Mozilla's inaugural "State of Open Source AI" report, based on analysis and a global survey of over 950 developers, indicates that open-source models have reached a performance parity with proprietary systems like ChatGPT and Claude, with the gap narrowing to approximately 3%. While model costs have plummeted by up to 50x over the last three years, a significant economic disparity persists: open models account for roughly one-third of real-world AI usage but capture only 4% of total revenue. Geopolitically, East Asia leads in adoption at 89%, while 47 countries have implemented restrictions on foreign processing for critical workloads. Despite high interest driven by privacy and cost, a deployment gap exists; while 79% of developers use open models, only 51% have deployed them in production, compared to 63% for closed models. This discrepancy is attributed to insufficient infrastructure and tooling rather than model quality. Furthermore, the report highlights a critical shift toward "agentic harnesses"—the software layer managing model interactions—which currently suffers from low oversight, with users approving AI agent requests by default up to 93% of the time. Ultimately, the findings suggest that for open source to remain viable, investment must shift from raw model capability toward robust deployment infrastructure, governance, and ownership of the agentic software layer.