Large Language Models in equity markets: applications, techniques, and insights

Recent breakthroughs in Large Language Models (LLMs) have the potential to disrupt equity investing by enabling sophisticated data analysis, market prediction, and automated trading. This paper presents a comprehensive review of 84 research studies conducted between 2022 and early 2025, synthesizing the state of LLM applications in stock investing. We provide a dual-layered categorization: first, by financial applications such as stock price forecasting, sentiment analysis, portfolio management, and algorithmic trading; second, by technical methodologies, including prompting, fine-tuning, multi-agent frameworks, reinforcement learning, and custom architectures. Additionally, we consolidate findings on the datasets used, ranging from financial statements to multimodal data (news, market trends, earnings transcripts, social media), and systematically compare general-purpose vs. finance-specialized LLMs used in research. Our analysis identifies key research trends, commonalities, and divergences across studies, evaluating both their empirical contributions and methodological innovations. We highlight the strengths of existing research, such as improved sentiment extraction and the use of reinforcement learning to factor market feedback, alongside critical gaps in scalability, interpretability, and real-world validation. Finally, we propose directions for future research, emphasizing hybrid modeling approaches, architectures that factor reasoning and large context windows, and robust evaluation frameworks to advance AI-driven financial strategies. By mapping the intersection of LLMs and equity markets, this review provides a foundation and roadmap for future research and practical implementation in the financial sector.