Singapore-Based Manus Is Independent Again. But Washington and Beijing May Still Set Its Limits

Manus is becoming independent again. That does not necessarily mean it is becoming free.

The Singapore-based AI start-up has recently announced it will resume independent operations upon completing its separation from Meta Platforms. As part of the separation, user data generated on or after December 29, 2025—the date Meta’s acquisition closed—will be deleted between August 23 and 24 (Singapore time). Users can back up their data until then, with data restoration set to begin on August 25.

On paper, the announcement moves Manus toward an outcome that appeared increasingly inevitable after China’s National Development and Reform Commission (NDRC) ordered Meta in April to unwind its $2 Bn-plus acquisition of the AI agent developer. Tencent Holdings has since entered talks to become the start-up’s largest shareholder. In July, media reports suggested the proposed restructuring might allow Manus to operate independently out of Singapore rather than face an outright acquisition by the Chinese technology conglomerate.

But the more compelling question is where this shifting ownership story will ultimately land.

Manus may regain control over its company. But its experience suggests that some of the decisions that make a technology business independent—who can finance it, who may acquire it, where its technology can travel and how much protection an offshore headquarters provides—are increasingly shaped by governments as much as corporate boards.

That tension extends far beyond one start-up.

Washington has steadily tightened national security boundaries around Chinese technology, investment and supply chains. Beijing, while opposing many of those restrictions, is also strengthening control over strategically important technology, data and talent leaving China. However, the policies are not equivalent and emerge from different political systems, legal frameworks and security priorities.

For companies straddling these two markets, a singular commercial reality is fast setting in. Owning the cheapest technology, securing the most willing investor, or finding the highest bidder may no longer be enough.

When Commerce Clashes with National Security

This contradiction is particularly stark in the US, where Chinese tech firms have historically built massive commercial demand on price and capability before Washington moves to restrict them.

Since December, the Federal Communications Commission (FCC) has added four broad categories of foreign-made hardware to its ‘Covered List’, effectively blocking new models from securing the equipment authorisations required for import, marketing, or sale in the US. This rapid expansion swept in drones in December 2025, consumer routers in March 2026, and, on July 28, a dual addition of advanced robotic devices and connected power inverters.

A conditional-approval pathway through the Departments of Defense and Homeland Security allows individual manufacturers to seek clearance regardless of where they are based. However, in practice, the restrictions fall most heavily on Chinese-made supply chains given Beijing’s dominance in these categories.

FCC Chairman Brendan Carr said that the restrictions on robots and power inverters were meant to accelerate domestic manufacturing as much as address national security risk, arguing it was “the right time to move investment back home”.

This second objective gets to the cold economics underpinning the geopolitical clash.

China completely dominates the global supply chain for power inverters, the essential grid-tied components that bridge solar installations and battery arrays to the utility network. Industry analyst Wood Mackenzie notes that Chinese giants Huawei and Sungrow swept 55% of the global market in 2024—marking a decade of consecutive market leadership—amid record global shipments of 589 GWac.

The trend points to a familiar playbook in which Chinese tech expands its share of Western markets by aggressively driving down prices. Now, as Washington institutes pre-emptive bans on new models to secure US data centres and critical energy infrastructure, the booming demand for the artificial intelligence buildout is forcing this supply-chain reality into the spotlight.

Artificial intelligence is making this exact tension difficult to ignore.

Chinese developers are aggressively expanding into international markets, leveraging not only rapidly improving capabilities but also substantially lower prices.

Start-ups like Moonshot AI, Z.ai (Zhipu) and MiniMax have steadily rolled out highly capable systems at a fraction of Western development costs. On the other hand, US cloud platforms, including Together AI and DigitalOcean, now offer commercial hosting for Chinese open-weight models, making frontier architectures like Moonshot’s Kimi and DeepSeek’s latest releases readily accessible to global enterprises.

Entrepreneur Asia Pacific has tracked this shifting pricing dynamic, highlighting how DeepSeek and Alibaba are successfully competing on raw performance and cost efficiency.

For businesses, the calculus is straightforward. If an AI model, a robotic system, or a piece of hardware infrastructure performs sufficiently well at a materially lower cost, buyers possess a rational commercial incentive to deploy it, irrespective of its origin.

National governments, however, are arriving at a vastly different conclusion.

