Durantic Launches as Managed Operating Layer for Heterogeneous GPU Fleets — Founded by Meta and HRT Infrastructure Engineers

OpenPR·medium signal

Durantic emerged publicly on May 19 as a managed infrastructure layer for organizations that have GPU compute but lack the operational depth to run AI workloads efficiently across customer-owned, colocated, leased, and hybrid environments. Founded by a former Meta Traffic Disaster Recovery team member who spent four years at Hudson River Trading building low-latency bare metal, and a co-founder who was Infrastructure Platform Lead at Replika. As AI infrastructure fragments across hyperscalers, GPU clouds, and sovereign clusters, the operational layer underneath becomes the bottleneck this company targets.