Meta Chief Executive Mark Zuckerberg has called for lower barriers to open-source artificial intelligence as Meta unveiled a new AI model that can run on a per...
Meta Chief Executive Mark Zuckerberg has called for lower barriers to open-source artificial intelligence as Meta unveiled a new AI model that can run on a personal computer. The announcement came as debate continues in the United States over whether access to advanced AI technology should be more restricted and as communities raise concerns about the development of data centres that support AI systems.
Zuckerberg outlined his position in a long essay published on Meta’s website alongside the launch. He said the company would make the weights of its latest AI model available to the public, allowing people to download it. In artificial intelligence, weights are the calculations or rules that determine how an AI system behaves.
The Meta chief executive also said the company would release a new class of open-source models designed to run on a laptop. The move places Meta’s latest AI efforts within a wider debate about how artificial intelligence should be developed, distributed and accessed.
Advertisement
Zuckerberg said in an Instagram video on Monday that Meta would open the weights of its latest model. He argued that artificial intelligence should not be concentrated among a small number of organisations. In his essay, he wrote: “Rather than centralizing superintelligence, we should distribute it widely and give every person the ability to direct it.”
His comments also focused on the balance of power created by different approaches to artificial intelligence. Zuckerberg criticised AI laboratories that are focused on building systems for companies, governments and other institutions. He argued that concentrating AI development in those areas could shift more power towards institutions rather than individuals.
Meta’s approach differs from that of competitors OpenAI and Anthropic, which largely focus on closed AI models. By making the weights of its latest model available for download and introducing models designed for personal computers, Meta is supporting a more open approach to access to AI technology.
The announcement comes as the United States considers how to manage the development and release of advanced artificial intelligence models. US President Donald Trump signed an executive order in June establishing a voluntary federal review of national security risks in advanced AI models before they are released.
Zuckerberg also raised concerns about the competitive position of US artificial intelligence companies compared with foreign laboratories. He said foreign labs have advantages over US companies that have to comply with restrictions on training data. He argued that limiting access to foreign open-source models would not provide an effective solution.
“I do not believe restricting access to foreign open source models is an effective solution,” Zuckerberg said. He added that the objective should be for American open-source models to become the best in the world. “This requires removing the hurdles that make it harder for American open source models to compete,” he said.
The debate over artificial intelligence is also taking place alongside growing public opposition to the construction of data centres in the United States. These facilities support the computing infrastructure needed for AI development. In response to the concerns around communities where Meta operates data centres, Zuckerberg said the company was launching a fund to directly support those communities.
A Meta spokesperson said the fund, launched on Monday, is worth $1 billion. The announcement comes as Meta continues to maintain plans for heavy investment in the data centres and computer chips that support its artificial intelligence work.
Meta expects its capital expenditure to reach as much as $145 billion this year. The figure is nearly double the amount the company spent in 2025 and reflects the scale of its continued investment in the infrastructure supporting its AI efforts.
The latest model launch therefore combines two parts of Meta’s artificial intelligence strategy: making some AI technology more accessible to the public while continuing to invest heavily in the infrastructure needed to develop and operate advanced systems. Zuckerberg’s comments also place open-source AI at the centre of his argument that individuals should have greater access to the technology.
The announcement comes at a time when governments, technology companies and communities are considering different aspects of artificial intelligence, including access to advanced models, national security risks, competition between countries and the development of data centres. Meta’s latest move adds its support to the argument for wider access to open-source AI while the company continues to expand its investment in AI infrastructure.
