WEKA has announced the launch of its next-generation WEKApod appliances, designed to address cost and performance challenges in data center storage for artificial intelligence (AI) and high-performance computing. The redesigned WEKApod Prime and new WEKApod Nitro target data center operators, cloud providers, and engineers deploying demanding AI workloads.
WEKApod Prime features a mixed-flash configuration that combines triple-level cell (TLC) and enterprise triple-level cell (eTLC) flash drives with WEKA’s AlloyFlash technology. This approach, according to WEKA, results in 65 percent better price-performance by automatically optimizing data placement, maintaining full write performance, and avoiding the penalties associated with typical tiered storage systems. Prime supports up to 20 drives in a 1U rack or 40 drives in a 2U form factor. WEKA claims efficiency gains of up to 4.6 times higher capacity density, five times better write input/output operations per second (IOPS) per rack unit, and 68 percent lower power consumption per terabyte compared to the previous generation, supporting workloads such as AI training and checkpointing.
WEKApod Nitro is built for large-scale AI deployments and GPU clusters, enabling two times faster performance and 60 percent better price-performance over prior models. It utilizes updated hardware including NVIDIA ConnectX-8 SuperNIC with 800 Gb/s throughput and 20 TLC drives in 1U. Both appliances are supplied as turnkey systems, certified for NVIDIA DGX SuperPOD and NVIDIA Cloud Partner solutions, reducing integration time for rapid deployment.
Both appliances integrate WEKA’s NeuralMesh storage software in pre-validated configurations. Features include distributed data protection, instant snapshots, multi-protocol access, automated tiering, encryption, and hybrid cloud support. The systems are designed to start with as few as eight servers and scale to hundreds, addressing expansion needs while minimizing integration effort.
The Danish Centre for AI Innovation reports direct benefits from deploying WEKApod, with Nadia Carlsten, CEO, stating, “Space and power are the new limits of innovation in data centres. WEKApod’s exceptional storage performance density allows us to deliver hyperscaler-level data throughput and efficiency within an optimised footprint—unlocking more AI capability per kilowatt and square metre.” Ajay Singh, Chief Product Officer at WEKA, added, “AI investments must demonstrate ROI. WEKApod Prime delivers 65% better price-performance without compromising on speed, while WEKApod Nitro doubles performance to maximize GPU utilization. The result: faster model development, higher inference throughput, and better returns on compute investments that directly impact profitability and time-to-market.”
Next-generation WEKApod appliances are available immediately. Further details are provided on the WEKA website.
Source: WEKA
