=== TAG === Policy === HEADLINE === China Blocks Meta's $2 Billion Manus Deal, Orders Unwind === META_DESC === China's NDRC on April 27, 2026 ordered Meta to unwind its $2 billion acquisition of AI agent startup Manus under China's Foreign Investment Law, after 100 employees had already moved into Meta's Singapore offices and CEO Xiao Hong was reporting directly to Meta's COO. === DATE === April 27, 2026 === AUTHOR === Jane Sterling === READ_TIME === 9-minute read === HERO_IMG === img/content.png === SCRIPT_LABEL === Video Script (9 min, clean transcript for captioning) === SCRIPT === On April 27, 2026, China's most powerful economic planning body issued a ruling that stopped one of the most closely watched AI acquisitions of the decade. China's National Development and Reform Commission, known as the NDRC, ordered Meta and Manus to withdraw their roughly $2 billion acquisition agreement and prohibited the foreign investment under Chinese law. The agency published exactly 1 sentence explaining its decision: "The National Development and Reform Commission has made a decision to prohibit foreign investment in the Manus project in accordance with laws and regulations." One sentence. No named legal basis beyond a vague reference to laws and regulations. No appeal process described. No timeline for unwinding what had already been built. Just a single line, issued from one of the most powerful regulatory offices in the world, ordering a $2 billion deal to stop. This is NOT a story about a deal that got blocked before it could get started. By the time that 1-sentence ruling landed, the acquisition had been running operationally for months. Manus had moved roughly 100 employees into Meta's Singapore offices. CEO Xiao Hong was already reporting directly to Meta's COO, Javier Olivan. Joint development work had begun on agent systems and product integration. The two companies had started becoming one. Now Beijing is ordering that process reversed, without explaining what a compliant reversal looks like, without offering a timeline, and without addressing the fact that IP and trained systems created jointly may not have a clean ownership boundary. Meta responded with a statement that was careful and deliberate: "The transaction complied fully with applicable law. We anticipate an appropriate resolution to the inquiry." Meta is not going quietly. The phrase "appropriate resolution" is diplomatic language for a legal and political fight that is not over. The NDRC's order is the controlling fact. Meta shares fell 2.41 percent on the day of the ruling. Investors understood immediately that this was not just a blocked transaction. It was a blocked strategy. Meta had structured plans around what Manus could provide, and those plans are now in a deeply uncertain state. To understand why Beijing moved this way, and why this ruling matters far beyond Meta's balance sheet, you have to understand what Manus actually built and how a Beijing-founded AI startup wound up at the center of a confrontation between the world's two largest technology powers. Manus was founded in Beijing in 2022 by Xiao Hong, Yichao Ji, and Tao Zhang, operating under a parent company called Butterfly Effect. The team built a general-purpose AI agent: not a chatbot, not a narrow single-task tool, but a system designed to autonomously browse the web, write and execute code, fill out forms, analyze documents, and complete multi-step tasks that would ordinarily require a human being making judgment calls at every stage. When Manus launched publicly on March 6, 2025, the AI industry took notice immediately. The reason was the benchmark numbers. Manus posted a score of 86.5 percent on GAIA Level 1, one of the most demanding evaluations of general AI agent capability in use today. That score beat the 74.3 percent that OpenAI's Deep Research agent had posted on the same benchmark. GAIA does not measure trivia recall. It tests whether an agent can actually accomplish real-world tasks that require chaining together multiple tools, multiple decisions, and multiple steps in a coherent sequence. Manus went further still: it posted 70.1 percent on GAIA Level 2 and 57.7 percent on GAIA Level 3, which are progressively harder tiers designed to push the outer limits of what autonomous agents can accomplish without human correction. The business performance matched the technical capability. Manus reached roughly $100 million in annualized recurring revenue within just 8 months of its public launch. Consumer subscriptions started at $39 per month, with a Pro tier at $199 per month. That is a company with genuine product-market fit, one that enterprise customers were deploying for real workflows. By April 2025, Benchmark led a $75 million Series B that valued Manus at $500 million. Seven months later, Meta announced an acquisition at more than $2 billion. From $500 million to over $2 billion in less than a year. Here is where the corporate architecture began to matter. Before the Meta deal went public, Manus had relocated its headquarters from Beijing to Singapore in mid-2025. This kind of move is not unusual for Chinese-founded technology companies positioning themselves for Western acquisition. Singapore has stable legal institutions, English-language courts, and a reputation as a politically neutral hub. The industry term for this strategy is Singapore washing: restructure the formal legal domicile so that a U.S. buyer can proceed without appearing to trigger Chinese national security or foreign investment review processes. It