Meta’s AI trajectory hinges on sustained momentum, regular launches, and monetization beyond ads amid competition from OpenAI, Anthropic, and Google.

Meta has begun offering AI-related subscriptions and aims to expand beyond advertising, but concerns persist about whether Muse Spark can deliver meaningful revenue quickly.

Meta spent more than $14 billion in 2025 to recruit Scale AI founder Alexandr Wang and his team, launching Muse Spark in April as Meta’s first major proprietary foundation model and signaling a shift away from open-weight approaches.

The pivot followed sharp criticism of the Llama open-weight strategy and disappointing developer uptake from late 2023 into 2024, triggering leadership changes and a move toward internal deployments and platform integration.

Muse Spark, released in April, represents Meta’s move to own a proprietary foundation model designed to integrate with Meta apps and devices and to unlock new revenue streams beyond ads.

Industry voices suggest developers currently favor competing models—notably Google's—over Muse Spark, raising questions about trust and the value of third-party access to Muse Spark’s underlying tech.

Developer trust has weakened, with limited access to Muse Spark and lower engagement than during the Llama era.

Investors and developers remain skeptical, seeking tangible AI-driven revenue beyond ad efficiency, even as Meta’s stock has fallen about 18% over the past year despite Q1 2026 revenue growth.

Some observers advocate a path focused on computational efficiency and cost reductions rather than sheer model size, though broad third-party developer support remains a question.

Meta says it will open Muse Spark’s API and underlying tech to outside developers, with early partner testing and plans for wider release.

Internal pressures and morale challenges are evident, including layoffs across multiple divisions, as Meta’s AI leadership debates how to translate investments into sustained revenue.

Ultimately, Muse Spark’s credibility hinges on monetization and a robust developer ecosystem to validate Meta’s AI bets.