The paper introduces SuperCluster, a novel method for efficient and scalable 3D panoptic segmentation of large point clouds. The approach reformulates the panoptic segmentation task as a scalable graph clustering problem, eliminating the need for resource-intensive instance-matching during training. SuperCluster is trained using local auxiliary tasks and can be adapted to the superpoint paradigm, further enhancing its efficiency. This allows the model to process scenes with millions of points and thousands of objects in a single inference. SuperCluster achieves state-of-the-art performance on two indoor scanning datasets (S3DIS and ScanNetV2) and sets new benchmarks on large-scale mobile mapping datasets (KITTI-360 and DALES). With only 209k parameters, SuperCluster is over 30 times smaller than competing methods and trains up to 15 times faster. The method's scalability, efficiency, and precision make it a promising solution for large-scale 3D analysis in various applications, including digital twin creation and geospatial analysis.The paper introduces SuperCluster, a novel method for efficient and scalable 3D panoptic segmentation of large point clouds. The approach reformulates the panoptic segmentation task as a scalable graph clustering problem, eliminating the need for resource-intensive instance-matching during training. SuperCluster is trained using local auxiliary tasks and can be adapted to the superpoint paradigm, further enhancing its efficiency. This allows the model to process scenes with millions of points and thousands of objects in a single inference. SuperCluster achieves state-of-the-art performance on two indoor scanning datasets (S3DIS and ScanNetV2) and sets new benchmarks on large-scale mobile mapping datasets (KITTI-360 and DALES). With only 209k parameters, SuperCluster is over 30 times smaller than competing methods and trains up to 15 times faster. The method's scalability, efficiency, and precision make it a promising solution for large-scale 3D analysis in various applications, including digital twin creation and geospatial analysis.