The Great Escape That Wasn't
How China's "Penetrating Supervision" Is Rewriting the Rules of Tech Capital Flight
On December 29, 2025, Meta announced a $2 billion acquisition of Manus, a Chinese AI startup. It should have been a textbook success story: a company acquired by a tech giant just nine months after launching its flagship product, its valuation quadrupling from $500 million to $2 billion. Meta’s third-largest acquisition ever. For the founders and their investors at Benchmark Capital, Manus looked like the perfect template for how Chinese AI entrepreneurs could navigate the treacherous waters between Washington and Beijing.
Ten days later, on January 8, 2026, everything changed. At a routine press conference, He Yadong, spokesman for China’s Ministry of Commerce, announced that authorities would launch an evaluation and investigation into the deal, examining its compliance with export control regulations, technology import and export laws, and outbound investment rules. The transaction’s fate was suddenly in limbo.
This was no ordinary regulatory review. Manus had already completed what industry insiders call “Singapore washing” back in June 2025. The company relocated its headquarters to Singapore, shuttered its Beijing and Wuhan offices, laid off 80 of its 120 employees, deleted its Chinese social media presence, and even blocked Chinese IP addresses from accessing its platform. By any traditional legal standard, Manus was no longer a Chinese company. So why could Beijing still claim jurisdiction?
Part I: The Rise of Penetrating Supervision
The answer lies in a new regulatory mechanism that China has been quietly building over the past five years: what we might call “penetrating supervision.”
To understand this shift, we need to trace the evolution of China’s tech capital controls through four distinct phases. In the first phase, from late 2020 to early 2021, the landmark case was Ant Group’s derailed IPO, halted just 48 hours before what would have been the world’s largest public offering. Regulators focused primarily on financial risk, concerned that Ant’s consumer lending products were essentially leveraging minimal capital to create massive credit exposure. The resulting fine of 7.12 billion yuan and forced restructuring set the template for what was to come.
The second phase, spanning late 2021 to 2022, centered on Didi. The ride-hailing giant went public on the New York Stock Exchange on June 30, 2021, raising $4.4 billion. Just two days later, Chinese regulators launched a cybersecurity review. Within weeks, Didi’s app was pulled from stores, and seven government departments jointly stationed investigators at the company. The eventual fine of 8.03 billion yuan established data security as the regulatory red line. Investigators found that Didi had illegally processed 64.7 billion pieces of personal information over seven years.
The third phase, from 2023 to 2024, saw the institutionalization of oversight through the overseas listing filing system. What had been emergency interventions became routine procedures. By the end of 2024, 199 Chinese companies had completed filings for overseas IPOs, with 62 percent choosing Hong Kong as their destination.
Now, in the fourth phase beginning in 2025, the Manus case signals a new priority: technology sovereignty. The core concern is no longer just data or financial risk, but the potential outflow of AI capabilities that Beijing considers strategically vital.
The concept of “penetrating supervision” represents a fundamental shift in how Beijing approaches cross-border tech deals. Regulators no longer simply check where a company is registered. Instead, they trace the true origin of the technology itself. In the Manus case, the investigation focuses on three critical questions: Was the core technology developed within Chinese territory? Does the transfer of personnel to Singapore constitute a de facto technology export? And should the company have applied for an export license under China’s controlled technology list?
This approach reflects a deeper concern in Beijing about precedent-setting. As Winston Ma, a law professor at New York University and partner at Dragon Capital, told the Wall Street Journal: “If this deal goes through smoothly, it validates an exit blueprint: relocate first, then sell globally.” Chinese authorities worry less about one company leaving and more about the demonstration effect. If Manus succeeds, dozens of AI startups might follow the same playbook, potentially draining the country’s most valuable technical talent.
Yet it would be a mistake to characterize China’s approach as simply restrictive. The regulatory picture is more nuanced than Western coverage often suggests. In 2024, the monthly volume of data export security assessment applications dropped by approximately 60 percent, as authorities streamlined procedures for non-sensitive data transfers. Companies like CATL, the battery giant, completed their overseas listing filings in as few as 25 days. Beijing’s logic is differentiated: relax controls for sectors that do not involve core technologies while tightening them for AI algorithms and other strategic areas. Technology companies and green economy firms received “fast track” treatment, while hard-tech companies accounted for over 60 percent of approved filings in 2025.
