Unleash AI Performance on
Oracle Cloud

NeuralMesh™ runs on a cluster of optimized bare metal Compute shapes (BM.Optimized3.36) with local SSD to create a high-performance storage layer. NeuralMesh software can then extend the single namespace to an OCI Object Storage bucket for high-capacity storage at the lowest cost. You can easily add more OCI compute shapes or more OCI Object Storage to scale the WEKA namespace up to support Exabytes of capacity or millions of IOPs of storage performance and scale back down when capacity or performance is not needed. Data stored with your WEKA environment is accessible to applications in your environment through multiple protocols, including NFS, SMB, POSIX, NVIDIA GPUDirect, and S3.

Augmented Memory Grid™ extends GPU high-bandwidth memory into a persistent, petabyte scale token warehouse, eliminating the memory wall that limits long-context and agentic AI workloads. Built on NeuralMesh and deployed with NeuralMesh Axon, Augmented Memory Grid continuously streams key-value cache data between GPU memory and flash storage using RDMA and GPUDirect Storage, leveraging OCI’s bare metal GPU infrastructure to achieve memory-speed performance without adding physical DRAM. This breakthrough eliminates costly token recomputation when GPU memory fills, delivering 20x faster time to first token, 1000x more KV cache capacity, and up to 7.5M read IOPs—enabling AI cloud providers, model builders, and enterprises to maximize GPU utilization, reduce inference costs, and profitably serve long-context AI applications.