Why China Is Rethinking Outbound AI Deals in the Age of Tech Sovereignty

Beijing’s scrutiny of Meta’s $2B Manus acquisition reflects a broader shift toward tighter control over cross-border AI flows

China is tightening its oversight of outbound artificial intelligence (AI) deals, indicating a broader policy shift that extends beyond standard regulatory checks. The reported review of Meta’s $2 B acquisition of Manus a Singaporean AI startup with links to China—highlights this change. Transactions that would earlier be seen as routine cross-border deals are now being assessed more carefully for their implications on national security, data governance, and control over critical technologies.

At the center of the issue is a fundamental question: who controls AI innovation developed within China’s ecosystem? While companies have increasingly adopted offshore structures to attract global capital and reduce geopolitical friction, regulators appear less willing to overlook the origin of core technologies. In the case of Manus, its Chinese founding background and potential access to sensitive data have reportedly triggered concerns, even though the company is headquartered outside mainland China.

From Capital Outflow to Technology Control

China has long maintained oversight of outbound investments, particularly in sectors tied to national interest. However, AI is now being treated differently. Unlike traditional industries, AI sits at the intersection of data, algorithms, and compute—making it both commercially valuable and strategically sensitive.

This evolving stance suggests that outbound deals involving AI are no longer just financial transactions. They are increasingly viewed as transfers of strategic capability. As a result, regulators are expanding their reviews to include: (1) potential export of sensitive technologies 2) data security implications, (3) ownership structures and control rights, (4) compliance with cross-border capital regulations. In practical terms, this means deals can face delays, restructuring, or even reversal if they are seen to conflict with national priorities.

The Limits of Offshore Structuring

For years, Chinese startups have used jurisdictions like Singapore to position themselves as global companies. This approach has been particularly common in the AI and deep-tech sectors, where access to international funding and partnerships is critical.

However, the Manus case highlights the limitations of this strategy. Relocation does not fully detach a company from its origins, especially when: (1) core R&D activities remain linked to China, (2) founding teams or early investors are Chinese, (3) datasets or training models are developed within China. Beijing’s increasing willingness to “look through” offshore structures suggests that regulatory scrutiny will follow the substance of the business, not just its legal domicile.

AI as a Strategic Asset

China’s tightening stance is part of a much larger shift: AI is no longer treated as just another high-growth tech sector—it is increasingly viewed as critical national infrastructure.

Unlike earlier waves of technology, AI sits on top of three highly sensitive layers: data, computing power, and algorithms. Together, these determine not just commercial outcomes, but also influence in areas such as defence, economic planning, and information control. This is why governments, including China’s, are becoming far more protective of how AI capabilities are developed and where they ultimately end up.

For China, keeping a firm grip on AI development serves several strategic goals:

• Staying competitive with the US:
  The global AI race is largely framed around US–China rivalry. Losing promising AI companies or core technologies to foreign ownership could weaken China’s position over time.
• Controlling data flows:
  AI systems are only as strong as the data they are trained on. If companies with access to Chinese datasets are acquired by foreign firms, it raises concerns about indirect data transfer or loss of control over how that data is used.
• Building domestic champions:
  China has been actively nurturing its own AI ecosystem. Allowing key startups to be absorbed by global tech giants could limit the growth of local leaders and reduce long-term economic value within the country.

This is why deals like Meta’s acquisition of Manus are drawing close attention. Even if the company is headquartered overseas, the question regulators are asking is deeper: where does the real value of this innovation sit, and who will ultimately control it?

In this context, ownership is only one part of the equation. What matters more is the flow of talent, intellectual property, and data. If these elements are seen to be moving out of China’s sphere of influence, regulators are far more likely to intervene.

Put simply, AI is no longer just about building products—it is about shaping national capability. And that is fundamentally changing how cross-border deals are evaluated.

Rising Friction in Cross-Border AI M&A

The implications extend well beyond a single transaction. China’s evolving stance is introducing a structural layer of uncertainty into cross-border AI dealmaking—one that is difficult to price, predict, or navigate.

For global investors and acquirers, the challenge is no longer just identifying strong technology or market fit. It is increasingly about assessing whether a deal can withstand regulatory scrutiny across jurisdictions, particularly when Chinese-origin technology is involved. This shift is slowing decision-making and, in some cases, discouraging deals altogether.

Several risks are becoming more pronounced:

• Unpredictable approvals:
  Regulatory reviews are becoming less transparent and more case-specific. Even deals that appear compliant on paper may face extended reviews or unexpected conditions, making timelines harder to manage.
• Retroactive intervention:
  There is growing concern that authorities could revisit or impose conditions on deals after they are announced—or even after completion—if new risks are identified. This raises the stakes for both investors and founders.
• Higher compliance burden:
  Startups with any linkage to China—whether through founders, R&D, or data—are now subject to deeper scrutiny. This includes how data is stored, where models are trained, and who ultimately controls the technology.

As a result, cross-border AI M&A is becoming more cautious and selective. Investors are likely to prioritise deals with cleaner structures and fewer geopolitical complications, while startups may need to redesign their corporate setups from the beginning. This includes clearer separation of markets, stricter data governance, and more transparent ownership models to reduce regulatory friction later on.

In effect, what was once a relatively straightforward global deal environment is evolving into a more fragmented system—where political boundaries increasingly shape how and where AI companies can scale or exit.

What It Means for Startups and Investors

For founders, especially those operating between China and global markets, the message is becoming clear: geopolitical considerations are now as important as product and market fit.

Startups may need to localise data and infrastructure more clearly, separate China and international operations and build transparent governance structures to ease regulatory concerns

For investors, particularly venture capital and private equity firms, due diligence will need to go deeper into regulatory exposure—not just financial or technical factors.

China’s reassessment of outbound AI deals marks a turning point in how cross-border technology transactions are viewed. The focus is shifting from enabling capital flows to safeguarding strategic capabilities. While the immediate trigger may be high-profile cases like Meta’s acquisition of Manus, the underlying trend is broader and likely to persist.

As AI becomes central to economic and national power, governments are asserting greater control over how it is developed, transferred, and commercialised. For the global startup ecosystem, this means adapting to a reality where innovation is no longer borderless—and where the rules of engagement are being rewritten in real time.

Quick takeaways:

• China is tightening oversight on outbound AI deals, signaling a shift from open capital flows to stronger tech control.
• Meta’s $2B Manus acquisition has become a test case, highlighting how even offshore-structured startups face scrutiny if they have Chinese roots.
• AI is now viewed as a strategic asset, not just a commercial technology, due to its link to data, infrastructure, and national competitiveness.
• Regulatory focus is expanding beyond ownership to include data flows, IP control, and where innovation ultimately resides.
• Cross-border AI M&A is becoming riskier, with slower approvals, higher compliance demands, and potential post-deal intervention.
• Offshore structures (e.g., Singapore setups) are no longer enough to avoid Chinese regulatory attention.
• Investors and startups must rethink deal structures, including clearer data governance and separation of markets.
• The broader trend reflects rising US–China tech tensions, where AI is central to geopolitical competition.
• Big picture: Global AI dealmaking is entering a more fragmented phase, where geopolitics increasingly shapes who can buy, build, and scale technology.