The potential of generative artificial intelligence (GenAI) in open source threat intelligence, especially in grid security, was explored in this paper. Large language model (LLM) as an innovative tool coupled with its powerful natural language processing capabilities, is expected to play a key role in the production of open source threat intelligence, improve intelligence quality and operational efficiency, and bring smarter and more efficient solutions for cybersecurity defense. In this study, a practical model example of using LLM for open source threat intelligence collection was described in detail, demonstrating that the potential of introducing LLM can significantly enhance the security and responsiveness of the current power monitoring system in the absence of a unified threat intelligence system. Despite the powerful information processing capabilities of LLMS, the realization of highly accurate, highly specialized, and unified structured models requires long-term, in-depth professional training. With the maturity of LLM model in open source threat intelligence field, power system will usher in the era of truly automated threat intelligence production.