The paper develops a model of costly competitive adjustment where firms reposition capacity in response to free cash flow shocks, analyzing strategic crowding and firm repositioning within a language-model embedding space. Building on Xu et al. (2022), the authors represent news articles and analyst reports from Nasdaq-listed companies (2018–2022) as dense vector clouds using large language models, capturing the information geometry of firm dissimilarity over time. A novel sandwich estimator evaluates the sign and magnitude of predicted capacity adjustments, validated through multi-agent simulation. The framework links publicized investment decisions — covering capital structure, product mix, leadership, risk exposure, and operating procedures — to shifts in discounted free cash flows for competitors and investors, offering a quantitative method to measure strategic interactions via semantic representations of financial disclosures.