failed here, decisively. China's Ministry of Commerce opened a formal review of the Meta-Manus transaction in January 2026. The antitrust regulator joined the process. By late March 2026, co-founders Xiao Hong and Yichao Ji were placed under exit bans. Both normally lived in Singapore. Beijing summoned them to the mainland and barred them from leaving the country. Exit bans are a tool Beijing uses when it wants direct personal leverage over specific individuals, not over corporate entities. The government was not interested in Manus's Singapore incorporation documents. It was interested in the people who built the technology, and those people were reachable regardless of where the company was registered. The exit bans were a signal that this was not a routine antitrust review. Standard antitrust asks whether a deal concentrates market power inappropriately. What Beijing was asking was different: whether Chinese-developed AI technology should be allowed to pass into foreign hands at all. That is a national security and industrial policy question, and it is one that China's Foreign Investment Law was specifically built to address. The exit bans were the visible enforcement mechanism for a much broader political judgment that had already been reached. The NDRC's formal ruling came under that same Foreign Investment Law, which gives Beijing broad discretionary authority to block foreign acquisitions of assets it considers strategically sensitive. Agentic AI, the category of technology capable of acting autonomously inside complex systems, has now been placed unambiguously in that category. A general-purpose agent that can browse, execute code, and complete real-world multi-step tasks is not just a software product. From Beijing's perspective, it is infrastructure, and infrastructure does not get transferred to an American company through a Singapore holding structure and a wire transfer. The immediate operational challenge for Meta is unlike anything the company has faced before. By March 2026, the integration between Meta and Manus was not just organizational. It was technical. Around 100 employees were working inside Meta's Singapore offices. Xiao Hong was inside Meta's reporting hierarchy. Systems had been trained, integrated, and developed jointly. How do you reverse that? What IP belongs to Manus as a stand-alone entity versus what was produced under Meta's ownership? Where do you draw the line on a trained system built using data and compute from both companies simultaneously? The NDRC has provided no answers to any of these questions. It has only issued an order. There is a customer complication layered on top of that. Earlier in 2026, some Manus enterprise customers publicly distanced themselves from the company after the Meta deal was announced, citing concern about their workflows and proprietary data ending up inside Meta's ecosystem. So the idea that Beijing's ruling simply restores a clean prior state is wrong twice: the technical integration is already deep, and the customer base Manus had before December 2025 is not intact and waiting at the door. The competitive implications across the AI industry are real and immediate. Meta's agent strategy had been built around what Manus could provide in the race against Google's Project Astra and OpenAI's growing agent products. That strategy is now disrupted. Google is expected to push harder on its own agent roadmap. Microsoft, which built its Copilot agent infrastructure on top of OpenAI technology inside U.S. and EU jurisdictions, looks comparatively insulated from NDRC exposure. Anthropic and other U.S.-headquartered vendors gain by default as the pool of Chinese-founded acquisition targets available to American buyers gets effectively smaller. OpenAI is in a complicated position: Manus had been outscoring its Deep Research agent on GAIA in a very public way, and the disruption of Manus reduces that competitive pressure in the near term, though not through any product achievement on OpenAI's part. The deeper signal from April 27 is about the limits of corporate restructuring as a geopolitical tool. A company founded in China, built by Chinese founders, with IP developed under Chinese law, does not escape Chinese regulatory reach by updating its corporate address. The founders carried Chinese passports. The original technology was developed in Beijing. The founding networks were Chinese. None of that evaporates when you file new documents in Singapore. Beijing has now shown it will reach through a corporate structure, past the closing date of a transaction, past operational integration, and order the entire arrangement reversed. That is what makes this ruling categorically different from a standard blocked deal. It is not a checkpoint at the border. It is a hand reaching into an already-running operation and pulling out the foundation. If you are building AI in China and thinking about a Western acquisition exit, you are studying what happened to Xiao Hong and Yichao Ji very carefully. And if you are at a U.S. tech company evaluating Chinese-founded AI startups as acquisition targets, the question is no longer just whether you can get the deal approved. The question is whether the approval can be trusted to hold once the employees have moved, the systems are running, and the integration is real. On April 27, 2026, Beijing answered that question. The answer is: not necessarily. === SCRIPT_HTML === === ANNOTATED_LABEL === Annotated Script (with b-roll & cut cues) === ANNOTATED_HTML === [TALKING HEAD — hook]

