Artificial intelligence is playing a growing role in satellite image processing and in the remote sensing market, as deep learning methods are improving the performance of many processing and inference tasks. This talk will overview recent advances in the field, covering tasks related to image generation (e.g., image denoising/despeckling and super-resolution) as well as image classification, event prediction and compressive imaging. It will also discuss key aspects in the application of deep learning to onboard satellite image processing.