Multi-Objective Portfolio Optimization for Energy Infrastructure Investment Under Uncertainty

C. Fraces, G. Moog

Abstract

The energy investment landscape is undergoing a structural transformation. On the demand side, artificial intelligence infrastructure, data center proliferation, and transportation electrification are driving load growth at rates unseen in decades. On the supply side, transmission bottlenecks, interconnection queues, and permitting delays constrain the pace at which new capacity can be brought online. Capital must be allocated across a widening universe of technologies—conventional oil and gas developments, geothermal baseload, utility-scale solar and wind, and battery storage—each with distinct risk profiles, revenue structures, and policy exposures. In this environment, the unit of decision is shifting from the individual project to the portfolio: the relevant question is no longer whether a given project clears a hurdle rate, but how a configuration of projects performs jointly under uncertainty. Yet the prevailing practice in capital allocation remains sequential, project-by-project approval. Each opportunity is evaluated in isolation against deterministic or lightly risked economics, ranked, and funded until the budget is exhausted. This practice systematically ignores the correlation structure that binds projects together—shared commodity price exposure, common regulatory regimes, geological dependencies among adjacent prospects, and shared infrastructure. As a result, correlated macro risks are underestimated, diversification benefits are left unquantified, and portfolios drift toward concentrated exposures that no individual project assessment would reveal (Adams et al., 2001; Ball and Savage, 1999; Savage, 2009).

Source: semanticscholar

Read the AI summary and key takeaways for traders on WOBR Quant Research.