Meta Releases Muse Glimmer AI and $1 Billion Community Fund
Meta Platforms released the open-source Muse Glimmer AI model and established a $1 billion fund for communities hosting its data centers.
Meta Platforms Inc. released Muse Glimmer, a distilled 30-billion parameter AI model designed to run on personal computers with a single graphics card. Available via Hugging Face, the model focuses on agent-like tasks such as schedule management and file organization. Meta also plans to release open weights for Muse Spark 1.2, while continuing to offer the original Muse Spark as a paid, closed model. The company is developing a more powerful model, internally named Watermelon, for future release.
Chief Executive Mark Zuckerberg accompanied the launch with a 6,500-word essay, "The Future is for Everyone," advocating for the democratization of personal superintelligence. Zuckerberg argues that distributing AI to billions of individuals prevents power from concentrating within a few corporations or governments. He envisions 24/7 AI agents, interfaced through smart glasses, that assist users with health, finances, and career management. This open-source strategy contrasts with the restrictive safety approaches favored by rivals OpenAI and Anthropic.
To mitigate local resistance to the resource-intensive nature of its infrastructure, Meta established a $1 billion fund for U.S. communities hosting its data centers. This initiative follows a model in Richland Parish, Louisiana, where tax revenues funded teacher bonuses. Zuckerberg also urged the U.S. government to reduce regulatory friction and accelerate energy infrastructure to maintain a competitive lead over China. Financially, Meta plans up to $145 billion in capital spending this year, potentially reaching $600 billion by 2028 for data center expansion.