Morgan Stanley says open-weight AI models could speed adoption by cutting costs

Morgan Stanley said the rise of open-weight AI models such as Kimi K3, Qwen, and Llama is likely to accelerate AI adoption by lowering costs, rather than weakening the competitive position of industry leaders. The bank said cheaper AI should encourage broader enterprise use and cited the Jevons paradox, the idea that improvements in efficiency can lead to higher overall demand instead of less consumption. The firm also expects open-weight and closed-source models to coexist over the long term. It said 63% of enterprises already use both types of models at the same time, pointing to a mixed deployment approach rather than a winner-take-all shift. Morgan Stanley added that Nvidia stands to be the biggest beneficiary regardless of which AI model architecture eventually becomes dominant. Its view is based on continued growth in demand for AI computing infrastructure, which the bank sees as a key underlying driver as model usage expands across enterprises.