On April 27, 2026, China's most powerful economic planning body issued a ruling that stopped one of the most closely watched AI acquisitions of the decade.

[STAT CARD: "Meta-Manus deal: ~$2B blocked by NDRC"]

China's National Development and Reform Commission, known as the NDRC, ordered Meta and Manus to withdraw their roughly $2 billion acquisition agreement and prohibited the foreign investment under Chinese law. The agency published exactly 1 sentence explaining its decision: "The National Development and Reform Commission has made a decision to prohibit foreign investment in the Manus project in accordance with laws and regulations." One sentence. No named legal basis beyond a vague reference to laws and regulations. No appeal process described. No timeline for unwinding what had already been built. Just a single line, issued from one of the most powerful regulatory offices in the world, ordering a $2 billion deal to stop.

[CUT] [VOICEOVER — scene 1] [B-ROLL: company-logo:Meta]

This is NOT a story about a deal that got blocked before it could get started. By the time that 1-sentence ruling landed, the acquisition had been running operationally for months.

[B-ROLL: news-studio]

Manus had moved roughly 100 employees into Meta's Singapore offices. CEO Xiao Hong was already reporting directly to Meta's COO, Javier Olivan. Joint development work had begun on agent systems and product integration. The two companies had started becoming one.

[B-ROLL: code-terminal]

Now Beijing is ordering that process reversed, without explaining what a compliant reversal looks like, without offering a timeline, and without addressing the fact that IP and trained systems created jointly may not have a clean ownership boundary.

[B-ROLL: data-center]

Meta responded with a statement that was careful and deliberate: "The transaction complied fully with applicable law. We anticipate an appropriate resolution to the inquiry." Meta is not going quietly. The phrase "appropriate resolution" is diplomatic language for a legal and political fight that is not over. The NDRC's order is the controlling fact.

[B-ROLL: finance-charts]

Meta shares fell 2.41 percent on the day of the ruling.

[STAT CARD: "Meta shares: -2.41% on ruling day"]

Investors understood immediately that this was not just a blocked transaction. It was a blocked strategy. Meta had structured plans around what Manus could provide, and those plans are now in a deeply uncertain state.

[/VOICEOVER] [TALKING HEAD — transition]

To understand why Beijing moved this way, and why this ruling matters far beyond Meta's balance sheet, you have to understand what Manus actually built and how a Beijing-founded AI startup wound up at the center of a confrontation between the world's two largest technology powers.

[CUT] [VOICEOVER — scene 2] [B-ROLL: ai-abstract]

Manus was founded in Beijing in 2022 by Xiao Hong, Yichao Ji, and Tao Zhang, operating under a parent company called Butterfly Effect.

[STAT CARD: "Manus founded: Beijing, 2022"]

The team built a general-purpose AI agent: not a chatbot, not a narrow single-task tool, but a system designed to autonomously browse the web, write and execute code, fill out forms, analyze documents, and complete multi-step tasks that would ordinarily require a human being making judgment calls at every stage.

[B-ROLL: code-terminal]

When Manus launched publicly on March 6, 2025, the AI industry took notice immediately. The reason was the benchmark numbers.

[B-ROLL: screen-capture:GAIA benchmark leaderboard]

Manus posted a score of 86.5 percent on GAIA Level 1, one of the most demanding evaluations of general AI agent capability in use today.

[STAT CARD: "Manus GAIA Level 1: 86.5%"]

That score beat the 74.3 percent that OpenAI's Deep Research agent had posted on the same benchmark.

[STAT CARD: "OpenAI Deep Research GAIA L1: 74.3%"]

GAIA does not measure trivia recall. It tests whether an agent can actually accomplish real-world tasks that require chaining together multiple tools, multiple decisions, and multiple steps in a coherent sequence.