Procurement decisions, once dictated solely by price, performance, reliability and technical support, must now account for a critical new variable: the technology’s provenance and whether its adoption creates a strategic dependency.

Consequently, geopolitical nationality is becoming embedded in the total cost of technology. And the same protectionist principle has bled into the movement of capital. Under outbound investment rules implemented by the US Treasury on January 2, 2025, American investors are prohibited from funding—or must formally notify the government of—certain transactions involving semiconductors, microelectronics, quantum information systems and advanced AI in China, Hong Kong and Macau.

The Treasury maintains these boundaries are vital to prevent US capital and strategic expertise from accelerating foreign capabilities that threaten national security. But as the unravelling of the Meta acquisition demonstrates, Manus just collided with the reverse side of that equation.

Beijing Builds a Ring-Fence of Its Own

Manus had already moved away from China. The company shifted its headquarters and operations to Singapore as it sought international investors, customers and a corporate structure better suited to operating globally amid tightening US-China technology restrictions. It was one of a growing number of China-founded businesses seeking Singapore as an operating base.

But relocating the company did not relocate every regulatory claim attached to it.

Weeks before Beijing’s formal order, its reach became personal. Co-founders Xiao Hong and Ji Yichao were barred from leaving China after being summoned to a meeting with the NDRC in March, The Financial Times reported, as regulators continued their review of the Meta deal.

When the NDRC ordered Meta to reverse its completed acquisition in April 2026, China demonstrated that Singapore incorporation would not necessarily put technology developed by China-founded companies beyond its reach. Chinese regulators had focussed on the company’s technology, intellectual property and talent, with Beijing viewing AI as a sector critical to national security.

China has since put a broader legal structure around that approach. Its new outbound investment rules under State Council Decree No. 837, the country’s first overarching administrative regulation on the subject, took effect on July 1 and expanded regulators’ authority over overseas transactions involving Chinese investors, technology, data and national security. The framework gives Beijing a formal basis to scrutinise and, under specified conditions, require the unwinding of completed overseas transactions.

The rules matter because the regulatory question no longer stops at ownership. Beijing has been seeking greater control over outbound flows of technology, intellectual property and talent, making it harder to separate strategically important know-how from its country of origin simply by transferring a corporate entity overseas.

China is also considering controls closer to the product itself. Authorities have discussed whether overseas access to some of the country’s most advanced AI models should be restricted as those models gain international users. Entrepreneur Asia Pacific reported in July that Beijing was considering curbs on overseas access to advanced AI models and tighter scrutiny around strategic technology transfers.

That creates a contradiction increasingly familiar on both sides of the Pacific.

China wants its AI companies to compete internationally. Lower prices and open models can help them win overseas developers and customers. But the more strategically valuable those technologies become, the stronger the argument inside Beijing for controlling where they go.

Washington is wrestling with the mirror-image problem. Businesses want access to capable, inexpensive Chinese technology, while policymakers worry about strategic dependence and security exposure.

Neither country is ending technology trade altogether.

Both are increasingly deciding that some technologies are too important to be governed purely by market logic.

The Limits of a Neutral Operating Base

Singapore sits squarely within this new corporate calculus. Many China-founded or China-linked companies have sought the city-state as a comparatively neutral operating ground as US-China tensions intensified. Singapore offers access to global capital, international customers, deep financial markets and extensive trade links without being part of either country’s domestic market. Reuters has reported that interest from Chinese companies looking to domicile there has accelerated across technology, biotechnology, data centres and critical minerals.

The trend has sometimes been described by analysts as ‘Singapore washing’, although the phrase risks understating the genuine operational reasons companies choose the city-state. The more important issue is whether relocating there can change how governments perceive a business’s nationality.

Several prominent companies suggest the answer is increasingly complicated.

Shein moved its headquarters from China to Singapore in 2022 and built an international corporate structure around the city-state. Yet its efforts to list in New York and London ran into political and regulatory resistance due to its China-linked supply chain, and the company still required Chinese regulatory approval for its current plans to list in Hong Kong.

China’s securities regulator cleared the listing on July 10. Shein is now seeking a valuation of $30-40 Bn, a steep reset from the $98.2 Bn it commanded in a 2022 funding round. Entrepreneur Asia Pacific has covered Shein’s latest listing push in detail.

