical tension heats up between the United States and China, and domestic regulations intensifies in China. Singapore is now a major hub for ByteDance, Google, Netflix, Shein, and others.

The Singapore Infocomm Media Development Authority (IMDA) announced a GenAI Sandbox program in February to support small and medium enterprises (SMEs) across industries in adopting AI and offering a grant to be part of a three-month trial run program. 

Tribe, a “startup ecosystem builder,” and Digital Industry Singapore, a joint office between the Economic Development Board, Enterprise Singapore, and the IMDA, also recently announced a collaboration with U.S. tech giant Nvidia in establishing an AI-focused accelerator program, showcasing multiple arms of the government bolstering the enhancement of the AI ecosystem in Singapore.

“Singapore is a good location for two reasons,” said Kelly An, co-founder of OpenCord.AI, a 24/7 on-demand social media manager. “First, it is easy to establish a company in Singapore, and for Chinese companies going global, Singapore allows more international and diverse exposure. Second, Singapore’s overall tech environment is relatively strong, with all the major companies having huge operations here. So it’s easier to build those networks and connections.”

When pro-democracy protests broke out in Hong Kong in 2019, Singapore saw an influx of capital, more specifically private equity money. Years later, it is a top choice for family offices and venture capitalists to set up shop. In comparison to the previous first choice, Hong Kong, Singapore boasts better tech entrepreneur visa rules, tax incentives, and often a lower cost of operations. 

Now, with China facing advanced chip and equipment bans from the United States, Singapore is once again the perfect place for Asia-based talent to set up AI companies, from software to hardware. Singapore checks all the boxes needed for a founder, as it offers a strong talent pool, a vibrant ecosystem, proximity to private equity and venture capitalist money, and a free and open internet.

A country of under 6 million people, Singapore made up 15 percent of Nvidia’s revenue in the October 2023 quarter, only trailing behind the United States (34.8 percent,) Taiwan (23.9 percent), and China including Hong Kong (22.2 percent). Whether the chips are used domestically for AI development or data centers, or they are being repackaged and shipped off elsewhere, Singapore has successfully positioned itself as one of the world’s leading AI players.

The concerted push to attract some of China’s most promising AI startups to relocate seems all too natural and a win-win for both the companies and for Singapore. Companies get access to international networks, public and private investments, and mature digital infrastructure, while Singapore gets another chance to court some of the brightest minds in the industry and leapfrog to obtain frontier technology. 

“Singapore is an attractive place for many founders to set up entities targeting the global market because it shares the same internet ecosystem as the rest of the world, has good internet connectivity, and serves as an easy-to-access travel hub so many Singaporean AI companies can easily adapt a global roadmap,” said Yichen Guo, CEO of CapGo.AI, a Singaporean AI productivity app for lead generation.

Many echoed his comment and said that the fact that the country can offer AI developers access to a free internet would be enough incentive to draw in Chinese talent. The issue with training large language models (LLMs) in China is that companies are limited to using data within the Great Firewall, due to both institutional censorship and self-censorship. 

Ultimately, LLM performance improves with scale – more parameters, more and better training data, more training runs, and more computation,” Goldman Sachs wrote in a recent note. There is simply less information available in a closed internet. So Chinese AI companies face an inherent disadvantage, as there is more data (in all languages, from all countries) to train on for companies outside of China. 

An, of OpenCord.AI, said that for Chinese companies going global, the United States or Europe could be intimidating while “Singapore feels like a safe harbor.” But companies setting up shop in Singapore are not problem free, Guo added. “The biggest challenge for AI applications in Singapore is that the market size of the region (Southeast Asia) is still relatively small and fragmented. Software businesses, in general, thrive better in developed countries with a higher willingness to pay.”

Metta Ni contributed to the research of this story.