Meta closed its $2 billion acquisition of AI startup Manus in December promising to 'scale this service to many more businesses.' Three weeks later, the deal's looking more like a liability. Existing customers are already heading for the exits, citing deep mistrust of Meta's data practices. The churn signals a bigger problem for Meta's AI ambitions: in enterprise, trust might matter more than money.
Meta closed its $2 billion acquisition of Manus in late December with grand plans. The company would take the Singapore-based startup's subscription AI agent platform and inject it into the broader Meta ecosystem, reaching small businesses, startups, and everyone in between. But the deal arrived with an uncomfortable problem already baked in: customers don't trust Meta.
Within weeks, existing Manus users started jumping ship. Seth Dobrin, co-founder and CEO of Arya Labs, was one of them. He'd loved Manus as an agentic AI platform - the kind that can independently execute complex tasks like market research, coding, and data analysis. Under its original ownership, Manus offered transparent terms of service. But under Meta? "I'm legitimately sad that this has happened," Dobrin told CNBC.
His concern cuts to the heart of Meta's problem in enterprise AI. "I do not agree with a lot of Meta's practices around data and how they essentially weaponize people's personal data against them," Dobrin said. "I don't want to engage with a company who I don't feel comfortable with how they're going to use data." It wasn't an isolated gripe. Karl Yeh, co-founder of consulting firm 0260.AI that advises startups on integrating AI tools, said he stopped recommending Manus to clients and pulled his own company off the platform entirely.
"Will the data policies of Meta apply to Manus? I would assume it will eventually," Yeh told CNBC. The fundamental uncertainty stung. "We don't know where Manus is going to fit into Meta's AI road map. We're not sure if Manus is going to still remain a separate company even though they said it would." Yeh's hedging to alternatives like Genspark. And he's not alone.
The exodus reveals something Meta executives might not have anticipated when they cut the check: in enterprise AI, institutional trust trumps scale. While Meta swings Google and -sized budgets around - the company announced in June it was to secure talent and stakes in Scale AI - it hasn't cracked the enterprise market.
