Everpure is enhancing its artificial intelligence (AI) capabilities at Nvidia’s GTC event by unveiling a complete AI data pipeline automation system, showcasing benchmark results for its FlashBlade AI hardware, and offering public cloud-like AI storage capacity provisioning driven by service-level agreements (SLAs). This comprehensive approach aims to address the growing demands of enterprise AI applications.
Central to Everpure’s AI pipeline automation is its Data Stream technology, a software suite designed to run on FlashBlade and Nvidia Blackwell GPU hardware. Data Stream automates and accelerates data ingestion, transformation, and optimization, essential for efficient enterprise AI pipelines. The company is also extending its Evergreen//1 service model, which revolutionized on-premises storage by allowing customers to subscribe to storage services rather than purchasing specific hardware.
Kaycee Lai, Everpure’s Vice President for AI and Analytics, emphasized the importance of treating AI as a “tier one application.” He stated, “Successful companies really treat AI as a tier one application; they make sure that you’ve got production grade, availability, reliability, performance, uptime.” The company’s Enterprise Data Cloud (EDC) for AI is designed to ensure robust performance and governance, akin to traditional enterprise environments like Oracle or SAP.
The Evergreen//1 model, introduced in 2018, shifted the focus from specific hardware ownership to a subscription-based approach, allowing customers to scale their storage capacity flexibly based on their needs. Everpure’s architecture has evolved from classic dual-controller setups to a disaggregated model that incorporates metadata storage controllers linked to scalable all-flash storage nodes. This approach is influenced by parallel file systems associated with high-performance computing (HPC).
The FlashBlade//EXA product, which utilizes existing FlashBlade technology for metadata control and NVMe fabric-connected data nodes, is now showing impressive performance metrics. According to recent benchmarks, EXA has outperformed competitors in the SPECStorage Solutions run and is set to score well in the upcoming ML Perf 2.0 results. Lai noted that EXA is expected to receive Nvidia-certified status later this year, alongside a co-engineered configuration with Supermicro aimed at simplifying access for organizations venturing into AI.
Everpure’s FlashBlade//S500 also posted a notable score of 7.2 million IOPS in the IO500 benchmark, reinforcing the performance credentials of the EXA system. Lai expressed confidence in EXA’s capabilities, stating, “On the preparation side, RAG pipelines can require 300,000 to a million IOPS each time you run it. Great. Flashblade delivers 400 million, 450 million IOPS.” He further highlighted that EXA supports multiple 100-gigabyte-per-second throughput during training and an exceptional 4.6 billion metadata operations per second for inference tasks.
The Data Stream technology integrates with Nvidia’s AI Data Platform, and Lai underscored the close partnership between Everpure and Nvidia. “We’re now seen as a tier-one storage partner with Nvidia,” he said, noting that the collaboration has intensified in the past year. Everpure boasts over 10 terabits per second in read performance, outpacing competitors such as Dell, DDN, Hammerspace, and others in the ML Perf 2.0 and SPECStorage benchmarks, while promising to lower total cost of ownership by as much as 58 percent.
Currently in beta testing, the Data Stream offering signifies Everpure’s commitment to AI, which has become its fastest growing business segment. Lai reported a “massive triple-digit growth year-on-year” in the AI sector alone, with many customers adopting Everpure solutions specifically for AI applications. “It’s a great one-two punch,” he added, highlighting the dual advantage of attracting new clients while also extending existing customer relationships into AI.
See also
AI-Driven Defense Technologies Face ITAR Compliance Challenges, Impacting Export Controls
Nvidia Announces Trillion-Dollar AI Sales Forecast, Unveils New Groq Chips and CPUs
Nvidia Reveals $1-Trillion AI Chip Sales Target as Inference Demand Soars
AI Computing Hardware Market Set to Expand from $45.51B to $172.15B by 2035 at 14.23% CAGR
HOPPR™ AI Foundry Launches NVIDIA’s NV-Reason and NV-Generate for Advanced Medical Imaging
