According to PR Newswire and coverage in Antara and Dealroom, Solid State Storage Technology Corporation (SSSTC), a subsidiary of Kioxia Corporation, is showcasing an expanded lineup of SSDs purpose-built for immersion cooling at Computex 2026. Reported product families include SATA ER3, ER4, ER5 and PCIe U.2 PJ1 and EJ5 series, with materials and structural enhancements to improve corrosion resistance in non-conductive dielectric fluids. Sources state enterprise eTLC drives offer endurance options of 1 and 3 DWPD over five years and maintain more than 90% random IOPS consistency under sustained workloads, while industrial models support operating ranges of -40 C to 85 C and anti-vibration designs. Reporting frames these releases as addressing thermal and reliability challenges in high-density AI deployments.
What happened
According to PR Newswire and Antara (republishing the PR), Solid State Storage Technology Corporation (SSSTC), a subsidiary of Kioxia Corporation, showcased an expanded portfolio of SSDs at Computex 2026 aimed at immersion-cooling use cases for AI data centers. Per the coverage, the announced lineup includes SATA ER3, ER4, ER5 series and PCIe U.2 PJ1 and EJ5 series, and emphasizes corrosion-resistant materials, component protection, and structural design for submersion in non-conductive dielectric fluids.
Technical details
Reporting describes the immersion approach as using liquids with higher heat capacity and convective heat transfer to improve heat removal and Power Usage Effectiveness (PUE). Per Antara and PR Newswire, the enterprise eTLC SSDs offer endurance choices of 1 and 3 DWPD over a five-year period and are claimed to maintain over 90% random IOPS consistency under sustained workloads. The industrial SSDs are reported to support operating temperatures from -40 C to 85 C, include anti-vibration and shock-resistant mechanical design, and use pSLC architecture options and a multi-tier Power Loss Protection framework (hardware PLP, firmware PLP, and PLN) for data integrity.
Industry context
Industry reporting frames these product updates as part of a wider move by storage vendors to adapt hardware to immersion environments as AI server densities rise. Industry coverage notes immersion cooling reduces reliance on air cooling and can materially lower rack-level thermal constraints, but it also creates new engineering requirements around corrosion resistance, connector reliability, and firmware behavior under continuous liquid exposure.
For practitioners
Observers evaluating immersion-ready SSDs should consider three practical dimensions: stated endurance and DWPD targets for write-heavy AI workloads; mechanical and connector qualifications for sustained liquid exposure; and the vendor's firmware features for consistent random IOPS under sustained load. Per the PR coverage, SSSTC highlights configurable firmware options and more than 18 years of in-house firmware development experience as supporting customization for diverse deployments.
What to watch
Reported claims on endurance, IOPS consistency, and corrosion-resistant materials are verifiable in vendor datasheets and independent test labs; readers should look for third-party validation of sustained workload performance and long-term reliability in immersion racks. Also monitor interoperability notes from system and rack vendors: immersion-ready SSDs require validated connectors, seals, and system-level thermal management to realize stated PUE gains.
Attribution note
All product names, endurance figures, temperature ranges, and design claims above are drawn from PR Newswire, Antara, and Dealroom summaries of SSSTC's Computex 2026 disclosures. SSSTC has not been quoted directly in the scraped coverage included in these sources.
Scoring Rationale #
This is a notable hardware update that expands options for immersion-cooled AI deployments; endurance and sustained-IOPS claims matter to architects and procurement. It is not a frontier-shifting research breakthrough, so it rates as "notable" for practitioners.
Practice interview problems based on real data
1,500+ SQL & Python problems across 15 industry datasets — the exact type of data you work with.