Tapping into a large amount of renewable generation considering the inherent variability of renewable energy sources (RES) can greatly increase the risk of supply and demand imbalances in electric power delivery. One of the major components of this risk is the intermittency of both wind and solar power generation. In this paper, we show that by strategically planning for geographical and technological diversification of renewable generation capacity it is possible to reduce such risk in a RES-only US energy portfolio. We consider wind and solar as the sole sources of generation and use risk-averse stochastic optimization with Conditional Value-at-Risk (CVaR) to optimize energy generation locations and capacities in an idealized case study. The optimal RES portfolios demonstrate a significant improvement in generation profile compared to non-pooled or non-optimized alternatives. This confirms that with smart policy planning one can push the limits of the risk of imbalances in RES-only portfolios within continental United States, and highlights the need for system-wide thinking when designing a large-scale energy portfolio.