Netweb Technologies has announced the launch of Tyrone ParallelStor Velox, a new unified data platform designed to overcome critical data bottlenecks in AI, HPC, and enterprise environments. By integrating flash, disk, tape, and cloud storage into a single global namespace, the platform aims to significantly improve GPU utilization and accelerate model training. This innovation is specifically engineered to meet the growing demand for sovereign, high-performance computing infrastructure in India.
Addressing the Modern AI Data Bottleneck
On April 29, 2026, Netweb Technologies unveiled Tyrone ParallelStor Velox, a solution created to solve one of the most pressing challenges in contemporary computing: the data bottleneck. As organizations deploy increasingly powerful GPU clusters, traditional storage architectures have struggled to maintain the necessary data velocity, leading to underutilized compute resources and increased operational complexity.
High-Performance Features
The new platform functions as a high-performance data backbone, unifying fragmented storage environments across various tiers including flash, disk, tape, and cloud. Key technical advantages include:
- High-throughput data pipelines tailored for efficient GPU cluster operation.
- Seamless support for NVIDIA GPUDirect Storage, which enables direct data transfer to GPU memory and minimizes CPU overhead.
- Multi-protocol access including POSIX, NFS, SMB, S3/Swift, and Hadoop on a single dataset.
- Intelligent policy-driven lifecycle management to optimize storage performance and costs.
Supporting Sovereign AI Infrastructure
According to Swastik Chakraborty, VP of Netweb Technologies, the effectiveness of any AI infrastructure is entirely dependent on the quality of its underlying data layer. Tyrone ParallelStor Velox is designed to empower organizations in high-growth sectors such as AI & HPC research, Government, and BFSI. By providing a scalable, secure, and unified environment, the platform supports the critical need for sovereign AI infrastructure within India, allowing for better control and governance of data-intensive workloads.
Source: BSE
