**[Opinion: Open-Source Models Account for Only 10% of Enterprise Spending, But Mature Production Environments Will Be Dominated by Open-Source Models]** According to monitoring by Beating, public opinion often touts that open-source large models are taking over everything, but enterprise spending data paints a different picture. Jesse Zhang, co-founder and CEO of the enterprise-grade AI customer service platform Decagon, pointed out that the share of open-source models in total enterprise large model spending has dropped to 11%. This decline stems from the fact that most enterprises' AI applications are still in the early exploratory stages, defaulting to closed-source models. However, he emphasized that once application scenarios mature, open-source models will dominate production environments due to their advantages in ultra-low latency and deep fine-tuning. In Decagon's own production environment, 90% of call volume has already transitioned to open-source weight models. The core driver of this shift is interaction speed and customization capability, rather than cost savings. In customer service scenarios, a single conversation delayed by 8 seconds can completely ruin the product experience. Since leading closed-source labs do not allow fine-tuning of flagship models, and smaller closed-source models cannot be deeply customized, small-scale open-source models fine-tuned for specific tasks have become the only option to support high-frequency real-time interactions. In the future, enterprise AI will exhibit a division of labor: leading closed-source labs will continue to dominate exploration and discovery in new fields, while open-source weight models will increasingly take over actual production in mature business scenarios. Due to the high demands of data and talent for model fine-tuning, the migration from closed-source to open-source will be a slow process spanning several years, with both seeing sustained growth during the transition period. [Original Link]