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 OpenAI-sized budgets around - the company announced in June it was spending more than $14 billion to secure talent and stakes in Scale AI - it hasn't cracked the enterprise market.
Flo Crivello, CEO of Manus competitor Lindy, offered a more sympathetic read. His company saw a bump in users after Meta's acquisition announcement, likely from people exploring the entire agentic AI category. He suspects Meta's real play isn't Fortune 500 companies but small businesses and independent contractors - where Manus has focused. The problem: "The way these companies think of these acquisitions, they're acquiring the company for a specific, strategic reason. They just don't know precisely what the integration might look like yet," Crivello said. "They cut a check, it's a new thing they add to the chess board and then they figure it out. And sometimes it takes them years to figure out what to do."
That timeline is a luxury Meta can't afford. Manus said it had reached millions of paying customers and a revenue run rate of more than $125 million at acquisition time. Losing customers within weeks undermines both figures. Meanwhile, Meta's enterprise track record is sobering. The company shuttered its Workplace communication platform in 2024, discontinued Portal video devices, and just announced it's sunsetting Workrooms, its VR collaboration app.
The bigger issue: Meta doesn't have a coherent enterprise AI strategy like OpenAI, Google, or Anthropic do. Navrina Singh, CEO of governance startup Credo AI, observed that Fortune 500 companies in regulated sectors like healthcare and financial services typically build on OpenAI or Anthropic models, then run them through cloud platforms from Microsoft or Amazon - providers with established trust, security, and accountability frameworks. Meta hasn't built that institutional credibility for enterprise.
The one bright spot: WhatsApp. Meta acquired the messaging platform for $19 billion in 2014, and WhatsApp for Business has become genuinely useful for customer interactions. Analysts project it could generate $40 billion in revenue by 2030. That's a product Meta understood and could scale. With Manus, Meta is trying to force-fit an AI startup into an enterprise strategy it hasn't figured out yet.
The Manus customer exodus is really about Meta's bigger problem in AI: it has the cash and the ambition but not the credibility. While OpenAI and Anthropic built enterprise trust through focus and transparency, Meta is trying to buy its way in - spending billions on acquisitions and talent without a clear vision for how they fit together. Dobrin's sadness isn't personal; it's the sound of an enterprise customer watching a promising startup disappear into a company they fundamentally don't trust. That's a harder problem for Meta to solve with money.
