When Singapore-based firms clash with Chinese regulators

When fast-fashion giant Shein moved its headquarters to Singapore in 2022, the shift was widely seen as an attempt to distance the company from rising US-China tensions.

The group presented itself as Singapore-based as it prepared for stock market listings in New York and later London. But those plans stalled after it failed to secure approval from Chinese securities regulators.

More recently, founder Xu Yangtian reaffirmed the company’s Chinese roots when Shein applied to list in Hong Kong – a reminder that moving headquarters does not break regulatory ties.

That tension is becoming clearer as Chinese companies, especially tech groups, reshape their global structures to manage the rivalry between the US and China. Singapore has become a favoured base thanks to its stable regulation and reputation as a neutral hub.

But recent scrutiny of an artificial-intelligence deal suggests relocation alone may not shield companies from geopolitical pressure.

The US$2 billion (S$2.6 billion) purchase of Singapore-based AI group Manus by US tech giant Meta in January raised concerns in Beijing that the deal may have sidestepped China’s technology export controls.

Other Chinese groups have made similar moves. ByteDance, owner of TikTok, has hired thousands of staff in Singapore, while pharmaceuticals company WuXi Biologics is building a US$1.4 billion research and development hub in the country.

The lesson is straightforward: relocation may reduce political exposure, but it rarely removes it. Companies such as Manus and Shein remain subject to Chinese oversight even after moving their headquarters to Singapore. Chinese rules claim jurisdiction based on where a company’s assets and operations originate, not simply where it is incorporated.

Relocation, in other words, does little to reduce a firm’s exposure to China’s technology export controls. Or to the broader contest between the US and China over advanced technologies.

The long arm of export controls

Manus provides a use case to understand the thinking of Chinese regulators. The company moved its headquarters to Singapore before its sale to Meta in an effort to reduce not only the geopolitical baggage that comes with being based in China but also possibly to sidestep Chinese laws that curb cooperation with US investors.

Beijing’s review focuses on whether Manus should have secured an export licence before moving staff and technology to Singapore – a step that would give China leverage over the deal despite the relocation. Beijing is also concerned the acquisition might prompt other start-ups to move abroad to evade local oversight. 

In AI, algorithms are now treated as strategic assets, and their “nationality” is shaped less by a company’s address than by where its code was written and its data trained. An example: China relied on similar export-control tools during Washington’s earlier push to force ByteDance to sell TikTok’s US operations. Clearly, shifting to a neutral jurisdiction offers less protection than many founders had hoped.

For Manus, the pressure has come not just from China. The group’s move to Singapore was driven by US investment restrictions too. A US-led funding round drew questions from the US Treasury, prompting the company to distance itself from China.

The immediate risk to Singapore is obvious. Its role as a bridge between the US and China in AI is no accident: Singapore has built the regulatory footing, digital infrastructure and connectivity that make it a practical base for AI firms. Those companies bring opportunities: more talent, more capital and a chance for Singapore to become a base for AI work across South-east Asia.

Yet Singapore’s position between the US and China comes with real exposure. A small, trade-dependent economy – with exports and imports worth more than 300 per cent of GDP – is vulnerable if either side turns the screws on technology flows. 

These developments also come at a time when Singapore is banking on AI to drive new jobs, draw investment and support the growth of its digital economy. Prime Minister Lawrence Wong recently announced a nationwide push into AI when delivering the most recent Budget statement.

Already, the Meta-Manus deal shows how the city-state could become a pressure point. If Singapore ends up being a place where Chinese AI firms repackage themselves for Western markets – shifting their headquarters while keeping core engineering or data assets in China – it risks triggering tougher US export and investment controls on more Singapore-headquartered firms, and possible retaliation from Beijing. 

The dynamic mirrors longstanding US efforts to stop Chinese goods being routed through third countries to dodge tariffs. Regulators are looking past the label to where the work – or the data – actually sits.

Shaping global AI rules

The longer-term risk of this strategy is that Singapore hosts AI activity without influencing how it is governed. Shaping global AI rules to secure an open and inclusive environment for AI investments is in its interest. 

For this reason, the nation has developed guidelines for responsible AI use, and a framework for testing how AI systems perform against them, giving it a chance to help set the standards. Singapore’s commitment to multilateralism and a rules-based order offers rare stability, at a time when global politics is fracturing.

Elsewhere, the US, Europe and China are taking three approaches that reflect fundamentally different views of how technology should be governed. The US is leaning towards faster deployment of AI, but with patchy, state-level oversight. The EU has imposed comprehensive rules under the AI Act, while China subjects AI to tight state supervision.

The divide between US, Chinese and European approaches means AI is likely to develop under a patchwork of rules instead of one global standard. Companies will need to meet different requirements for data, safety and model behaviour in each market, raising both costs and the compliance burden. This fragmentation is pushing smaller economies to look for practical models they can follow, even if full alignment is not realistic.

Being neutral is not enough

The biggest risk is reputational: if “Singapore washing” truly takes hold, the nation’s credibility with global investors – and its appeal as a neutral, predictable place to do business and deploy capital – could quickly weaken. So how to avoid that fate? 

Singapore’s ability to stay neutral will hinge on what is, in effect, a masterclass in geopolitical agility. It needs to remain economically useful to both Washington and Beijing while widening ties elsewhere to avoid over-reliance on either side. 

Diplomatic consistency matters too: Singapore’s support for a rules-based order allows it to act as an honest broker in global forums. That balance will be tested as more AI firms like Manus relocate, whether by choice or pressure. 

Neutrality by itself offers little protection. What will determine Singapore’s resilience is whether it can pair its rules-based stance with credible oversight and a stronger voice in shaping global AI norms, making it harder for either side to see it as a loophole.