Morningstar

Meta: With Muse Personal AI Agent, Firm Offers First Glimpse at Revamped AI Products and Fortunes

We like the firm’s current trajectory, with both the Muse Spark model and the Muse agent showing clear, tangible progress.

Meta logo is displayed during the Viva Technology show.
Chesnot via Getty
Securities in This Article
Meta Platforms Inc Class A
(META)

Key Morningstar Metrics for Meta Platforms

  • Fair Value Estimate
    : $850.00
  • Morningstar Rating
    : ★★★★
  • Morningstar Economic Moat Rating
    : Wide
  • Morningstar Uncertainty Rating
    : High

On Sept. 8, 2026, Meta Platforms META released Muse, a personal AI agent that can interact with the digital world and help users complete real-world tasks, such as booking travel, making purchases, and managing communication. Shares of Meta are up more than 6% today.

Why it matters: After spending most of the last year playing catch-up with the major AI labs, Meta’s product and AI model velocity is picking up. We like the firm’s current trajectory, with both the Muse Spark model and the Muse agent showing clear, tangible progress.

  • We see Muse, which can be accessed via WhatsApp and as a stand-alone app, as a clear upgrade to other personal agents available to users, including Gemini Spark and Instinct.
  • To encourage adoption, Muse has a generous 100 million weekly token free tier, with two further tiers ($20/$100 per month) for users who want more. We expect Muse monetization to evolve, with transaction fees and revenue sharing with businesses as potential levers.

The bottom line: We maintain our $850 fair value estimate for wide-moat Meta. We see the launch of Muse, as well as the impending launch of Watermelon, the firm’s frontier AI model, as further proof points for our monetization thesis for Meta.

  • Much like with Alphabet in 2025, we believe the market is underappreciating Meta’s multivariate monetization opportunity in AI. We see the opportunity spread across digital ads, personal intelligence, enterprise adoption, and selling excess compute to other labs.
  • We see two scenarios. One: Meta has frontier models and can charge enterprises for them. Two: Meta has excess capacity that it can sell to external AI labs. In either case, billions of dollars in Meta’s top line are missing from the market’s estimates.

Coming up: We expect the upcoming launch of Watermelon, a frontier model, to be an important near-term milestone in shifting the perception of Meta as an AI laggard.

Editor’s Note: This analysis was originally published as a stock note by Morningstar Equity Research.

The author or authors do not own shares in any securities mentioned in this article. Find out about Morningstar’s editorial policies.

Sponsor Center