TRUSTA Unveils Advanced AI Scaler Toolkit at SuperComputing 2026
TRUSTA's Showcase at SuperComputing 2026
TRUSTA, ADATA's enterprise storage brand, made a significant entrance at the SuperComputing 2026 event, previously known as SC26, cementing its position in the competitive field of high-performance computing. With over 450 exhibitors and more than 18,000 attendees expected, the event held from November 17-19 at McCormick Place in Chicago is a pivotal platform for technology advancements. TRUSTA emphasizes the importance of scaling AI efficiently across enterprises, unveiling its upgraded AI Scaler Toolkit, which integrates GPU, DRAM, and SSD resources.
The AI Scaler Toolkit: Innovations and Features
The upgraded AI Scaler Toolkit stands out by utilizing vLLM's KV cache offloading, significantly enhancing AI's operational memory from GPU to DRAM and TRUSTA's enterprise SSDs. This novel approach allows for the quick reloading of long contexts from SSDs rather than recomputing them, thereby drastically reducing response times for AI applications and allowing for on-premises deployment at a lower cost, especially when GPU resources are limited.
In addition, the toolkit merges the Strata inference engine with Intel® AMX acceleration, which is claimed to speed up initial prompt processing (prefill) by as much as 25%. This blend of technology aims to streamline AI processes, making it easier for enterprises to harness machine learning effectively.
As artificial intelligence shifts from training to inference, the demand for efficient storage solutions elevates. Recent data from Counterpoint Research indicates that enterprise SSDs constituted 48% of global NAND flash capacity shipments in Q2 2026, a substantial increase from the previous year's 26%. Underlining this growth, TRUSTA's toolkit provides a seamless experience, with a one-click installation system that supports multiple models and inference engines like vLLM, llama.cpp, and Strata. This comprehensive interface allows for efficient allocation and management of GPU, DRAM, and SSD resources.
High-Performance SSDs and DDR5 R-DIMMs
Alongside the toolkit, TRUSTA featured its PCIe Gen5 enterprise SSDs and DDR5 R-DIMMs, designed to meet the demanding requirements of AI servers and high-performance computing (HPC) environments. The TD7P51 ECO/T7P5 PCIe Gen5 Enterprise SSD showcases impressive capacities up to 15.36TB with an endurance of up to 3 DWPD while maintaining energy efficiency below 20W. This advancement not only enhances storage solutions but also aligns with the industry's shift towards greener data centers.
The enterprise DDR5 R-DIMM offers up to 128GB per module at speeds of 6,400MT/s, incorporating built-in ECC for improved reliability even during intense workloads.
A Future-Forward Strategy
As enterprises continue to embrace AI technology, the optimized integration of computational resources will be crucial. TRUSTA asserts that businesses need not invest heavily in costly GPU systems; instead, maximizing the utility of CPUs, memory, and SSDs can facilitate more economical and faster AI capabilities. The commitment to innovative solutions reflects TRUSTA's mission to empower businesses to create adaptable, efficient, and scalable AI infrastructures.
In summary, TRUSTA's showcase at SuperComputing 2026 illustrates its unyielding dedication to innovating AI storage solutions tailored to the evolving landscape of data-driven industries. By equipping organizations with the necessary tools to enhance their AI strategies, TRUSTA positions itself as a leader in the next era of storage technologies aimed at redefining the future of enterprise computing.