WEKA builds data infrastructure software for the world’s most demanding accelerated computing environments. Its technology underpins AI training and inference, high-performance computing, life sciences, financial services, and engineering workloads at national laboratories, research institutions, neocloud providers, and Fortune 500 enterprises — wherever data has to move fast enough to keep GPUs busy.
Founded in 2013, WEKA began with a straightforward conviction: traditional storage architectures were never designed for the scale and performance demands of modern unstructured data and compute-intensive workloads, and incremental improvement would not close the gap. What started as a software-defined parallel file system has grown, over more than a decade of research and development and 150+ patents, into a company shaping how data infrastructure is built for AI at scale.
Today, WEKA’s NeuralMesh™ architecture applies a containerized microservices approach to storage, interconnecting data, compute, and AI services across edge, core, on-premises, hyperscale cloud, and neocloud environments from a single software foundation. It is deployed by organizations training frontier models, operating AI factories, and running large-scale HPC — and it runs on standard hardware, giving European research and industrial computing initiatives a portable, vendor-neutral data layer.