[B-ROLL: data-center]

Manus went further still: it posted 70.1 percent on GAIA Level 2 and 57.7 percent on GAIA Level 3, which are progressively harder tiers designed to push the outer limits of what autonomous agents can accomplish without human correction.

[STAT CARD: "GAIA Level 2: 70.1%"] [STAT CARD: "GAIA Level 3: 57.7%"] [B-ROLL: finance-charts]

The business performance matched the technical capability. Manus reached roughly $100 million in annualized recurring revenue within just 8 months of its public launch.

[STAT CARD: "$100M ARR — just 8 months post-launch"]

Consumer subscriptions started at $39 per month, with a Pro tier at $199 per month.

[STAT CARD: "Consumer tier: $39/month"] [STAT CARD: "Pro tier: $199/month"]

That is a company with genuine product-market fit, one that enterprise customers were deploying for real workflows.

[B-ROLL: finance-charts]

By April 2025, Benchmark led a $75 million Series B that valued Manus at $500 million.

[STAT CARD: "Benchmark Series B: $75M at $500M"] [STAT CARD: "Manus valuation: $500M at Series B"]

Seven months later, Meta announced an acquisition at more than $2 billion. From $500 million to over $2 billion in less than a year.

[B-ROLL: ai-abstract]

Here is where the corporate architecture began to matter. Before the Meta deal went public, Manus had relocated its headquarters from Beijing to Singapore in mid-2025.

[STAT CARD: "Manus HQ moved to Singapore: mid-2025"]

This kind of move is not unusual for Chinese-founded technology companies positioning themselves for Western acquisition. Singapore has stable legal institutions, English-language courts, and a reputation as a politically neutral hub. The industry term for this strategy is Singapore washing: restructure the formal legal domicile so that a U.S. buyer can proceed without appearing to trigger Chinese national security or foreign investment review processes.

[/VOICEOVER] [TALKING HEAD — transition]

It failed here, decisively.

[CUT] [VOICEOVER — scene 3] [B-ROLL: courtroom]

China's Ministry of Commerce opened a formal review of the Meta-Manus transaction in January 2026.

[STAT CARD: "MOFCOM formal review: Jan 2026"]

The antitrust regulator joined the process. By late March 2026, co-founders Xiao Hong and Yichao Ji were placed under exit bans.

[B-ROLL: stills:China exit ban enforcement]

Both normally lived in Singapore. Beijing summoned them to the mainland and barred them from leaving the country. Exit bans are a tool Beijing uses when it wants direct personal leverage over specific individuals, not over corporate entities. The government was not interested in Manus's Singapore incorporation documents. It was interested in the people who built the technology, and those people were reachable regardless of where the company was registered.

[B-ROLL: military]

The exit bans were a signal that this was not a routine antitrust review. Standard antitrust asks whether a deal concentrates market power inappropriately. What Beijing was asking was different: whether Chinese-developed AI technology should be allowed to pass into foreign hands at all.

[B-ROLL: courtroom]

That is a national security and industrial policy question, and it is one that China's Foreign Investment Law was specifically built to address. The exit bans were the visible enforcement mechanism for a much broader political judgment that had already been reached.

[B-ROLL: data-center]

The NDRC's formal ruling came under that same Foreign Investment Law, which gives Beijing broad discretionary authority to block foreign acquisitions of assets it considers strategically sensitive. Agentic AI, the category of technology capable of acting autonomously inside complex systems, has now been placed unambiguously in that category.

[B-ROLL: ai-abstract]

A general-purpose agent that can browse, execute code, and complete real-world multi-step tasks is not just a software product. From Beijing's perspective, it is infrastructure, and infrastructure does not get transferred to an American company through a Singapore holding structure and a wire transfer.

[B-ROLL: data-center]

The immediate operational challenge for Meta is unlike anything the company has faced before. By March 2026, the integration between Meta and Manus was not just organizational. It was technical. Around 100 employees were working inside Meta's Singapore offices. Xiao Hong was inside Meta's reporting hierarchy. Systems had been trained, integrated, and developed jointly.

[B-ROLL: code-terminal]

How do you reverse that? What IP belongs to Manus as a stand-alone entity versus what was produced under Meta's ownership? Where do you draw the line on a trained system built using data and compute from both companies simultaneously? The NDRC has provided no answers to any of these questions. It has only issued an order.

