Excelero
Excelero was an Israeli high-performance storage startup that developed NVMesh, a distributed NVMe-over-Fabrics block storage solution enabling shared NVMe flash performance at scale for data-intensive applications including AI/ML, HPC, and real-time analytics.
Visit WebsiteCompany Overview
Excelero developed NVMesh, a software-defined distributed block storage solution built on NVMe-over-Fabrics technology that disaggregated NVMe flash resources and presented them as shared, high-performance block volumes accessible across the network. NVMesh delivered bare-metal NVMe performance with the flexibility of shared storage—enabling AI/ML training, high-performance computing, real-time analytics, and other latency-sensitive workloads to access storage at millions of IOPS with sub-100-microsecond latency.
Commercially, Excelero competed in the high-performance distributed storage market alongside VAST Data, WekaIO, Pure Storage, and Lightbits Labs. Founded in 2014 in Tel Aviv by Lior Gal (CEO) and a team of storage industry veterans, the company raised approximately $30M from investors including Dell Technologies Capital, Western Digital Capital, Samsung Catalyst Fund, and Hitachi Ventures. In 2022, NVIDIA acquired Excelero for its distributed NVMe technology to enhance storage performance for AI and GPU-accelerated computing workloads.
From a defense and national security perspective, high-performance distributed storage is critical for military intelligence processing, real-time sensor data analytics, large-scale simulation environments, and AI/ML workloads for defense applications. The ability to deliver NVMe performance at scale enables rapid processing of intelligence data, satellite imagery, and signals intelligence that requires extreme storage throughput.
Dual-Use Assessment
High-performance NVMe storage is critical for military AI/ML workloads, intelligence processing, real-time sensor analytics, and defense simulation. Distributed NVMe-over-Fabrics enables the storage performance required for processing satellite imagery, SIGINT, and large-scale defense datasets.
Key Technologies
- NVMe-over-Fabrics distributed block storage
- Software-defined NVMe flash disaggregation
- Sub-100-microsecond latency distributed storage
- Multi-million IOPS shared block volumes
- Bare-metal NVMe performance with shared storage flexibility
- GPU-optimized storage for AI/ML training workloads
Use Cases & Applications
- AI/ML training storage with extreme throughput and low latency
- High-performance computing (HPC) shared storage
- Real-time analytics and streaming data storage
- Database acceleration for latency-sensitive applications
- Military intelligence data processing and storage (dual-use)
- Defense AI/ML and simulation environment storage infrastructure (dual-use)
Strategic Value to U.S.-Israel Alliance
High-performance storage infrastructure underpins military AI/ML, intelligence processing, and simulation capabilities. NVMe-over-Fabrics technology enables the data throughput required for processing defense intelligence workloads at scale.
Interested in this startup?
Learn more about our investment approach or get in touch to discuss opportunities in dual-use technology.