The signal of stricter enforcement in sensitive sectors, however, is unmistakable. A newly revised Foreign Trade Law, passed on December 27, 2025, and set to take effect in March 2026, dramatically increases penalties. Violators now face fines of one to five times their illegal gains, and in serious cases, criminal prosecution. For founders like those at Manus, the stakes have never been higher. Technology sovereignty protection is entering a new phase.
Part II: The Vanishing Gray Zone
Penetrating supervision is only half the story. Chinese tech companies are actually caught in a two-way squeeze, facing pressure from both Beijing and Washington simultaneously. The space for companies to exploit regulatory gaps between the two superpowers is shrinking rapidly.
Consider the full extent of Manus’s “identity engineering.” The company did not just relocate; it undertook an extreme makeover. It moved headquarters to Singapore, closed offices in Beijing and Wuhan, reduced staff from 120 to about 40, scrubbed its Chinese social media accounts, and blocked access from Chinese IP addresses. Analysts at the Internet Governance Project identified a five-layer compliance stack that Chinese AI companies must now navigate to access American capital: geographic separation, revenue isolation, distance from state funding, data partitioning, and brand repositioning. Manus checked every box. And still, it was not enough to escape Beijing’s scrutiny.
From Washington’s direction, the pressure is equally intense. The Outbound Investment Security Program, often called “reverse CFIUS,” took effect on January 2, 2025, restricting American capital from investing in Chinese AI, semiconductor, and quantum technology companies. Benchmark’s $75 million investment in Manus, made in April 2025, prompted an inquiry from the U.S. Treasury Department. This scrutiny accelerated Manus’s decision to relocate. But even the company’s defense that it was merely a “Claude wrapper” leveraging existing AI models from Anthropic and Alibaba’s Qwen, rather than developing its own foundation models, did not shield it from examination. In Washington’s eyes, Chinese origins do not disappear simply because a company reincorporates in Singapore. The TikTok saga had already demonstrated that structural compliance does not guarantee acceptance.
The squeeze extends beyond the US-China axis. Consider Malaysia, which has emerged as a key transit point for restricted Nvidia chips heading to China. Between January and April 2025, GPU imports to Malaysia surged to $6.45 billion, a staggering 3,400 percent increase year-over-year, exceeding the country’s total GPU sales for all of 2024. The surge triggered alarm bells. In Singapore, authorities prosecuted three individuals in a $390 million chip smuggling case linked to DeepSeek, the Chinese AI company. Under American pressure to “monitor every shipment of Nvidia chips,” Malaysia’s trade minister announced a crackdown. In July 2025, the country issued Directive No. 1/2025, imposing immediate controls on the export and transshipment of high-performance AI chips. What some had hoped might be “Malaysia washing” was itself coming under siege.
The implications are stark. Companies like Shein, the fast-fashion giant, pursued listings in both the United States and the United Kingdom, only to encounter political resistance in both markets. The company is now reportedly considering a return to China to seek a Hong Kong listing instead. The gray zone that once allowed tech companies to straddle both worlds is systematically disappearing. Companies and countries alike are being forced to make clearer choices about where they stand.
Part III: Strategic Choices for the Global South
For nations standing between the two superpowers, what does this tightening landscape mean?
The core insight is that the question is not whether to choose China or America. That framing is a trap. The real question is how to develop domestic capabilities while navigating the pressure from both sides. China’s “penetrating supervision” offers an important lesson: technology sovereignty cannot be protected through legal formalities alone. What matters is substantive local technological capacity. Registration addresses and corporate structures are paper shields. Actual technical capabilities are the real armor.
There are several paths worth considering. India’s Digital Public Infrastructure model demonstrates how developing economies can build foundational technology layers, like the UPI payments system, that serve domestic needs while remaining open to international integration. Regional coordination mechanisms, whether through the African Union framework or the BRICS partnership, can provide collective bargaining power that individual countries lack. The New Development Bank’s launch of a $5 billion digital sovereignty fund and the BRICS Pay initiative to bypass SWIFT represent early experiments in alternative infrastructure.
And strategic flexibility matters. The countries that will negotiate the best terms from both superpowers are those that maintain options without betting everything on gray zone arbitrage. Resource-rich emerging economies, in particular, understand that playing both sides can yield technology access, tailored trade agreements, and preferential financing. But this strategy only works if it is backed by genuine domestic capacity building, not just clever legal structuring.
The vanishing gray zone is both a challenge and an opportunity. It forces nations to think seriously about their own technology sovereignty strategies rather than relying on regulatory arbitrage. The countries that will thrive are those that use this moment to invest in genuine capabilities, not those that hope the cracks in the system will remain exploitable forever.