PCIe Gen5 SSDs Take Center Stage for AI PCs and Gaming
Across the show floor, PCIe Gen5 SSDs dominated the storage narrative, targeting both high-end gaming rigs and emerging AI PCs. KIOXIA’s new XG10 client NVMe family steps up to a PCIe 5.0 x4 interface and NVMe 2.0d, delivering up to 14,000 MB/s sequential read and 12,000 MB/s write, with random performance rated as high as 2,000K/1,600K IOPS. This puts the drive firmly in the upper tier for OEM notebooks, desktops, and workstations that juggle large datasets, AI-assisted applications, and heavy content creation. addlink’s virtual showcase mirrors this trend with its G57 and G55 PCIe Gen5 SSDs, also positioned for AI PCs and performance workstations, while Biwin’s Black Opal X570 PRO PCIe Gen5 SSD pushes up to 14,000 MB/s reads and 13,000 MB/s writes. Together, these PCIe Gen5 SSDs mark a clear shift toward storage tuned for AI, workstation, and enthusiast gaming workloads.
AI-Ready DDR5 Memory and Enterprise Workstation Focus
Memory vendors are aligning their roadmaps with AI servers and professional workstations. G.SKILL is pivoting from its roots in DIY and extreme overclocking toward high-performance RDIMM and ECC UDIMM solutions designed for AI servers, enterprise workstations, and advanced industrial PCs. The company emphasizes reliable DRAM for database training and continuous workstation computation, catering to organizations that want to keep sensitive data on internal AI platforms rather than the cloud. ADATA, through its broader AI ecosystem, complements this with enterprise memory and storage coordinated under its TRUSTA brand, which introduces an AI Scaler memory-storage architecture that dynamically balances GPU, DRAM, and SSD resources to reduce deployment costs for AI training and inference. addlink also contributes to the DDR5 memory Computex story with its SC5 DDR5 CUDIMM, supporting speeds up to 8,400 MT/s to match upcoming platform requirements for AI PCs and high-end desktops.
Thunderbolt 5 Storage, Stackable Expansion, and AI Acceleration
Thunderbolt 5 storage and connectivity formed another major pillar of the event. OWC showcased an extensive lineup, including the Envoy Ultra, billed as the first certified 8TB Thunderbolt 5 portable SSD, and the StudioStack, a stackable Thunderbolt 5-powered solution offering up to 32TB of expansion for creative pros. The company’s Thunderbolt 5 hubs and docks extend this ecosystem with port expansion and dual 10GbE networking for demanding workflows. The standout, however, is OWC Stack AI, a Thunderbolt 5 AI accelerator and storage hub that effectively expands AI working memory for compatible PCs. By offloading larger models from constrained GPU memory, Stack AI aims to let businesses and power users run more complex AI workloads locally, maintaining privacy and reducing reliance on cloud inference. Together, these Thunderbolt 5 solutions blur the line between fast external drives, NVMe RAID, and full-fledged AI infrastructure nodes.
Docking, Portable Storage, and Verbatim’s Storage-and-Power Vision
Beyond raw PCIe Gen5 performance, vendors are rethinking how storage plugs into mobile and hybrid workflows. addlink’s B31 Magnetic NVMe docking station integrates an NVMe enclosure, 100W PD charging, HDMI 4K60, USB-C, USB-A, and SD/TF slots into a single USB-C NVMe docking station designed for one-cable laptop desks. Its P30 USB4 portable SSD pushes up to 4,000 MB/s, targeting creators who need near-internal speeds on the go. Biwin’s Amber PX4000 portable SSD takes a rugged approach with up to 3,900 MB/s reads, IP67 protection, and capacities up to 8TB for cross-device use. Verbatim frames its portfolio as S’n’P (Storage and Power), updating its Pocket SSD line with durable metal enclosures, integrated hub functions, write protection, and ultra high-speed USB 3.2 Gen 2x2 drives up to 2,000 MB/s, while also introducing power banks and GaN chargers to keep these storage-heavy setups running.
From Cloud to Edge: Memory, Storage, and AI Workstation Workloads
A unifying theme across exhibitors is optimizing both gaming and AI workloads from cloud to edge. ADATA’s “Activate the AI Core” ecosystem spans cloud hardware, industrial edge solutions, and XPG gaming components, with its TRUSTA AI Scaler toolkit orchestrating GPU, DRAM, and SSD resources to cut AI deployment costs and improve utilization. G.SKILL highlights growing demand from small and medium enterprises for AI workstation memory that keeps proprietary data local, using server-grade RDIMM and ECC UDIMM modules in internal AI platforms. Meanwhile, OWC’s Jellyfish Nomad shared storage, along with its broader RAM portfolio, targets 3D, VFX, and collaborative editing workloads that increasingly overlap with AI-assisted tools. Together with high-bandwidth PCIe Gen5 SSDs and versatile Thunderbolt 5 storage, these solutions show a market converging on the same goal: enabling creators, gamers, and enterprises to run AI and data-heavy workflows locally with workstation-class responsiveness.
