GreeNN is an optical neural network designed with the goal of reducing the carbon footprint of image classification. Increasing the complexity of computations for image classification in traditional silicon based hardware is associated with a significant energy cost. The team’s solution implements the most computationally intensive task of image classification in free space optics with high energy efficiency. The setup encodes an image onto a green laser beam using a digital micromirror device and performs a convolution calculation on a spatial light modulator. Throughout the year, the concept has been developed and tested on handwritten digit images, showing potential for application on more sophisticated datasets.