Press Release · ADATA Technology Co. Ltd
TRUSTA Debuts at SuperComputing 2026 with Upgraded AI Scaler Toolkit
Distributed by AI AI List · Originally issued via PR Newswire

Integrating GPU, DRAM, and SSD Resources to Accelerate Enterprise AI Inference and Deployment
TAIPEI, Nov. 18, 2026 — TRUSTA, the enterprise storage brand of memory leader ADATA, today makes its debut at SuperComputing 2026 (SC26), the world's premier event for high-performance computing. This year's SC26 is expected to draw more than 450 exhibitors and over 18,000 attendees. With a focus on helping enterprises scale AI efficiently, TRUSTA is showcasing AI solutions that bring together software, memory, and storage. Highlights include the upgraded AI Scaler Toolkit, which uses vLLM's KV cache offloading to extend AI's working memory from the GPU to DRAM and TRUSTA enterprise SSDs. Long contexts are reloaded from SSD instead of being recomputed, significantly shortening response wait times and helping enterprises deploy AI on-premises at lower cost with limited GPU resources. The toolkit also integrates the Strata inference engine with Intel® AMX acceleration, speeding up prompt processing (prefill) by up to 25%*.
Upgraded AI Scaler Toolkit: vLLM KV Cache Offloading to SSD for Enterprise AI
As AI shifts from model training to inference, storage demand is growing rapidly. According to Counterpoint Research, enterprise SSDs accounted for 48% of global NAND flash capacity shipments in Q2 2026, nearly double the 26% recorded a year earlier. During inference, an LLM keeps a KV cache of every token it has read, and long documents, multi-turn conversations, and AI agents quickly grow it beyond GPU memory. Once evicted, it must be recomputed, which keeps users waiting. The upgraded AI Scaler Toolkit integrates vLLM, the open-source engine widely used for enterprise AI serving, and offloads its KV cache across three tiers: GPU memory for active requests, DRAM as a fast cache, and TRUSTA PCIe Gen5 enterprise SSDs as a large, low-cost tier that persists across restarts. TRUSTA internal testing shows:
- Much faster on long documents: Long documents are reloaded from SSD instead of recomputed, significantly shortening the wait before AI responds.
- Context that survives reboots: After a server restart or reboot, conversations resume from the KV cache saved on SSD instead of being recomputed from scratch.
The toolkit also offers a one-stop experience: one-click installation, support for multiple models and inference engines (vLLM, llama.cpp, and Strata), and one interface for allocating GPU, DRAM, and SSD resources, tuning parameters, and testing models.
Strata with Intel® AMX: Up to 25% Faster Prompt Processing
For specific deployments that must fit very large models into limited GPU memory with efficient use of DRAM and SSD, the toolkit also includes the lightweight Strata engine. On Intel® Xeon® processors, Intel® AMX lets the CPU process model data directly in DRAM, speeding up prompt processing by up to 25%*.
High-Performance Enterprise SSD and DDR5 R-DIMM: Strengthening AI Infrastructure
Alongside the AI Scaler Toolkit, TRUSTA is also showcasing PCIe Gen5 enterprise SSDs and DDR5 R-DIMMs. These high-performance, high-capacity, energy-efficient memory and storage products meet the diverse needs of AI servers and HPC environments.
- TD7P51 ECO / T7P5 PCIe Gen5 Enterprise SSD: Capacities of up to 15.36TB in mainstream U.2, E1.S, and E3.S form factors, with endurance of up to 3 DWPD. Flexible Data Placement (FDP) technology improves storage efficiency and extends drive lifespan. With power consumption below 20W and 1.7x better energy efficiency than comparable products, it helps data centers reduce power usage and supports the global trend toward greener data centers.
- Enterprise DDR5 R-DIMM: Speeds of up to 6,400MT/s and capacities of up to 128GB per module. Built-in ECC automatically detects and corrects data errors, keeping AI servers stable and reliable under sustained heavy workloads.
"As enterprise AI adoption continues to grow, effectively integrating compute, memory, and storage resources has become key to improving AI deployment efficiency and controlling costs," TRUSTA stated. "Enterprises don't need to keep buying expensive GPUs to adopt AI. By making better use of CPUs, memory, and SSDs, they can run AI faster and more cost-effectively. With the upgraded AI Scaler Toolkit, enterprise SSDs, and memory technologies, TRUSTA helps enterprises make the most of their existing hardware and build more flexible, efficient, and scalable AI computing environments, so they can scale AI efficiently."
*Test notes: Based on TRUSTA internal testing in November 2026, comparing the AMX-enabled TRUSTA version of Strata in the AI Scaler Toolkit with upstream Strata v0.2.11. Test platform: Intel® Xeon® w5-3535X processor with a single NVIDIA RTX 6000 Ada GPU, running a Mixture-of-Experts (MoE) LLM. Only the prefill stage was measured. Each test was run at least three times, and median results are reported. With a single input of about 8K tokens, speed improved by approximately 11%. In multi-turn conversations with about 4K new tokens per turn, speed improved by approximately 25%. For single inputs above about 12K tokens, there was no significant difference. Actual performance varies by hardware configuration, model, and input length.
Exhibition Information
- Event: Supercomputing 2026 (SC26)
- Date: November 17–19, 2026
- Venue: McCormick Place West – Exhibit Halls, Chicago, USA
- Booth#: 7214
SOURCE ADATA Technology Co. Ltd
About TRUSTA
TRUSTA is the enterprise storage brand under ADATA Technology, dedicated to serving the AI server and data center markets. Guided by the brand spirit "Together, Empowering Tomorrow," TRUSTA is built on a foundation of trust and partnership, delivering stable, high-speed, energy-efficient, and scalable storage solutions tailored to the evolving needs of data-driven enterprises. Backed by ADATA's extensive global distribution network, strong R&D capabilities, and advanced manufacturing resources, TRUSTA is strategically positioned to lead the next wave of storage innovation for the AI era—empowering organizations worldwide to build a resilient foundation for future enterprise computing.
Global Media Contact
Brand Marketing – Media Relations
ADATA Technology Co., Ltd.
T: +886-2-8228-0886
F: +886-2-8227-1760
E: [email protected]