[B-ROLL: company-logo:Manus]

There is a customer complication layered on top of that. Earlier in 2026, some Manus enterprise customers publicly distanced themselves from the company after the Meta deal was announced, citing concern about their workflows and proprietary data ending up inside Meta's ecosystem. So the idea that Beijing's ruling simply restores a clean prior state is wrong twice: the technical integration is already deep, and the customer base Manus had before December 2025 is not intact and waiting at the door.

[B-ROLL: ai-abstract]

The competitive implications across the AI industry are real and immediate. Meta's agent strategy had been built around what Manus could provide in the race against Google's Project Astra and OpenAI's growing agent products. That strategy is now disrupted.

[B-ROLL: company-logo:Google]

Google is expected to push harder on its own agent roadmap. Microsoft, which built its Copilot agent infrastructure on top of OpenAI technology inside U.S. and EU jurisdictions, looks comparatively insulated from NDRC exposure. Anthropic and other U.S.-headquartered vendors gain by default as the pool of Chinese-founded acquisition targets available to American buyers gets effectively smaller. OpenAI is in a complicated position: Manus had been outscoring its Deep Research agent on GAIA in a very public way, and the disruption of Manus reduces that competitive pressure in the near term, though not through any product achievement on OpenAI's part.

[B-ROLL: stills:Beijing regulatory offices]

The deeper signal from April 27 is about the limits of corporate restructuring as a geopolitical tool. A company founded in China, built by Chinese founders, with IP developed under Chinese law, does not escape Chinese regulatory reach by updating its corporate address. The founders carried Chinese passports. The original technology was developed in Beijing. The founding networks were Chinese. None of that evaporates when you file new documents in Singapore.

[B-ROLL: news-studio]

Beijing has now shown it will reach through a corporate structure, past the closing date of a transaction, past operational integration, and order the entire arrangement reversed. That is what makes this ruling categorically different from a standard blocked deal. It is not a checkpoint at the border. It is a hand reaching into an already-running operation and pulling out the foundation.

[/VOICEOVER] [TALKING HEAD — sign-off]

If you are building AI in China and thinking about a Western acquisition exit, you are studying what happened to Xiao Hong and Yichao Ji very carefully.

[STAT CARD: "NDRC ruling: April 27, 2026"]

And if you are at a U.S. tech company evaluating Chinese-founded AI startups as acquisition targets, the question is no longer just whether you can get the deal approved. The question is whether the approval can be trusted to hold once the employees have moved, the systems are running, and the integration is real. On April 27, 2026, Beijing answered that question. The answer is: not necessarily.

