Meta places big bet on autonomous AI — but the toughest test lies ahead

Meta places big bet on autonomous AI — but the toughest test lies ahead

Meta buys Manus in a landmark AI agency deal worth over $2bn. The Singapore-based startup’s general-purpose agent technology could redefine how AI works — but success depends on whether Meta can turn autonomy into measurable business value.


Meta has acquired Singapore artificial intelligence startup Manus in a deal reportedly worth more than US$2 billion, marking one of the largest Asian AI exits to date. The move signals Meta’s ambition to shift from prompt-based chatbots to fully autonomous AI agents capable of performing complex tasks on behalf of users and businesses.

Announcing the acquisition, Manus said it would “continue to operate and sell the Manus service” while its technology and talent are integrated into Meta’s product ecosystem. Its general-purpose agent, launched earlier this year, can independently execute tasks such as coding, data analysis, and market research — capabilities Meta hopes to embed across its consumer and business platforms, including Meta AI.

“Manus has built one of the leading autonomous general-purpose agents that can independently execute complex tasks,” the company said in its statement. “We plan to scale this service to many more businesses.”

Unlike large language models that require specific prompts and human oversight, agentic AI systems are designed to plan and act with limited intervention. They can chain multiple actions together — for instance, analysing market data, drafting a strategy, and producing a report — without the user micromanaging each step. In practical terms, these agents act more like digital employees than conversational assistants.

This autonomy has made the field one of the fastest-growing areas in generative AI investment. Meta’s move follows similar initiatives from OpenAI, Google, and Anthropic, all of which are developing agent frameworks to automate higher-order business processes.

But the value of such technology will hinge on what Meta does next, said Bill Conner, President and CEO of Jitterbit. “This acquisition is attempting to prove that AI agents can generate returns, not just burn capital,” he said. “The real test won’t be ownership, branding, or demos — it will be whether Meta can successfully integrate Manus’ AI agents into its products to materially change user behaviour and deliver measurable value.”

Conner warned that even though Meta has said Manus will continue to run independently, integration challenges are likely. “Unlocking AI value sounds great on paper, but the real challenge isn’t in acquiring data infrastructure, it’s in actually using it,” he added.

His comments reflect broader investor caution: after years of hype, AI ventures are now expected to demonstrate return on investment, governance, and secure delivery. The so-called “agent sprawl” — a proliferation of overlapping AI tools — risks eroding user trust unless companies can show tangible outcomes.

The acquisition highlights how Southeast Asian AI innovation is becoming a target for global platforms. Manus’s rapid growth, reportedly serving millions of users and processing more than 147 trillion tokens, highlights the region’s emerging role in foundational AI development.

However, the deal could draw regulatory scrutiny beyond Singapore. Reports indicate that Chinese authorities are reviewing whether the sale involves export-sensitive technology, underscoring the geopolitical sensitivity surrounding general-purpose AI systems.

For Meta, the integration of Manus’s team and technology may determine whether its AI ambitions can move beyond consumer chat experiences and into enterprise-scale automation. For the market, it offers a clear test of whether autonomous agents can deliver real economic value or remain a costly experiment in digital independence.



  • TUI holds outlook as customers book later

    TUI holds outlook as customers book later

    TUI has maintained guidance despite weaker third-quarter profit and bookings. Customers are making travel decisions later as geopolitical disruption and consumer caution reshape demand, while the group continues to expect €1.1bn–€1.4bn of annual underlying EBIT.


  • Zero-hours rules could cost employers £2.9bn annually

    Zero-hours rules could cost employers £2.9bn annually

    New analysis puts zero-hours reform costs into much sharper focus. Employer costs could reach £2.9bn annually depending on how ministers implement guaranteed hours, shift notice, and cancellation-payment rights.


  • Maeving secures £3m backing for export growth

    Maeving secures £3m backing for export growth

    Maeving has secured £3m financing to expand overseas motorcycle sales. HSBC UK funding backed by UKEF will support increased production for American and European demand, with 13 new Coventry jobs planned.