TikTok provides another version of the same problem. Its chief executive, Shou Zi Chew, is Singaporean, and the platform has a substantial presence in Singapore. But US scrutiny centred on its ownership by China-based ByteDance rather than the nationality of its CEO or where parts of the business operated.

A 2024 law threatened TikTok with a US ban unless its ownership underwent a qualified divestiture. ByteDance ultimately agreed to a restructuring under which American and global investors hold 80.1% of TikTok USDS Joint Venture. In comparison, ByteDance retains 19.9%, with the new entity controlling US user data, apps and the recommendation algorithm.

Manus completes the circle from the other direction.

It moved to Singapore and was acquired by an American technology company. Beijing still intervened.

The lesson is not that Singapore failed these businesses. It is that a neutral headquarters can provide an international operating platform without necessarily neutralising the nationality governments attach to technology, ownership, intellectual property or talent.

For founders, that shifts geopolitical risk much earlier into company building. Where was the IP developed? Who created it? Where do key engineers work? Whose capital funded the company? Which chips, models and infrastructure does it depend on? Where is its data held? Which regulator may claim jurisdiction if it is sold?

Questions once left largely to lawyers during an IPO or acquisition are increasingly influencing how a technology company is structured from the beginning.

India Shows Both Sides of the Equation

India confronted a similar problem earlier. New Delhi began blocking Chinese-linked apps after its 2020 border confrontation with China. The restrictions ultimately extended beyond companies headquartered in mainland China.

Free Fire, published by Garena, was one of the clearest examples. The popular mobile game belonged to Sea Ltd, a Singapore-headquartered technology company. However, India included it among the 54 apps blocked in February 2022 over concerns that user data was being sent to servers in China.

Singapore raised the matter with New Delhi, asking why an app belonging to a Singapore-based company had been caught in a crackdown on Chinese apps. Sea’s market value dropped by more than $16 Bn in a day—an 18% single-session decline, its steepest on record—after news of the ban.

The episode demonstrated how quickly a company’s legal nationality and its perceived technological nationality can diverge. But India’s subsequent policy shift also shows the limits of trying to completely separate security from economics.

On March 10, 2026, New Delhi’s cabinet eased some of the investment restrictions imposed in 2020. Proposals from land-bordering countries in specified manufacturing sectors, including electronic components and capital goods, can now be processed within 60 days, provided that majority ownership and control remain with resident Indians. Investors with non-controlling beneficial ownership of up to 10% can use the automatic route, subject to sectoral limits.

The government said the changes were intended to improve access to technology and integration with global supply chains. But industry pressure was part of the reason.

Earlier restrictions had constrained manufacturers dependent on Chinese technology and capital. At the same time, India’s trade deficit with China had reached a record $99.2 Bn in FY25, driven heavily by electronics, components and machinery.

That is an important counterweight to a simple decoupling narrative. Governments can erect security barriers. Businesses and supply chains continue to respond to cost, capability and availability. When economic dependence becomes sufficiently important, those barriers can be selectively redrawn.

The emerging technology order may therefore be less about complete separation than about a constantly shifting boundary between what markets are allowed to choose freely and what governments consider too strategic to leave entirely to them.

What Independence Means for Manus Now

That brings the argument back to Manus. Commercially, it has already demonstrated demand for what it built. A June report by The Information, cited by Reuters, said Manus’s annualised revenue run rate had risen to $400-500 Mn, from about $100 Mn when Meta acquired it. The figure is a run rate rather than confirmed recurring revenue, but the scale of the increase shows that Manus’s immediate problem was not a lack of commercial momentum. Nor was it unable to attract global capital or find a buyer.

It did both.

What it could not do was ensure that the transaction’s commercial logic would be sufficient to determine the outcome. The $2 Bn deal had closed in December, but by May, Meta and Manus had already completed their operational split and cut off data sharing following Beijing’s order to unwind it.

That is what makes its return to independence more significant than another turn in an unusual M&A dispute. Manus is testing whether a generation of technology companies can continue to combine Chinese engineering roots, Singapore operating structures, global capital and international customers without eventually being classified by one government or another as strategically belonging somewhere.

The market continues to push towards lower costs, global pools of talent and technologies that cross borders quickly. Governments are pushing towards trusted supply chains, domestic technological capacity and tighter control over assets they consider strategic.

Those forces are increasingly meeting inside individual companies.

Manus may be getting its ownership back. What it may not get back is the freedom to operate like an ordinary global business.
With inputs from Agencies