Meta CEO Mark Zuckerberg has outlined an ambitious plan to transform the company's digital ecosystem. During the second-quarter 2026 earnings call, he stated that in the coming months and years, personal AI agents will become the foundation of Meta's products and revenue. These systems will work on behalf of users around the clock, performing tasks to improve health, finances, relationships, and achieve personal goals.

From Code to Mass Adoption

Zuckerberg noted that AI agents first gained popularity in the programming sector. Developers with technical expertise are already actively using such tools to configure and optimize code. However, for Meta's billion-strong audience, this is not enough. The company aims to create a consumer product that works "out of the box" and does not require deep technical knowledge from the user.

In the future, according to the head of Meta, people will interact with multiple agents simultaneously. Key interfaces for these services will be WhatsApp and other company messengers. This will allow artificial intelligence to be integrated into daily communication and routine tasks.

Business Agents and New Monetization Models

The company has already begun implementing tools for businesses. Meta Business Agent became available worldwide on WhatsApp and Messenger in the second quarter of 2026. According to Zuckerberg, more than 1 million companies use this tool weekly. Meta is now deploying similar solutions on Instagram as well.

In the long term, the company plans to monetize AI services through several channels:

  • Subscriptions for premium agent features;
  • Payment based on the volume of computational resources used;
  • Models where businesses pay for the results of the agent's work.

Zuckerberg compared the potential monetization mechanics to Meta's advertising auction, which has already brought the company billions of dollars.

Record Investments in Infrastructure

The announcement of new products came against the backdrop of a sharp increase in Meta's spending on AI infrastructure. In the second-quarter report, the company stated that capital expenditures, including financial lease payments, amounted to $31.08 billion. Meta also raised the lower bound of its capital expenditure forecast for 2026: the expected range is now $130–145 billion, compared to the previous $125–145 billion.

Free cash flow for the quarter dropped to $784 million from $8.55 billion a year earlier. Following the release of the report, Meta's stock fell nearly 10% in after-hours trading, reflecting investor concerns about the return on such massive investments.

The Battle for Superintelligence

Alongside technical plans, Zuckerberg published a column in The Wall Street Journal, arguing against the concentration of future superintelligence in the hands of a small number of companies or institutions. In the industry, superintelligence refers to a hypothetical AI system that surpasses humans in a wide range of tasks. Meta uses the term "personal superintelligence" to describe AI that should work in the interests of a specific user.

"Healthy societies also recognize that security requires checks and balances. If only a few institutions possess superintelligence, they will inevitably exert controlling influence over the economy, science, and politics. Even with the best intentions, this concentration will limit people's ability to choose their own future," Zuckerberg wrote.

In his view, with the spread of superintelligence, more jobs will be created, and starting a business will become easier without attracting large amounts of capital. "The development of superintelligence will be the deepest technological progress we will see in our lifetime," he concluded.

Context and Competition

In April, Meta introduced Muse Spark — the first model of the new Meta Superintelligence Labs lineup, which is intended to be part of the movement toward personal superintelligence. In July, Zuckerberg also spoke out against the blocking of Chinese AI models in the US, believing that such restrictions will not help American companies maintain their advantage. Previously, Meta, Nvidia, Microsoft, AMD, and other companies urged authorities not to restrict the development and distribution of AI models with open weights.