Meta Unveils New AI as Zuckerberg Pushes Open Source

Meta’s new AI model arrived Monday with a clear political message: the United States should do more to help homegrown open-source developers keep pace with China. CEO Mark Zuckerberg used the launch to press Washington to lower barriers he says are holding American companies back in the race for AI leadership.

Zuckerberg argued that U.S. policy should be designed to make American open-source models the strongest in the world. In a manifesto titled “The Future is for Everyone: The Path to a Positive AI Future,” he wrote that the United States and its allies need to lead the open source AI ecosystem because it could account for a large share of global AI use. He also said foreign labs have an edge because American labs must follow additional restrictions on training data.

The remarks reflect a widening policy and business debate around artificial intelligence, one in which questions about openness, competition and national security are increasingly intertwined. Open-source and open-weight models have become especially important because they can be widely adopted, modified and embedded into commercial products more easily than tightly controlled systems. For companies building applications on top of AI, those qualities can translate into faster development and lower costs.

Meta’s move comes as Chinese open-weight models are gaining traction with U.S. businesses. Companies in the United States have been drawn to them because they can deliver increasingly competitive performance at a lower cost than some American alternatives. That shift has added urgency to the debate in Washington, where policymakers are weighing new restrictions over national-security concerns while American firms continue to face pressure to innovate quickly.

Among the Chinese firms drawing attention are DeepSeek, Alibaba and Z.ai. Their models have emerged as low-cost alternatives, and some U.S. tech companies have begun embracing them for their flexibility and cheaper computing costs. In an industry where access to computing power remains a major expense, the appeal of models that can perform well without the same price tag has helped boost their profile.