A recent McKinsey survey shows that more than half of organizations globally are now using open source AI technologies in their operations.
The report, “Open Source Technology in the Age of AI,” surveyed over 700 technology leaders and developers across 41 countries.
According to the findings, over 50 % of organizations employ open source AI tools for data management, AI models, and developer workflows. In the technology sector specifically, around 72 % of companies reported active use of open source AI models.
Across multiple layers of the AI tech stack, 58 % of organizations used open libraries for data and feature engineering, 63 % ran open models in production (rising to 72 % in tech companies), 51 % embedded open components in chatbots, copilots, or analytic dashboards, and 55 % employed open orchestration frameworks.
The report also highlights workforce trends: approximately 81 % of developers stated that experience with open source AI tools is important for their careers.
Cost efficiency is another key factor cited: organizations report that open source AI solutions are often less expensive to implement and maintain compared to proprietary alternatives.
Despite widespread adoption, the survey notes that security, regulatory compliance, and intellectual property concerns remain among the main challenges when using open source AI technologies.
Despite the challenges, the demand for such tools is rising. The 2026 Consumer Electronics Show (CES), that took place in Las Vegas between January 6 and January 9, proved that many tech companies are among the most active supporters of open-source innovation in the sphere of artificial intelligence (AI). Nvidia CEO Jensen Huang noted that the advancement of open models ensures that digital intelligence does not “leave anyone behind.”
The McKinsey research concludes that over 75 % of surveyed leaders expect their organizations to expand the use of open source AI tools in the coming years.
Another recent McKinsey report showed that 88% of organizations use AI in at least one business function, but only 23% say they are scaling at least one most innovative type of AI applications – agentic system somewhere in their enterprise