=== ARTICLE_HTML === === YOUTUBE_DESC === China just pulled off one of the most consequential regulatory moves in AI history. On April 27, 2026, the NDRC issued a single sentence ordering Meta to unwind its $2 billion acquisition of AI agent startup Manus — after 100 employees had already moved into Meta's offices and the CEO was reporting directly to Meta's COO. Sterling Intelligence covers the AI moves that actually matter. Subscribe for weekly deep dives that go beyond the headlines. This episode breaks down the full story of how a Beijing-founded AI company became the flashpoint for a defining U.S.-China tech confrontation — and why the ruling reaches far beyond Meta's balance sheet. Manus wasn't just another AI startup. Its flagship general agent scored 86.5% on GAIA Level 1 — ahead of OpenAI's Deep Research. It reached $100 million in annualized recurring revenue within just 8 months of public launch. By December 2025, Meta paid more than $2 billion to acquire it — valuing Manus at over 4x the $500 million valuation it had carried at its Benchmark-led Series B just months earlier. The problem: Manus was founded in Beijing in 2022 by Xiao Hong, Yichao Ji, and Tao Zhang under a parent company called Butterfly Effect. The team relocated to Singapore in mid-2025 in what the industry calls "Singapore washing" — restructuring your legal domicile to appear beyond Beijing's regulatory reach. It didn't work. China's Ministry of Commerce opened a formal review in January 2026. By late March, co-founders Xiao Hong and Yichao Ji were placed under exit bans, physically prevented from leaving the mainland. On April 27, the NDRC ruled under China's Foreign Investment Law and ordered the deal killed. What makes this ruling extraordinary is the operational depth it's trying to reverse. By the time the ruling landed, Manus had been running inside Meta for months. Around 100 employees were in Meta's Singapore offices. CEO Xiao Hong reported directly to Meta COO Javier Olivan. Joint development had produced systems that may not have a clean ownership boundary. The NDRC offered no instructions on how to untangle any of it — just one sentence. The competitive fallout is immediate. Meta's agent strategy is now in disarray. Google will accelerate its own agent roadmap around Project Astra. Microsoft's Copilot infrastructure, built inside U.S. and EU jurisdictions, looks comparatively insulated from NDRC exposure. Anthropic and other Western-headquartered vendors gain by default as the pool of acquirable Chinese-founded AI assets effectively shrinks. And OpenAI — whose Deep Research agent was being publicly outscored by Manus on GAIA — faces reduced near-term pressure, though not through any product win of its own. The deeper story is what Beijing just signaled to every AI founder and every U.S. acquirer: corporate restructuring cannot insulate Chinese-built technology from Chinese jurisdiction. If you built it in Beijing, it stays in Beijing's reach — regardless of where you file your incorporation papers. ⏱ Chapters: 00:00 - Hook: The NDRC's one-sentence ruling 01:30 - What Manus built and why the industry noticed 03:32 - Singapore washing and the regulatory escalation 05:47 - Operational fallout: unwinding the impossible 08:00 - Sign-off: what this means for AI acquisitions #AI #MetaAI #ManusAI #China #AIAgents #AgenticAI #NDRC #SingaporeWashing #AIAcquisition #TechRegulation #ForeignInvestmentLaw #GAIAbenchmark #AINews #MetaPlatforms #ChinaTech #AIPolicy #OpenAI #DeepResearch #TechMA #AIStrategy === TITLES_HTML ===
  • Top Pick
    China Blocks Meta's $2B Manus Deal, Orders Unwind49 chars
    Factual, search-friendly, captures the core action and the dollar figure in one line.
  • Alternate 1
    Beijing Reached Into a Running $2B Deal and Killed It53 chars
    Drama-leaning; emphasizes the unprecedented nature of reversing an already-integrated acquisition.
  • Alternate 2
    Why Singapore Washing Failed to Protect the Manus Deal54 chars
    Analyst angle; hooks the audience that wants to understand the structural failure and its implications.
    • === KEYWORDS === AI, AI agents, Meta, China, AI news, tech acquisitions, artificial intelligence, Manus AI, NDRC, Butterfly Effect, Xiao Hong, Yichao Ji, Singapore washing, Foreign Investment Law, GAIA benchmark, agentic AI, Javier Olivan, China tech regulation, Meta Platforms, AI acquisition blocked, agentic AI startup, China US tech war, China blocks Meta acquisition, Manus AI deal blocked, AI agent GAIA benchmark, China Singapore washing AI, Meta agentic AI strategy === THUMBNAIL_HTML ===

    Expression. Quietly alarmed, composed — the look of a reporter who just read a one-sentence order that killed a $2 billion deal.

    Head position. Slight forward lean, chin level, eyes direct to camera.

    Wardrobe. Fitted dark charcoal blazer, no jewelry, hair back.

    Eye direction. Direct to lens.

    Lighting. Cool key from upper-left at ~4600K, low fill, no rim; background falls into near-black.

    Scene setup. Austere corridor backdrop, institutional charcoal with faint steel-blue (#2A4D70) gradient edge accent; no competing visual elements.

    Best
    CHINA KILLS THE DEAL

    Position. Top-left, stacked two lines; sub-text bottom-center.

    Font. Extra-bold condensed sans-serif, all-caps; sub at medium weight normal case.

    Color scheme. Headline: #FFFFFF with 2px #FF3B30 drop-shadow; sub: #E0E0E0; background slab: rgba(0,0,0,0.55).

    Accent detail. 3px left-border bar in #FF3B30 running the height of the headline block; sub-text reads "$2 Billion Manus Acquisition BLOCKED".

    Alternate 1
    THEY ALREADY MOVED IN

    Position. Top-right, two-line stack; sub bottom-center.

    Font. Bold condensed sans-serif, all-caps headline; sub at regular weight.

    Color scheme. Headline: #FFFFFF; accent flash: #FFB800; sub: #CCCCCC; backing pill: rgba(0,0,0,0.60).

