AI agent startup Manus raises over $500 million at $4 billion valuation
Chinese startup Manus secures funding led by Boyu Capital with Tencent participation, following failed Meta acquisition.
Key points
- Manus raised more than $500 million in a new funding round led by Boyu Capital.
- The company is valued at approximately $4 billion, according to Bloomberg reports.
- Manus 2.0 launched recently with efficiency improvements and new cloud capabilities.
Chinese artificial intelligence startup Manus announced today that it has raised more than $500 million in a new funding round. The investment was led by private equity firm Boyu Capital, with additional contributions from Tencent Holdings Ltd. and several other unnamed investors. The company confirmed the details of the financing in a public statement released today.
Bloomberg reported last month that the transaction would value Manus at $4 billion. This figure represents roughly double the amount Meta Platforms Inc. reportedly offered to acquire the company last December. Meta’s acquisition attempt was abandoned in April after regulators in China blocked the deal, prompting Manus to resume operations as an independent entity.
Company History and Product Evolution
Manus originally launched as Butterfly Effect Co. Ltd. in 2022, initially offering a browser extension that provided access to multiple AI models through a single interface. In 2024, the company raised $10 million and shifted its focus toward developing the standalone AI agent that would later bear the company’s name. This strategic pivot laid the groundwork for the current product iteration.
The Manus agent entered the market last March through an invite-only program. Videos showcasing the agent performing complex, multistep tasks on user computers went viral, attracting significant attention from the venture capital community. This visibility led Benchmark, a prominent venture capital firm, to invest $75 million in Manus in April 2025, coinciding with the company’s relocation of its headquarters to Singapore.
New Features in Manus 2.0
Manus released version 2.0 of its agent just days before announcing the latest funding round. This update introduces a new harness, which consists of software modules designed to customize the behavior of underlying large language models. The company states that this feature reduces token usage by 23.2% for certain tasks while increasing processing speed by more than 28%, addressing key efficiency concerns for users.
The update also includes a tool for generating video clips and games, along with a Cloud Computer feature. Cloud Computer hosts backend code for AI-generated games in a managed environment, allowing the system to run websites and perform long-running tasks even when a user’s local machine is offline. This capability extends the agent’s utility beyond local execution.
Agent Collaboration and Resources
A third major addition to the platform is Cue, a feature that allows users to equip AI agents with resources such as email accounts and digital wallets. Users can also create message boards for their Cue-powered agents, enabling multiple agents to collaborate on complex tasks. This functionality suggests a move toward more autonomous and interconnected AI workflows within the Manus ecosystem.
What should you do?
Security professionals should monitor how AI agents like Manus interact with external resources, particularly when handling email accounts or digital wallets. Organizations adopting such tools should verify the isolation mechanisms used for cloud-hosted workloads to ensure that backend code execution does not expose sensitive data. Reviewing the permissions granted to agents performing multistep tasks is essential to prevent unauthorized actions.
Background: Large language models
Large language models are a specific type of neural network trained on text. Generative AI is the broader umbrella term for any system that creates new content, including images and code. Not all generative AI uses large language models, but most modern text generators do.
What to do and how to stay safe: Manus
- Audit the permissions and resource access granted to AI agents, especially those interacting with email or financial accounts.
- Verify that cloud-hosted workloads are properly isolated from sensitive internal networks and user data.
- Monitor agent activity logs for unexpected multistep actions or deviations from expected task parameters.
- Ensure that any AI-generated content or code is reviewed before deployment in production environments.
Step-by-step guide: Prompt Injection Warning Signs: How to Spot AI Manipulation
General security guidance from the Threat Times newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
Who led the latest funding round for Manus?
Private equity firm Boyu Capital led the investment, with Tencent Holdings Ltd. and other firms contributing.
What is the current valuation of Manus?
Bloomberg reported last month that the deal values Manus at $4 billion.
Why did the Meta acquisition fail?
The acquisition was scrapped in April after Chinese regulators blocked the deal.
