WEKA and Oracle Cloud Infrastructure validate 10x throughput increases for long-context AI inference

Common OCI H100 infrastructure benchmarks showed 10x more concurrent users, 10x higher token throughput, and 7x more tokens deployed without adding GPUs

CAMPBELL, CA., June 9, 2026 /PRNewswire/ — WEKAthe AI ​​data and storage infrastructure company, today announced production-scale benchmarks that show how companies can improve the economics of long-context AI inference by serving more users and tokens on the same GPU footprint. The benchmarks show that WEKA’s NeuralMesh™ platform with Augmented Memory Grid™ To Oracle Cloud Infrastructure (OCI) serves ten times more concurrent users, delivers ten times higher token throughput, and produces seven times more tokens per GPU than pure DRAM configurations without additional infrastructure. Results were validated on a nine-node OCI bare metal H100 cluster with 100,000 token context windows.

“Enterprise AI workloads push contextual windows and GPU utilization to new limits,”…