01 The Final Verdict: Unprecedented Regulatory Severity

The narrative of "AI global expansion" has met a harsh reality. On April 27, 2026, China's National Development and Reform Commission (NDRC) officially prohibited the foreign acquisition of Manus, ordering the parties to rescind the transaction.

The severity of the language marks a turning point: AI competition is no longer just a technical issue but a core element of national power.

For a team that was recently courted by Silicon Valley giants and praised by Microsoft’s CEO, the fall is a stark reminder that technology cannot be decoupled from geopolitics.

02 From Wuhan to Silicon Valley: A Flawless Growth Trajectory

Founded by 1993-born entrepreneur Xiao Hong, Manus’s parent company, Butterfly Effect, started in Wuhan.

Its early successes, like the AI plugin Monica, attracted top-tier Chinese venture capital (ZhenFund, Sequoia China, and Tencent), reaching a valuation of $85 million by late 2024.

At this stage, the company was firmly rooted in China, specifically the Wuhan "Optics Valley" headquarters.

03 The Benchmark Shackles: Regulatory Entanglement

The turning point came in April 2025. Following the release of the Manus product, Silicon Valley’s Benchmark led a $75 million Series B round, pushing the valuation to nearly $500 million.

However, this capital came with the "Reverse CFIUS" (U.S. Outbound Investment Security Program) oversight.

To satisfy American investors and de-risk the deal, Manus was forced to relinquish its autonomy regarding its headquarters and strategic direction.

04 The Risky Maneuver: "Singapore Washing"

In a bid to appease investors and secure high-end GPU clusters, Manus attempted a "re-domiciliation" strategy in mid-2025:

• The three co-founders relocated to Singapore.

• The entity was renamed Butterfly Effect Pte. Ltd., with new offices in San Francisco and Tokyo.

• The Chinese team was slashed from 120 to 40, Chinese social media accounts were purged, and Chinese IPs were blocked from the website.

Legally, this was a classic case of "Singapore Washing"—migrating R&D and data to a neutral economy to bypass U.S.-China investment screenings. However, regulators noted that the core technical milestones were achieved using Chinese engineering resources and capital prior to the move.

05 The Failed Meta Acquisition: A Miscalculation of Geopolitics

By late 2025, Meta proposed a $2 billion acquisition of Manus. The deal was built on the miscalculation that a "compliance window" existed before stricter regulations took hold. Manus underestimated a fundamental logic: regulators no longer focus solely on the place of registration, but on the "ancestry" of the technology itself. On April 27, 2026, the deal became the first AI acquisition publicly blocked under China's 2020 Measures on Security Review of Foreign Investment.

06 Conclusion: The Cost of Walking the Tightrope

The tragedy of Manus lies in its attempt to walk a geopolitical tightrope—leveraging Chinese engineering dividends while seeking Silicon Valley labels and a U.S. "exit" via acquisition.

• The Result: It lost favor on both sides. By the time the deal collapsed, the Chinese domestic market was already dominated by giants like ByteDance, Alibaba, and Tencent, leaving no room for Manus to return.

• The Lesson: In the current era of great power competition, attempting to bypass oversight through "shell-switching" is increasingly becoming a dead end. For AI startups, the path forward requires a singular, committed choice rather than an attempt to play both sides.