This paper presents a hybrid forecasting framework that combines historical commodity price data with semantic signals from global economic news, extracted using an agentic generative AI pipeline. The architecture integrates dual-stream Long Short-Term Memory (LSTM) networks with attention mechanisms to fuse structured time-series inputs with semantically embedded, fact-checked news summaries covering 1960 to 2023. The agentic pipeline employs a manager agent that coordinates a news specialist and a fact-checking agent to autonomously retrieve, summarize, and verify yearly economic news, which are then embedded and reduced via PCA before fusion with price data.

Evaluated on a 64-year dataset of normalized commodity prices and aligned news embeddings, the model achieves a mean AUC of 0.94 and accuracy of 0.91, significantly outperforming traditional baselines such as logistic regression (AUC=0.34), random forest (AUC=0.57), and support vector machines (AUC=0.47). Ablation studies reveal that removing the news component causes a steep drop in AUC to 0.46, underscoring the critical value of incorporating real-world context through unstructured text. The findings demonstrate that integrating agentic generative AI with deep learning can meaningfully improve early detection of commodity price shocks, offering a practical tool for economic planning and risk mitigation in volatile market environments.