The Shocking Block: China's NDRC Halts Meta's Major AI Deal
In a move that has sent ripples through the global technology sector, China's National Development and Reform Commission (NDRC) announced on April 27, 2026, that it has officially blocked Meta Platforms' $2 billion acquisition of Manus AI, a promising Singapore-based startup specializing in autonomous artificial intelligence agents. This decision comes after months of intense scrutiny and represents a significant escalation in Beijing's efforts to safeguard its burgeoning AI industry from foreign influence.
Manus AI, known for its innovative 'action engine' technology that enables AI to execute complex tasks autonomously, had been a rising star. The deal, initially announced in late December 2025, was seen as a strategic win for Meta in its race to dominate the agentic AI space—a field where AI systems not only think but act independently on user commands, from research to workflow automation.
The NDRC's statement cited national security concerns, invoking laws on foreign investment security reviews. This unwind order requires both companies to terminate the transaction, marking one of the most high-profile interventions in cross-border tech mergers involving Chinese-linked firms.
Understanding Manus AI: The Tech at the Heart of the Controversy
Founded by Chinese nationals Xiao Hong and Ji Yichao, Manus AI emerged from the remnants of Butterfly Effect, a Beijing and Wuhan-based entity. Relocating to Singapore in a bid to operate globally, Manus developed a browser-based sandbox environment allowing users to oversee AI agents in real-time, much like supervising an intern. This setup enables the AI to handle multi-step tasks, outperforming benchmarks set by competitors like OpenAI's DeepResearch in generating detailed reports and automating workflows.
The startup's core product, Manus, bridges the gap between ideation and execution. Users input goals, and the agent navigates web tools, data sources, and applications to deliver results. Its traction was rapid: by mid-2025, it sparked an AI agent boom in China, with applications in research, e-commerce, and personal productivity. Valued at over $2 billion, the acquisition promised Meta access to this cutting-edge tech amid its push for Llama-powered agents.
Despite assurances from Meta that no Chinese ownership would persist post-deal, regulators viewed the Singapore structure as insufficient to sever ties, labeling it a potential 'Singapore washing' tactic to bypass controls.
Timeline: From Announcement to Abrupt Halt
- December 2025: Meta announces the $2 billion acquisition, highlighting Manus as key to agentic AI advancements.
- January 2026: Chinese authorities launch a probe into potential violations of technology export controls and investment security laws.
- March 2026: Co-founders Xiao Hong and Ji Yichao are barred from leaving China amid ongoing review, raising alarms in tech circles.
- April 2026: Reports emerge of Beijing planning broader restrictions on US investments in AI firms.
- April 27, 2026: NDRC formally blocks the deal, ordering its unwind.
This sequence underscores the protracted regulatory process, transforming what seemed a done deal into a cautionary tale.
China's Rationale: Safeguarding AI Sovereignty
Beijing's decision stems from heightened national security priorities in artificial intelligence, viewed as a cornerstone of future economic and military power. The NDRC emphasized that the acquisition posed risks to core technologies, potentially allowing sensitive algorithms and data to flow to a US adversary amid escalating tensions.
China's AI strategy, outlined in recent policies, prioritizes self-reliance. With the US imposing chip export bans, Beijing fears reverse brain drain—talented engineers and proprietary models migrating abroad. Manus's dual structure, with roots in China but operations in Singapore, triggered reviews under the Foreign Investment Law and updated security guidelines for critical sectors like AI.
Experts note this aligns with a pattern: similar blocks on foreign buys in semiconductors and biotech, now extending to generative and agentic AI.
Photo by Aleksey Kashmar on Unsplash
Meta's AI Ambitions Disrupted
For Meta, the loss stings. CEO Mark Zuckerberg has pivoted heavily toward open-source AI agents via Llama models, aiming to rival Google and OpenAI. Manus's tech would have accelerated this, providing ready-to-deploy autonomous capabilities for Meta AI assistants on Facebook, Instagram, and WhatsApp.
A Meta spokesperson expressed disappointment but affirmed commitment to ethical AI development. The company may appeal or seek alternatives, though unwinding could involve refunds and IP disentanglement. Stock dipped 1.2% post-announcement, reflecting investor concerns over supply chain vulnerabilities in global AI talent.
Internally, Meta's $40 billion+ AI capex in 2026 now faces recalibration, potentially boosting in-house agent R&D.
Industry Reactions: Shockwaves in Silicon Valley and Shenzhen
Tech leaders voiced concerns. Venture capitalist Benchmark, an early Manus backer alongside Tencent, called it a 'wake-up call' for cross-border deals. Analysts at Goldman Sachs predict a 15-20% chill on US-China AI M&A, pushing startups toward domestic funding.
In China, state media praised the move as protecting 'core tech sovereignty.' Founders now hesitate on offshore relocations, fearing extraterritorial reach. Singapore's tech hub status is questioned, as regulators there cooperated with Beijing.
Quotes: 'This redraws the map for AI globalization,' said Wedbush analyst Dan Ives.
US-China Tech Rivalry Intensifies
This block fits a larger pattern. The US accuses China of 'industrial-scale' AI IP theft; Trump administration vows crackdowns on Chinese exploitation of US models. China counters with investment curbs, requiring approval for US capital in top AI firms.
Statistics: China's AI patents surged 30% YoY to 88,000 in 2025; US leads but gap narrows. Agentic AI market projected at $50B by 2030, making control pivotal.
Geopolitical stakes: AI underpins military apps like autonomous drones, cyber ops—hence dual-use security flags.
| Aspect | US Position | China Position |
|---|---|---|
| Investment Rules | Entity List, chip bans | Security reviews, exit bans |
| AI Patents 2025 | ~120,000 | 88,000 (+30%) |
| Market Share Goal | Lead AGI | World-class by 2030 |
Implications for Global AI Startups
Startups face a bifurcated world: Chinese firms risk founder detentions for foreign sales; US ventures navigate export controls. Alternatives emerge—Europe, India—but scale lags.
In China, domestic giants like Baidu, Alibaba ramp agentic AI investments, with Tencent pledging higher 2026 capex despite chip curbs. Policy signals: 'Local-first' AI ecosystem, favoring state-approved open-source.
For founders: Dual HQ risky; full relocation insufficient without tech transfer audits.
Photo by Bernd 📷 Dittrich on Unsplash
What's Next: Legal Battles and Policy Shifts
Meta may challenge via arbitration; Manus could pivot to Asian buyers. Beijing expected to formalize AI investment blacklist by Q3 2026.
Optimists see negotiation room—China supports 'law-abiding' deals. Pessimists warn of full decoupling.
Watch: Upcoming US AI safety bill, China's GAI pathways plan.
Future Outlook: A Divided AI Landscape
The Manus saga heralds fragmented AI development: US open innovation vs. China's controlled ascent. Benefits—diverse models; risks—slowed global progress, higher costs.
For investors: Diversify beyond borders. For users: Expect specialized agents, slower universal adoption.
Ultimately, this underscores AI as national strategic asset, where geopolitics trumps commerce.