Dell is using Kioxia’s 245.76 TB LC9 QLC SSD with its PCIe Gen 5 interconnect in its PowerEdge servers. Arun Narayanan, SVP, Compute and Networking, Dell Technologies, said: “SSDs like the KIOXIA LC9 Series combined with Dell PowerEdge servers offer high-capacity, power-efficient solutions tailored for advanced AI workloads while optimizing TCO and data center footprints.”
…
Research house Dell’Oro projects the worldwide Data Center Physical Infrastructure (DCPI) market to grow at a 15 percent CAGR from 2024 to 2029, reaching $63.1 billion by 2029 as AI-ready capacity accelerates through the mid-decade. This outlook reflects stronger-than-expected deployments to support accelerated computing workloads. Service providers (cloud and colocation) are set to grow at a 20 percent CAGR through 2029, while Enterprise increases to a 6 percent CAGR as enterprise leaders favor colocation partners to host AI infrastructure. Growth is broad-based, with North America leading and EMEA/China peaking around 2026 before moderating; AI sovereignty and export-policy shifts support momentum. Operators increasingly combine utility ties with on-site generation and other tactics. Overall, it expects only a modest impact on capacity expansion from power constraints. Buy its report here.
…
Content collaborator Egnyte announced its Agent Builder with a no-code framework. Users can create customized AI agents to automate time-consuming, content-intensive tasks. They have secure access to their company’s internal content and information from the public web. Egnyte’s existing AI agents serve as templates that allow non-technical users to leverage them as launching points for their agent creation that can be tested and shared across the organization to ensure consistent, reliable outputs. Customers can build custom AI agents or use agent templates.
Egnyte’s customers have already begun implementing AI agents to transform their workflows, tailored to the needs of their organization. Examples include creating agents that generate investment memos from extensive CIM documents and compare potential investment opportunities in Financial Services, crafting RFPs for subcontractors in Architecture, Engineering, and Construction (AEC), and, across industries, HR departments have created agents that help answer HR-related questions from employees, leveraging their internal policies and procedures.
To learn more about Egnyte’s AI Agent Builder, click here.
…
EnterpriseDB announced its sovereign data and AI platform, EDB Postgres AI integrates with Nvidia AI software. It says EDB PG AI, built with Nvidia accelerated computing and Nvidia AI Enterprise, including NIM microservices and NeMo Retriever, enables organizations to build AI and data platforms using PostgreSQL. This integration enhances EDB PG AI as a secure, high-performance platform for deploying generative and agentic AI with enterprise-grade governance, observability, and cost efficiency. Measurable outcomes for customers include:
- 2-3x throughput improvement for data embedding tasks
- 1.5-2x throughput improvement for retrieval performance
- Bring new, sovereign AI applications into production 3x faster
Read more information here.
…
DigiTimes reports Micron is drawing back from the Chinese market and stopping its UFS 5.0 development. It says this is a positive move from the Phison and SIMO UFS controller market point of view.
UFS 5.0 is a flash storage standard for mobile and similar devices. It will succeed UFS 4.0 and provide almost double its 5.8 GBps bandwidth. The JEDEC standard is expected to be released in 2027.
…
TrendForce reports DRAM fabber Nanya and IC design house Etron Technology are forming a Hsinchu City, Taiwan-based JV to build High Bandwidth Memory (HBM). Nanya plans to launch new HBM products with its partners by the end of 2026, targeting applications such as AI PCs, smartphones, robotics, and automotive systems. Nanya is working with packaging and testing partner Formosa Advanced Technologies to establish 3D Through-Silicon Via (TSV) processes and multi-chip stacking capabilities.
Comment: SK hynix is making so much money from HBM, with Nvidia using more and more of the stuff. Micron and Samsung have piled into the HBM market as well. It’s likely AI Inferencing will spread to desktop, handheld, and embedded devices in the automotive and other areas. If such devices use GPUs, HBM will speed the inferencing workloads and specialized designs could be needed, providing scope for new entrants.
…
Quantum, the troubled tape system, object archive, video storage, deduping backup target and file management supplier, has expanded into China, India, and the ASEAN market via distribution. It has agreements with ChangHong IT (CHIT) in China, Rashi Peripherals Limited in India, Hibino Graphics Corporation (formerly NGC) in Taiwan, and ACA Pacific in ASEAN. The new channel model is designed to expand Quantum’s market reach and enhance the customer experience with extended local service and support coverage, faster time-to-delivery, and tailored technical support. Each distributor will lead go-to-market efforts in their territory, invest in sales and marketing growth, and collaborate closely with Quantum to offer regional customers tailored support and services.
…
Korea’s Alpha Economy outlet reports Samsung is going to supply 12-hi HBM3E memory to Nvidia, which will receive 30,000 to 50,000 units for use in water-cooled servers in the near future. Samsung did not comment.
…
Germany’s Heise IT news outlet reports Seagate has found a counterfeit disk drive factory, which it alleges fraudulently sells used drives as new ones by deleting SMART on-drive usage data. Malaysian officials and the drive maker said in a statement: “Members of the Seagate security team from Singapore and Malaysia, together with officials from the Malaysian Ministry of Domestic Trade, unearthed a first counterfeiting workshop in a cramped storage room outside Kuala Lumpur back in May. According to the company, this counterfeiting workshop was taking in thousands of US dollars every month.” Used WD and Toshiba drives were also allegedly being processed to appear new. Officials claimed: “The fraudsters sold the counterfeit drives online via Shopee and Lazada, two of the largest e-commerce platforms in Southeast Asia.”
…
Silk, which supplies software enabling databases and similar workloads to use extremely fast ephemeral block storage in the public cloud, has hired Ed Filippine as president to lead Silk’s go-to-market functions, customer success operations, and global expansion as the company accelerates adoption across mission-critical enterprise and AI workloads. He previously held executive leadership roles at Vertica, Carbon Black, and CloudHealth Technologies, where he helped scale both companies before they were acquired by HP and VMware.
Silk CEO Dani Golan said: “As enterprises push AI into production and face growing performance and cost challenges in the cloud, Ed’s leadership in GTM will be critical in expanding the Silk platform to a broader set of global customers.”
…
A study by researchers from UCL, UC Davis, and Mediterranea University of Reggio Calabria, to be presented and published as part of the USENIX Security Symposium, looked at GenAI browser assistants and privacy. The researchers analyzed ten of the most popular generative AI browser extensions, such as ChatGPT for Google, Merlin, and Microsoft Copilot. It uncovered widespread tracking, profiling, and personalization practices that pose serious privacy concerns, with the authors calling for greater transparency and user control over data collection and sharing practices. One assistant, Merlin, even captured form inputs such as online banking details or health data.
The paper by Yash Vekaria et al. “Big Help or Big Brother? Auditing Tracking, Profiling, and Personalization in Generative AI Assistants”’” is published in the proceedings of the 34th USENIX Security Symposium. Read the abstract here.
