Gcore targets idle AI spend with on-demand NVIDIA Hopper GPU VMs in Europe

Gcore has introduced NVIDIA Hopper-based GPU virtual machines from its sovereign AI region in Portugal, focusing on a practical concern: unused compute costs. The service allows teams to pause GPU billing when workloads stop, while retaining configurations. Aimed at European startups and research groups, the model reflects growing pressure to balance AI performance needs with tighter budgets and stricter data residency requirements.