Meta Plans Major Push into Personal AI Agents

Meta is gearing up to make a significant push into personal AI agents that can perform tasks on users' behalf. During Wednesday's Q2 2026 earnings call, CEO Mark Zuckerberg outlined the company's vision: soon there will be agents that work 24/7 to help achieve goals and improve life—health, relationships, finances, and more.

Zuckerberg noted that agents have so far taken off in coding, but engineers are more technical and willing to spend time making them work. To build great personal agents, they must be great consumer products that work out of the box and are easy enough for billions of people to adopt. He seems to differentiate this vision from companies like Anthropic and OpenAI, which focus on coding and enterprise agents. Google has also pushed into personal AI agents with its Gemini Spark tool.

However, Meta may face an uphill battle. Its AI efforts still lag behind larger labs. Unlike Google or Microsoft, it lacks ecosystem access to people's email or documents. Additionally, Meta is dealing with trust issues related to smart glasses, and general distrust might deter users from adopting its agents.

Zuckerberg said the company will share more about personal agents “soon,” calling them “the foundation for our next wave of products and revenue lines.” Meta already offers business agents on WhatsApp and Messenger, used by over 1 million businesses weekly, and is rolling them out to Instagram.

This move follows last quarter's launch of the Muse Spark AI model, recently updated with version 1.1 for better coding. Meta has restructured to focus on AI, moving 7,000 staff to new AI initiatives, though it laid off about 8,000 people in May. The company also plans $130–$145 billion in capital expenditures in 2026 and yesterday announced a 1-gigawatt data center campus with BlackRock.

During the call, Zuckerberg also said Instagram now has over 2 billion daily active users.