Meta’s Manus Deal Signals Shift in AI Data Centre Strategy

Meta has started 2026 by reinforcing its AI strategy with the acquisition of Manus, a Singapore-based developer of general-purpose autonomous AI agents. 

The deal, reportedly valuing Manus at more than US$2bn, marks one of Meta’s largest AI-focused purchases to date and reflects a shift from pure model development towards deployable software that can be monetised at scale.

For data centre operators and infrastructure investors, the transaction underlines how hyperscalers are seeking to extract value from the vast compute capacity already built to support AI training and inference. 

Meta’s global data centre estate has expanded rapidly in recent years as it invested tens of billions of dollars into AI infrastructure. The Manus acquisition signals a clearer path to turning that capacity into revenue-generating services.

The deal values Manus at roughly four times its April 2025 valuation of US$500m when it raised funding led by Benchmark. It also continues a broader trend across the AI sector where the focus has moved from foundation models towards agent ecosystems capable of executing real tasks for enterprise users.

Manus and the execution layer

Founded out of Singaporean startup Butterfly Effect in early 2023, Manus has emerged as a notable player in the agentic AI space. Its platform is designed to reason, plan and act autonomously, breaking down complex objectives into multiple steps and executing them with limited human oversight.

Manus is positioned as an execution layer that sits on top of large AI models. It can conduct research, analyse data, write and refine content, automate workflows and even operate virtual machines.

The platform is aimed primarily at enterprise subscribers and is used for activities ranging from market research and analytics to coding and automation.

“Joining Meta allows us to build on a stronger, more sustainable foundation without changing how Manus works or how decisions are made,” says Xiao Hong, CEO of Manus.

“We’re excited about what the future holds with Meta and Manus working together and we will continue to iterate the product and serve users that have defined Manus from the beginning.”

For data centres, this type of software is tightly coupled with infrastructure demands.

Agentic platforms require persistent access to compute, storage and high-bandwidth connectivity, particularly as they scale across enterprise customers.

Monetising AI infrastructure

Murthy Grandhi, Company Profiles Analyst at GlobalData, describes the acquisition as a clear signal of changing priorities.

“Meta is making one of its clearest wagers yet on autonomous AI, agreeing to acquire Manus in a move that underscores how quickly the AI race is shifting from foundation models to deployable software agents that execute real work,” he says.

“The purchase gives Meta a functioning business with paying customers, meaningful revenue and infrastructure already proven at scale.”

Murthy points to the imbalance between Meta’s infrastructure spending and its AI revenues to date.

“For Meta, the strategic logic is straightforward. The company has committed tens of billions of dollars to AI data centres and model development, but monetisation has lagged its infrastructure ambition.

“Manus offers a ready-made, high-margin software layer that can be sold directly and integrated across Meta’s consumer and enterprise products.”

This alignment is significant for the data centre sector.

Hyperscale campuses built to support model training are increasingly being optimised for inference-heavy workloads tied to deployed AI services. Platforms like Manus provide a clearer utilisation model for that capacity.

Implications for data centre demand

Meta is expected to keep Manus operating as a standalone service while embedding its agent capabilities into Meta AI and its wider portfolio including Facebook, Instagram and WhatsApp. 

Murthy believes this will create “both immediate revenue and longer-term platform leverage”.

For infrastructure planners, this suggests sustained demand for resilient, high-density data centre environments capable of supporting always-on autonomous systems at global scale. 

It also reinforces the importance of energy efficiency and network performance as agentic workloads move closer to end users.

The acquisition reflects a maturing phase in the AI economy, where value is increasingly derived from execution rather than experimentation. 

As hyperscalers like Meta connect their expanding data centre footprints to deployable AI agents, the emphasis shifts towards operational platforms that can consistently turn compute into billable services.