    Accent detail. Amber (#FFB800) underline stroke beneath headline; sub reads "100 Manus Staff Inside Meta — Beijing Orders Unwind".

    Alternate 2
    SINGAPORE WASHING FAILED

    Position. Bottom-center, centered two-line block with sub above Jane's shoulders.

    Font. Bold sans-serif, all-caps; sub in italic medium weight.

    Color scheme. Headline: #FFFFFF with steel-blue (#2A6DD9) glow; sub: #D0D0D0; background gradient band: rgba(0,0,0,0.65).

    Accent detail. Steel-blue 2px border-bottom on the headline text container; sub reads "Beijing Reached Through the Corporate Structure Anyway".

    === HEYGEN_LOOK === A photorealistic headshot photo of a poised woman in her early 30s, wearing a fitted dark charcoal blazer, hair swept back cleanly, no reflective jewelry, minimal wardrobe. Background: an austere government-corridor atmosphere on a near-black charcoal base, with a faint render of institutional marble columns at roughly eight percent opacity in deep slate grey, a single cold overhead fixture suggesting bureaucratic authority; accent is a steel-blue gradient (#2A4D70) washing the right edge of the frame at low intensity. Mood: quietly-alarmed and measured — the expression of a policy reporter who has just watched a regulator issue a one-sentence order with no appeal. Key light from upper-left at approximately 4600K, low-ratio fill, no rim. Framing: tight headshot, shoulders visible, eyes direct to camera. Ultrarealistic, sharp focus, clean rendering, artifact-free, shallow depth of field. === MOTION_LOWER_THIRD === name: Jane Sterling role: Policy Reporter org: Sterling Intelligence === MOTION_OUTRO === eyebrow: If the receipts hit — main: Subscribe. sub: New episodes every week. No filler. platform1: YouTube handle1: @SterlingIntel platform2: X / Twitter handle2: @SterlingIntel platform3: Newsletter handle3: sterling.ai === MOTION_FUNDING_1 === category: Meta-Manus Acquisition symbol: $ amount: 2 scale: B desc: China's NDRC ordered the deal withdrawn on April 27, 2026 badge: ▼ BLOCKED — Foreign investment prohibited by NDRC === MOTION_STAT_1 === category: NDRC Public Notice value: 1 unit: desc1: Sentences in official ruling — zero named legal basis, no timeline desc2: NDRC · April 27, 2026 badge: No appeal process described === MOTION_STAT_2 === category: Integration Depth value: 100 unit: desc1: Manus employees inside Meta's Singapore offices by March 2026 desc2: CEO reporting to Meta COO · March 2026 badge: Operationally fused — not just signed === MOTION_STAT_3 === category: Meta Platforms (META) value: 2.41 unit: % desc1: Share price decline on NDRC ruling day desc2: Meta Platforms · April 27, 2026 badge: ▼ Market priced a blocked strategy, not just a blocked deal === MOTION_FINANCE_CHART_1 === title: Meta vs S&P 500 — Last 6 Months subtitle: Through the Manus review and NDRC ruling ticker_a: META label_a: Meta Platforms color_a: cyan ticker_b: ^GSPC label_b: S&P 500 color_b: purple period: 6mo footer: Source: Yahoo Finance · April 2026 === MOTION_COMPARISON_1 === benchmark: GAIA Level 1 model_a: Manus score_a: 86.5 model_b: OpenAI Deep Research score_b: 74.3 unit: % source: Source: GAIA Benchmark · March 2025 === MOTION_MULTI_1 === title: Manus GAIA Benchmark val1: 86.5% lbl1: Level 1 val2: 70.1% lbl2: Level 2 val3: 57.7% lbl3: Level 3 === MOTION_MULTI_2 === title: Manus Business Snapshot val1: $39/mo lbl1: Consumer Tier val2: $199/mo lbl2: Pro Tier val3: $100M lbl3: ARR in 8 Months === MOTION_FUNDING_2 === category: Series B — Benchmark led symbol: $ amount: 75 scale: M desc: Valued Manus at $500M — April 2025 badge: ▲ $500M → $2B+ in under 7 months === SOURCES_HTML ===

    Media

    Analyst & Independent

    Prior Context