Renewable Energy Output Uncertainty and Risk Transmission in Sustainable Power Supply Chains

Uncertainty in renewable energy output changes how risk is allocated among generators and electricity retailers. We develop a mean–variance model with a thermal plant, a green plant, and a retailer under retailer-led Stackelberg and generation-side Cournot arrangements. In the Stackelberg arrangement, greater retailer risk aversion raises the purchase price and both outputs by compressing the retailer’s unit-margin exposure. In the Cournot arrangement, greater thermal-plant risk aversion lowers thermal and total output but raises green output, the expected purchase price, and renewable share. At the baseline endpoints, green share falls from 55.9% to 51.6% in the former arrangement and rises from 54.8% to 74.1% in the latter. The green plant’s optimal mean output is independent of its own risk coefficient under additive zero-skewness, quadratic-profit, and decision-independent uncertainty conditions; a skewed-shock counterexample shows where this shielding can fail. Local, global, and distributional sensitivity checks and 5000 feasible parameter draws reproduce the analytical directions while showing that magnitudes are calibration dependent. Because renewable share can move opposite to absolute renewable generation, sustainability assessment should report generation levels, shares, procurement prices, and risk allocation together. The model does not estimate emissions, welfare, or reliability.