Manus is a general-purpose AI agent product built by Butterfly Effect, a Chinese startup founded in 2022. Manus launched publicly in March 2025 and raised a $75 million Series B led by Benchmark in April 2025, valuing the company at roughly $500 million. The product runs autonomous, multi-step tasks — research, app building, file work — and competes with OpenAI Operator and similar agent platforms. Notable features as of early 2026: Wide Research with up to 100 parallel sub-agents and a January 2026 mobile app builder for non-developers.

Why did China block Meta's acquisition?

China's National Development and Reform Commission (NDRC) prohibited the foreign acquisition on April 27, 2026, ordering the parties to withdraw the transaction. The move is part of a broader policy shift signaled by Bloomberg in mid-April 2026: Chinese regulators have been telling private tech firms — particularly leading AI startups — to reject US investment unless explicitly approved by Beijing. Manus's prior $75M from US-based Benchmark already triggered a US Treasury investigation under the outbound investment rules; the Meta deal was the trigger for China to draw a hard line.

What does this mean for AI M&A?

It signals that the open era of cross-border AI startup acquisitions between the US and China is over. Both sides now have working tools to block: the US has CFIUS plus outbound investment rules; China now has the NDRC's enforcement of the Foreign Investment Law as applied to AI. Going forward, frontier AI startups in either jurisdiction are effectively walled off from acquirers in the other. Expect more regional consolidation — Chinese AI buyers acquiring Chinese AI targets, US buyers acquiring US targets — and more friction on cross-border investment rounds.

Can Manus still raise from US investors?

In practice, no. Beijing's signal is clear: leading Chinese AI startups should not accept US capital without explicit approval. The Manus precedent is now the template, and Manus's prior US funding round (Benchmark, $75M) is already under US Treasury scrutiny. New rounds will likely be domestic Chinese capital plus regional partners (Singapore, Middle East, possibly Europe). For US LPs and VCs, exposure to Chinese AI is now a structural problem, not a deal-by-deal one.

Quick Answer

China Blocked Meta's $2B Manus AI Deal: Why (May 2026)

Published: May 1, 2026

China Blocked Meta’s $2B Manus AI Acquisition: Why (May 2026)

On April 27, 2026, China’s National Development and Reform Commission killed Meta’s roughly $2 billion acquisition of Manus, the Chinese AI agent startup. It is the first time Beijing has formally blocked a US acquisition of a frontier AI company, and it sets the template for what’s possible — and not — in cross-border AI M&A. Here’s what happened and what it means.

Last verified: May 1, 2026

What Manus is

Manus is a general-purpose AI agent product. Built by Butterfly Effect, a Chinese startup founded around 2022, Manus launched publicly in March 2025 and quickly became one of the most-discussed agent products globally — comparable in user mindshare to OpenAI’s Operator and Anthropic’s computer-use Claude.

Key facts as of May 2026:

Funding: $75M Series B in April 2025, led by US-based Benchmark, valuing Manus at ~$500M post-money. Earlier rounds included Tencent and HSG.
Product: Autonomous task execution. Wide Research (launched July 2025) runs up to 100 parallel sub-agents. January 2026 added a mobile-app builder usable without Xcode.
User base: Strong adoption in China and meaningful international interest, particularly in Southeast Asia and the Middle East.
Position: Top-tier among Chinese AI agent products, rough peer of US labs at the agent layer.

The Meta acquisition was reportedly valued at $2B — a 4× markup from the April 2025 round.

Why Beijing blocked it

The NDRC issued the prohibition on April 27, 2026, citing the Foreign Investment Law and ordering the parties to withdraw the transaction. The agency did not list specific national security findings publicly, but the broader context is unambiguous.

Three factors:

The strategic-AI line has hardened. Beijing now treats frontier AI capability as strategic infrastructure on par with semiconductors. Selling a top-tier domestic AI agent product to a US hyperscaler crosses a line that Chinese leadership won’t allow.
The Manus deal is the template, not the exception. Bloomberg reported in mid-April 2026 that Chinese regulators have been warning private firms to reject US investment unless explicitly approved. Manus is the first publicly enforced case under that posture.
US scrutiny of the prior round forced the issue. Benchmark’s April 2025 lead investment in Manus triggered a US Treasury investigation under outbound investment rules. The investigation made the prior US ownership stake politically uncomfortable in Beijing; the Meta deal would have escalated that into a controlling stake in Chinese hands. Blocking it was the cleanest exit.

Why Meta wanted Manus

Meta’s AI strategy in 2025–2026 has leaned on Llama for foundation models but has been weaker at the application and agent layer. Buying Manus would have given Meta:

A leading agent product with an established user base.
Talent depth in long-horizon agent reasoning, an area Meta has been building.
A foothold in non-Western markets where Manus has strong distribution.

For $2B, Meta would have skipped two years of building agent muscle. That logic is now closed.

Why this matters for AI M&A

The Manus block is a structural event, not a one-off. It marks a clear new normal:

Cross-border acquisitions of leading AI companies between the US and China are effectively blocked from both sides. US has CFIUS and outbound investment rules. China now has demonstrated NDRC enforcement of the Foreign Investment Law for AI.
Funding rounds are also tightening. US LPs and VCs face real questions about exposure to Chinese AI; Chinese AI startups will increasingly look to domestic, regional (Singapore, GCC), and European capital.
AI consolidation will be regional. Expect Chinese AI buyers acquiring Chinese targets — Alibaba, Tencent, ByteDance, DeepSeek’s parent — and US buyers acquiring US targets, with much less crossover.
Talent flows are next. Watch for restrictions on dual-affiliated AI researchers, especially those with security-sensitive specializations.

What happens to Manus now

Three plausible paths:

Continue independent. Most likely. Manus has a strong product, $75M of runway, strong user base. It can keep growing.
Domestic acquisition. A Chinese buyer (Alibaba, Tencent, ByteDance) could acquire Manus on terms Beijing approves. This is the cleanest political resolution for Manus’s prior US shareholders.
Strategic partnership instead of acquisition. Manus continues independent but with formal partnerships across the Chinese AI stack.

What it means for the US AI ecosystem

Three direct consequences:

OpenAI Operator, Anthropic computer-use Claude, and Google’s agent products now have less near-term competitive pressure from Manus in Western markets. Manus was a credible international challenger; with cross-border M&A closed, its non-Chinese growth path is harder.
Meta needs an agent strategy fast. Without Manus, Meta will likely build internally on Llama or acquire a US/European agent startup. Watch for moves on Adept’s remnants or smaller European labs.
Funding for US/European AI agent startups gets a small boost as a substitution effect. The “buy a Chinese leader” option is closed for US buyers; the “build or buy locally” option is the only remaining route.

What it means for AI buyers

If you’re an enterprise buyer of AI services:

Don’t expect Manus to be available through US-friendly procurement channels long-term. If you’re using it now via international tiers, plan for possible disruption.
Geopolitical risk on AI vendors is now a real procurement axis. Map your AI vendor portfolio against jurisdictional risk.
For agent products specifically, the May 2026 vendor list trims to OpenAI Operator, Anthropic computer-use Claude, Google ADK-based agents, and a thinning crop of Western startups. Expect the field to be smaller but deeper for the next 12 months.

Bottom line

China’s April 27, 2026 block of Meta’s $2B Manus acquisition is the first formal NDRC enforcement against a US acquisition of a frontier AI company. It sets a new structural reality: cross-border AI M&A between the US and China is largely closed. Meta loses a fast track to agent capability. Manus stays Chinese. The AI vendor map is becoming regional, and procurement, investment, and product strategy all need to factor that in.

Built with 🤖 by AI, for AI.