The computing infrastructure for the LHC data handling (Worldwide LHC Computing Grid-WLCG) was well prepared for Run 2 and delivered a steady data processing since the first collisions. Over the past years including the Run 1 period of the LHC the WLCG developed a unique expertise in building and operating a very large scale infrastructure for unprecedented amounts of LHC-produced data. Together with other research projects the WLCG entered the era of Data Intensive Science. The cost of Computing infrastructures and services reached the level needed for building the LHC detectors. In this contribution, we will present ongoing and planned devel-
opments to evolve the computing models towards the needs for the High Luminosity (HL) LHC, including the disappearance of the distinction between “online” and “offline”, analyses of the physics performance and corresponding computing costs, and inter-operation between commercial cloud services and publicly funded e-infrastructures.