GPUs are only as fast as the data feeding them. WEKA NeuralMesh is a software-defined storage and memory platform purpose-built for AI, scaling linearly from ten nodes to ten thousand and running the same software on-premises, in the cloud and everywhere between. XENON has deployed WEKA for Australian AI research since the platform’s early days, including at Deakin University’s Applied Artificial Intelligence Institute, where epoch times fell from around 40 minutes to 6. XENON also deployed the first WEKApod in APAC in 2026.
AI Factories
WEKApod Nitro appliances are certified for NVIDIA DGX SuperPOD, pairing naturally with XENON’s SuperPOD certification for APAC. WEKA’s Augmented Memory Grid extends GPU memory into a persistent token warehouse, with WEKA citing 1,000 times more KV-cache capacity and up to 6.5 times higher token throughput for inference and agentic AI. NeuralMesh Axon embeds the data layer directly inside GPU servers.
HPC
NeuralMesh delivers parallel-filesystem performance, up to 162 GB/s and 2 million IOPS to a single client per WEKA’s benchmarks, with zero tuning and no metadata bottlenecks. InfiniBand and Ethernet, MPI-IO, HDF5 and NetCDF are all supported, with automated tiering to object storage keeping costs in check.
HFT
WEKA publishes dedicated solutions for high-frequency and quantitative trading: consistent ultra-low latency through peak market activity, and parallel access that cuts backtesting times dramatically, with WEKA citing 66% lower database run times versus local NVMe.
Data Management
One exabyte-scale namespace with POSIX, NFS, SMB, S3, GPUDirect Storage and Kubernetes CSI access ends redundant data copies. Contractually guaranteed data reduction of up to 6x, encryption in flight and at rest, multitenancy, and Snap-to-Object snapshots to any object store cover backup, DR and cloud bursting.
Services
XENON designs, deploys and supports WEKA end to end, on WEKApod appliances or on qualified server hardware XENON builds, and can operate the platform as part of a managed AI or HPC environment.
Proven at exabyte scale worldwide. Talk to XENON’s experts about accelerating your data.