Meta Platforms has released a new downloadable artificial intelligence model capable of running on a personal computer, as chief executive Mark Zuckerberg published a 6500-word essay laying out his vision for a future where AI power is placed in the hands of individuals rather than corporations and governments.
The move deepens an already heated debate in Washington and Silicon Valley over whether increasingly powerful AI technology should be freely available or kept under tighter control by a small number of organisations.
What is the new Meta AI model?
The new model, called Muse Glimmer, is a distilled version of Meta's more powerful Muse Spark 1.2 model. It has been engineered for efficiency, meaning it places lower demands on computing hardware than its parent system. At 30 billion parameters, it is small enough to operate on a single graphics card.
Meta says Muse Glimmer is designed primarily for agent-like AI tasks — things like schedule management and file organisation. The model's weights, which are the internal values that drive an AI system's decision-making, will be made freely available through the Hugging Face collaboration platform at no cost, though users will need to cover their own computing expenses. Meta has also indicated it plans to release weights for a version of the more powerful Muse Spark model.
Shares in Meta rose around one per cent in New York trading following the announcement.
What did Zuckerberg argue in his essay?
Zuckerberg's lengthy essay positions Meta squarely against rivals such as OpenAI and Anthropic, which largely favour keeping their AI systems proprietary and under tighter controls. Without naming those companies directly, he criticised those "focused on building AI for companies, governments, or other institutions."
"Rather than centralising superintelligence, we should distribute it widely and give every person the ability to direct it," Zuckerberg wrote, adding that such an approach could usher in "a new era of personal empowerment where individuals can use this powerful new capability to reach their full potential."
He described Meta's ambition as creating "personal superintelligence for everyone."
The open AI debate and China's challenge
The timing of Meta's release is significant. Low-cost open AI models from Chinese companies — including DeepSeek, Moonshot and Alibaba — have intensified competitive pressure on US firms that charge for access to their technology. Moonshot AI's Kimi K3 is among the latest models drawing attention from policymakers and industry figures alike.
Some US policymakers have explored imposing restrictions on open-weight models originating from China. That prospect has drawn pushback from technology industry leaders, including Nvidia chief executive Jensen Huang, who argue that open-weight models benefit the long-term development and security of AI. Zuckerberg aligned himself with that view.
"I do not believe restricting access to foreign open-source models is an effective solution," he wrote. "Our goal should be for American open-source models to be the best globally."
The Trump administration has so far held back from restricting open AI systems. Under a newly completed AI safety framework, cutting-edge models would be subject to a 30-day voluntary safety review window before release — but open-weight models have been explicitly excluded from that testing requirement for now.
What happens next?
Zuckerberg used his essay to caution against what he called a "rigid process and review timeline" for AI safety, arguing instead for closer collaboration between government and industry. Meta has spent years promoting open-weight AI, having previously released a series of models under the Llama name to mixed reviews.
As the debate over AI regulation intensifies and the future of AI-driven technologies becomes central to economic policy discussions globally, Meta's latest move signals the company intends to keep pushing open access as a defining principle — and a competitive strategy.
