Meta CTO Andrew Bosworth said the new AI lab delivered key models internally this month, Reuters reported. Enterprise AI demand is colliding with data governance expectations, forcing procurement slowdowns from policy ambiguity.
Meta is accelerating its AI build internally, but its enterprise ambitions face a trust gap highlighted by customer pullback after acquisitions
Before this update, Meta was spending heavily on AI and trying to define an enterprise strategy while competing with model leaders. Early enterprise platform adoption often tightened around data rights language, not model release cadence. Meta is investing heavily in AI to compete with OpenAI, Google, and Anthropic. Enterprise buyers often require clear data handling and governance commitments for AI tools.
Meta Superintelligence Labs delivered first high-profile AI models internally this month, Bosworth said. Some Manus customers say they stopped using the product after Meta agreed to buy the company. CNBC reported Meta planned to scale Manus subscription AI agent service to more businesses, per Axios. "very good," said Andrew Bosworth, Chief Technology Officer at Meta Platforms, according to Reuters (via Yahoo Tech).
Internal model progress strengthens Meta capabilities, but customer distrust around data use can limit enterprise adoption of acquired products as scaling starts. Zoom clarified terms in Aug. 2023 after AI training fears, showing how perceived overreach can force revisions. CNBC described customer concerns about Meta data practices while Meta planned broader Manus subscriptions. Meta pointed to a Manus blog post promising continued subscription operations from Singapore. The same promise sits beside uncertainty about Manus future under Meta, per customer comments.
The Reuters internal delivery claim raises product supply, while CNBC documents demand leakage tied to data policy fears. Zoom called its ToS language a process failure, showing how wording can become the demand shock. Meta gets almost all of $200 billion annualized revenue from ads, keeping data practices inside buyer evaluation. Manus said it reached millions of paying customers with a revenue run rate above $125 million. That run rate collides with customer reports of switching, leaving subscription scaling exposed to churn.
Whether Meta data policies will formally apply to Manus product data and when remains unclear. The scale of Manus customer churn and revenue impact after the deal stays undisclosed. Customer distrust around data use constrains enterprise adoption of acquired products. Bosworth only said models were very good, limiting any quality comparison. Meta cannot be judged losing the enterprise AI race from anecdotal customer quotes alone. How formal data policy scope, churn magnitude, and continuity commitments evolve will determine Manus subscription scaling over the next 6–18 months.
