Clean renewable distributed energy resources (DERs), such as wind and photovoltaic (PV) power generation, are being installed worldwide to reduce the carbon emissions generated by fossil fuel power plants. Their intermittencies hinder their integration into the power grids. Various technologies, such as energy storage systems (ESS) and microgrids, are developed for effective grid integration of DERs. However, achieving a high percentage installation and grid integration of DERs remains a challenge primarily because of the low feed-in tariffs. To resolve the issue, virtual power plants (VPPs) are introduced to aggregate and coordinate large numbers of DERs, ESS, and microgrids. Recognising the potential business opportunity, many power retailers have established centralised VPPs to increase DER installation and financial resilience. While these VPPs can provide adequate buffers to power price volatility (as demonstrated in the recent power price surge) and improve the distribution network resilience, the centralised coordination fails to provide sufficient incentives to attract the enthusiastic participation of DER, ESS and microgrid owners. This talk presents our vision and recent studies of VPPs for high percentage integration of DERs in smart grids, including machine-learning-based energy management and internal market operations.