Weka Uses Flash Storage to Replace Costly GPUs

Weka has launched NeuralMesh 6, a software platform paired with its first self-designed hardware line, Wekapod 3, designed to reduce GPU memory strain in AI workloads. The system uses cheap NAND flash storage to cache pre-calculated tokens, eliminating redundant recomputation in long context windows and multi-turn conversations. The technology, called Augmented Memory Grid, aggregates flash storage to behave like GPU memory at a fraction of the cost. Weka enters a competitive market alongside Dell, NetApp, Pure Storage, and VAST, all of which have repositioned toward AI infrastructure in recent years.