To design weather-resilient electricity systems that rely on variable renewable energy sources (VRES), spatially-explicit modeling with several decades of historical weather data is necessary to avoid over-optimization to a single weather year. Here, we apply the electricity system model EXPANSE for Switzerland to search for resilient cost-optimal and near-optimal system designs in 2035 under various policy scenarios, using data from 25 historical weather years (1995–2019) for electricity demand, solar photovoltaics (PV), wind power, hydropower dams and run-of-river hydropower. In particular, we demonstrate how modeling to generate alternatives (MGA) could be used to define a range of potential system designs for reaching different levels of weather resilience. We find that solar PV capacity of 35 GW in Switzerland in 2035 without any wind power would allow the country to reach its target of 35 TWh yr ⁻¹ on new renewable electricity in 95% of the 25 historical weather years. Further limiting winter electricity import to a maximum of 5 TWh yr ⁻¹ as currently envisioned by policy, would require additional 10 GW of wind power to have a comparable level of weather resilience. Within the range of alternative MGA scenarios, such 95% resilience level could also be achieved using 12 GW of solar PV and 18 GW of wind power, 20 GW of solar PV and 13.5 GW of wind power. This methodology could be transferred to other countries and continents that intend to reach high shares of VRES in their electricity systems.