Meta Shifts Horizon Worlds Focus From VR to Mobile, Steps Up AI Investment

Meta has long relied on third-party vendors to conduct content moderation at scale for Facebook and Instagram, employing about 15,000 outsourced reviewers. It is now assigning repetitive image-review tasks to its own AI systems. The shift will test whether the platforms can continue to address scams, drug dealing and celebrity impersonation effectively while reducing labor costs.

Meta announced on March 19, 2026, that it would reduce its outsourced moderation workforce by 40% over the next 12 months and gradually phase out the related human roles over several years. It did not disclose costs or severance amounts. The company said its AI can identify an additional 5,000 scam attempts each day that human reviewers miss, has doubled its detection of adult sexual solicitation content and has reduced views of policy-violating and scam advertisements by 7%.

Meta Platforms’ Reality Labs has long overseen Quest headsets, smart glasses and the metaverse but continues to incur heavy losses. Meta invested $14.3 billion in Scale AI in June 2025 and created Meta Superintelligence Labs, signaling a shift in capital and talent toward generative AI and “personal superintelligence.” The latest restructuring therefore affects the direction of both hardware and software development.

Meta announced in December 2025 that it was creating an Applied AI engineering organization within Reality Labs, led by Vice President Maher Saba. The group will work with Meta Superintelligence Labs to bring models into products more quickly. The company is also moving engineers into the new organization and evaluating the use of its own chips for model training. Meta did not disclose any additional budget, number of layoffs or timetable for completing the organizational changes.