The bank said cheaper models such as Kimi K3, Qwen and Llama could broaden enterprise AI use, with 63% of companies already using both open-weight and closed-source systems.
Open-weight AI models such as Kimi K3, Qwen and Llama are likely to speed up the spread of artificial intelligence by lowering costs rather than weakening the competitive position of major industry leaders, Morgan Stanley said on Aug. 4. The bank argued that cheaper AI should encourage broader corporate uptake and cited Jevons paradox, the idea that efficiency gains can increase overall demand. It added that open-weight and closed-source models are likely to coexist over the long term, with 63% of enterprises already using both. Morgan Stanley also said Nvidia stands to be the biggest beneficiary regardless of which AI model architecture ultimately dominates, because demand for AI computing infrastructure is expected to keep rising.