Meta CEO Mark Zuckerberg outlines a vision for “personal AI agents” that act on a user’s behalf, as the company continues investing in AI infrastructure. Speaking during Meta’s Q2 2026 earnings call, Zuckerberg previews how the company is thinking about these agents and what they could do for people. TechCrunch reports that Zuckerberg predicts billions of people will have personal AI agents within roughly five years, framing the effort as a major long-term bet that will require significant spending now. The Verge similarly characterizes Meta’s approach as an upcoming push into personal agents and notes Zuckerberg’s high-level preview of the concept during the call. Across both accounts, the focus is on scaling agent capabilities so they can perform tasks for individuals, and on convincing investors that the investment in agents will pay off over time. Neither source provides specific product names, timelines beyond the multi-year horizon, or detailed technical specifications, but both describe the same executive messaging: Meta is preparing to expand from broader AI research into user-facing personal agent systems.