EVERKIN Introduces New AI Data Cycle Storage Framework to Help Customers Capture the Value of AI

Shenzhen, Guangdong – June 2, 2025 – EVERKIN, a global leader in connectivity and data storage solutions, today announced the launch of its AI Data Cycle Storage Framework, a comprehensive architecture designed to address the full spectrum of enterprise AI requirements. By combining cutting-edge hardware with intelligent data management strategies, EVERKIN empowers organizations to optimize every stage of the AI lifecycle—from data ingestion and preprocessing to model training, inference, and long-term retention.

Growing Suite of Enterprise AI Storage Solutions for Compute- and Storage-Intensive Workloads

As enterprises accelerate AI adoption, they encounter increasingly complex demands for high throughput, low latency, and massive capacity. EVERKIN’s expanding portfolio of AI-optimized storage products is engineered to meet these requirements head-on, delivering performance and reliability across a variety of use cases:

High-Performance NVMe SSDs for Training and Inference

• Gen4/Gen5 NVMe Solutions: EVERKIN’s NVMe SSDs leverage PCIe Gen4 and Gen5 interfaces to deliver sustained sequential read/write speeds of up to 12 GB/s and random read IOPS exceeding 1.8 million. These drives are purpose-built for multi-node GPU clusters, ensuring that large-scale model training jobs proceed without data bottlenecks.
• Endurance-Optimized NVMe Drives: For real-time inference workloads—where consistent low latency is critical—EVERKIN offers endurance-rated SSDs with up to 10 Drive Writes Per Day (DWPD). Built-in power-loss protection and advanced wear-leveling algorithms guarantee predictable performance over millions of inference operations.

Tiered Storage Architectures for Cost-Efficient Data Management

• Hybrid SSD/HDD Enclosures: EVERKIN’s tiered storage modules combine high-speed NVMe SSDs with high-capacity enterprise HDDs in a single chassis. Intelligent tiering software automatically migrates “hot” datasets to SSD tiers during compute-intensive phases, while “cold” data resides on power-efficient HDD tiers—optimizing both performance and total cost of ownership.
• Automated Workload Balancing: The framework’s software layer monitors data usage patterns in real time, dynamically reassigning storage resources to maintain optimal throughput. This ensures that AI pipelines—such as data labeling, feature extraction, and batch inferencing—always access the appropriate storage tier for maximum efficiency.

Edge-Ready Connectivity Kits

• Rugged USB-C and PCIe Expansion Hubs: To support distributed AI inference at remote or industrial locations, EVERKIN’s edge-optimized connectivity kits provide multiple M.2 NVMe slots, Gigabit Ethernet ports, and optional 5G modules in compact, IP54-rated enclosures. These hubs accelerate the transfer of sensor data and enable local preprocessing before sending subsets to core compute clusters.
•   AI-Ready Media Readers: Photographers, videographers, and data-gathering teams can offload high-resolution imagery and video directly into AI preprocessing nodes using EVERKIN’s Thunderbolt 4 card readers, which support simultaneous transfer of CFexpress, SD, and microSD media at up to 40 Gbps.

High-Capacity HDDs for Long-Term Retention

• 22 TB Enterprise HDDs: When it comes to archiving multi-petabyte datasets—such as historical logs, large-scale image repositories, or compliance archives—EVERKIN’s helium-sealed 22 TB HDDs offer sustained transfer rates up to 270 MB/s, a 7,200 RPM spindle speed, and a mean time between failures (MTBF) exceeding 2.5 million hours. These drives minimize power consumption (as low as 5 W idle) and maintain industry-leading error rates (less than 1 nonrecoverable error per 10^16 bits read).
• Cold Storage JBOD Systems: For massive-scale archival needs, EVERKIN’s 4U 60-bay chassis supports up to 1.32 PB per enclosure. Built-in power-management controls spin down inactive drives to reduce energy usage, while integrated temperature sensors and redundant cooling ensure data integrity over long retention periods.

“Enterprises can no longer afford to treat storage as a generic resource,” said Austin Zhu, CEO of EVERKIN. “With ever-larger AI datasets and increasingly complex models, organizations need a storage framework that delivers both top-tier performance and cost-effective scalability. Our AI Data Cycle Storage Framework brings together enterprise-grade SSDs, HDDs, and intelligent tiering to streamline every phase of the AI journey.”

About EVERKIN
EVERKIN is a global provider of advanced connectivity and enterprise storage solutions, including USB-C Hubs, Docking Stations, Adapters, Cables, and high-performance SSD/HDD systems. Serving cloud operators, data centers, AI developers, and industrial customers in over 80 countries, EVERKIN is committed to innovation, reliability, and sustainability—accelerating digital transformation and empowering enterprises to capture the full value of artificial intelligence.