AI-Driven Intellectual Property Portfolio Optimization and Concentration Risk – Evidence from Global Consumer Brands

Marlena Jankowska, M. Pawełczyk, Sebastian Dziarmaga-Działyński

Abstract

This paper demonstrates that artificial intelligence (AI)-driven optimization of intellectual property (IP) portfolios affects on firm-level concentration risk, revenue volatility and stock price crash risk. While AI enhances the efficiency of IP allocation decisions, we argue that algorithmic optimization may systematically increase portfolio concentration, thereby amplifying downside risk exposure under demand uncertainty. We develop a theoretical framework that extends Modern Portfolio Theory (MPT) to IP asset portfolios and introduces an AI-driven allocation mechanism. Using patent portfolio data from global consumer brands, including Nike, Pop Mart, Huawei, and Xiaomi, we construct IP concentration indices (Herfindahl-Hirschman Index), estimate revenue volatility, and model downside risk using semivariance and negative conditional skewness (NCSKEW). Monte Carlo simulation and sensitivity analysis reveal a fundamental tension between AI-optimized efficiency and portfolio resilience. AI-driven IP allocation increases concentration in high-expected-return technology segments, which reduces idiosyncratic forecasting error but amplifies tail risk during demand shocks. The study makes three contributions: it bridges IP management and financial risk modeling, introduces an AI-induced concentration risk channel, and provides actionable evidence for corporate risk managers and policymakers navigating the intersection of AI deployment and IP strategy in an uncertain global economy.

Source: semanticscholar · PDF

Read the AI summary, key takeaways and discussion on WOBR Quant Research.