The GPU-as-a-Service (GPUaaS) solution integrates WEKA‘s advanced data storage software with Nebius‘ full-stack AI cloud platform. The collaboration enables customers to scale compute and storage resources on demand with ultra-high performance and microsecond latency for efficient AI model training and precision AI inference.
Organizations that run AI model training and inference processes at scale often face challenges related to compute, memory, storage, and data management, which can impede innovation. As demand for modern AI infrastructure grows, organizations are embracing specialized neoclouds for access to turnkey infrastructure that can power their AI ambitions. In turn, neoclouds are seeking innovative ways to optimize the performance and efficiency of their GPU data center infrastructure.
Nebius AI Cloud delivers a cutting-edge, cost-optimized neocloud environment that empowers innovators of all sizes, from enterprises to startups to research institutions to operationalize AI workloads. To fuel the premium tier of its next-generation platform, Nebius selected WEKA’s high-performance storage software to turbocharge its AI Cloud performance while effortlessly scaling from petabytes to exabytes of data.
A leading research institution selected Nebius’ purpose-built AI infrastructure to power its large-scale experimentation and model development efforts, reserving a multi-thousand-GPU cluster and leaning on Nebius AI Cloud’s developer-friendly platform optimized for AI/ML workloads.
To further tailor the environment to its exacting operational needs, the customer requested the integration of WEKA’s data platform, citing previous success with WEKA and the need for features such as user and directory quotas. With 2PB of WEKA storage deployed alongside Nebius’ compute infrastructure, the institution now benefits from a high-performance, scalable, and fully managed platform that supports the rigorous demands of cutting-edge AI research.
“WEKA exceeded every expectation and requirement we had,” said Danila Shtan, CTO at Nebius. “The WEKA solution not only delivers outstanding throughput, IOPS, and low latency at scale while effortlessly managing mixed read and write workloads, but it also provides exceptional metadata management and streamlined multitenancy.”
“We are proud to be collaborating with Nebius to deliver high-performance, cloud-based solutions that maximize their AI innovation while minimizing infrastructure complexity,” said Liran Zvibel, cofounder and CEO at WEKA. “Together, Nebius and WEKA are redefining what’s possible when high-performance storage meets AI-first infrastructure, providing a unified solution that is a catalyst for enterprise AI and agentic AI innovation.”
