Technology · United States

Dell Expands AI Data Platform to Ground Agents in Enterprise Context

Credited to HPCwire · hpcwire.com · 18h ago

Dell AI Data Platform, the data foundation of the Dell AI Factory, expands with new data orchestration and storage capabilities

ROUND ROCK, Texas, Oct. 6, 2026 — Dell Technologies is expanding the Dell AI Data Platform so AI systems deliver better answers, cost less to run and reach production faster, giving applications and agents trusted, current context from a company’s own data.

In most cases the model isn’t the bottleneck to achieving enterprise scale for AI. The data is. A company’s most useful information sits in files, databases, cloud services and systems across multiple sites. Most of it was created long before anyone expected an AI agent to read it. It was never labeled, connected or organized with a machine reader in mind. As a result, agents burn tokens reconstructing answers that should already exist. Every query means more steps, more compute, more cost and slower, less trustworthy results.

The Dell AI Data Platform is built end to end, from where data is stored to where agents use it. Storage moves critical data to accelerated compute without bottlenecks; orchestration connects and prepares it; analytics and search transform and index it. NVIDIA Nemotron Retriever models enable document parsing, embedding and reranking. NVIDIA cuVS accelerates vector indexing and search. Because every layer runs on one platform, everything it produces traces back to governed data the customer already owns. Agents return to the same documents, history and decisions repeatedly, and most systems rebuild their understanding from scratch every time. That means answering the same two questions on every request: what does this term mean and how does it relate to everything else? Dell is extending the Dell AI Data Platform with three new features to answer both at once.

The Unified Semantic Layer, Knowledge Graph and Knowledge Agents hold a customer’s most valuable and sensitive information, so they stay inside the platform in the customer’s data center. The platform builds context from that data, keeps it current as the business changes and shares it across every application and agent, without locking customers into any single model, data or storage provider.

Acceleration and Security Built into the Data Path

Business context is only reliable if data is accurate, current and governed from source to application. Customers should not have to adopt another proprietary layer to accelerate different parts of their data environment. Dell brings acceleration into the data engines themselves, working with NVIDIA across the data path. Data must be transformed and analyzed, then reach accelerated compute fast enough to keep GPUs saturated and stay protected along the way. Dell data and storage advancements include:

Dell is expanding Dell AI Data Platform implementation services across its data and storage engines to move organizations from deployment to production results quickly and reliably. Dell Services establish a production-ready foundation, activating the platform’s analytics, processing, search and orchestration capabilities while keeping the platform tuned to the demands of AI workloads.

“Data without context is just noise. Most enterprises have spent years making their data accessible,” said Arthur Lewis, president of Dell’s Infrastructure Solutions Group. “That’s not the same as making it usable. An agent that can find a customer record but doesn’t know what it means, how it connects to everything else, or whether it can be trusted isn’t intelligent. It’s just fast. The companies that solve it won’t just deploy AI faster or run more agents. They’ll get more out of the data they already have.”

Dell Technologies (NYSE: DELL) helps organizations and individuals build their digital future and transform how they work, live and play. The company provides customers with the industry’s broadest and most innovative technology and services portfolio for the AI era.

TechnologyUnited States

Cisco Expands Validated SONiC Across Nexus N9000 for Open AI Fabrics

Credited to Converge Digest · convergedigest.com17h ago

Useful

Cisco is expanding its validated SONiC network operating system across the Nexus N9000 portfolio, creating a supported open-networking model that spans both conventional leaf-spine networks and high-performance AI fabrics. The move extends SONiC beyond the build-your-own model commonly associated with hyperscalers by pairing a precompiled, security-hardened SONiC image with Cisco hardware lifecycle management, Nexus Dashboard visibility, and end-to-end Cisco TAC support. Cisco says the model is intended for enterprises, neoclouds, web-scale operators, and AI infrastructure providers that want open networking without maintaining their own SONiC distribution.

The first configurations are available now on Nexus N9000 systems using Cisco Cloud Scale silicon for high-density 100G and 400G leaf-spine networks, as well as the N9100 family based on NVIDIA Spectrum-X Ethernet silicon for loss-sensitive backend AI fabrics. Cisco says support for Silicon One-powered Nexus N9000 platforms is coming next, which would extend the validated SONiC model across its multi-silicon Nexus switching portfolio. The same underlying Nexus hardware can support different software paths, including NX-OS, ACI, or Cisco-validated SONiC, depending on the platform and deployment model.

The operational layer is an important part of the announcement. Cisco takes upstream SONiC builds and produces a precompiled image that it says has undergone security hardening and regression testing across supported silicon. Nexus Dashboard provides common visibility across NX-OS, ACI, and Cisco-validated SONiC nodes, including telemetry, anomaly detection, and health monitoring. Cisco TAC then covers the hardware, optics, Switch Abstraction Interface (SAI), Free Range Routing (FRR) stack, and compiled SONiC image. Cisco positions the architecture for AI infrastructure ranging from roughly 8,000 to more than 32,000 GPUs, with 400G/800G N9100 and Cloud Scale configurations supporting RoCEv2 congestion-control mechanisms including PFC and ECN.

Available now — Cisco-validated SONiC on N9000 with Cloud Scale for 100G/400G leaf-spine networks and N9100 with NVIDIA Spectrum-X for backend AI fabrics.

Coming next — Cisco Silicon One-powered N9000 support, extending validated SONiC across Cisco’s multi-silicon Nexus portfolio.

Software — Precompiled, security-hardened Cisco-validated SONiC binary with regression testing across supported silicon.

Operations — Nexus Dashboard visibility across SONiC, NX-OS and ACI, including telemetry, anomaly detection and health monitoring.

Support — Cisco TAC supports switches, optics, SAI, FRR and the compiled SONiC image under a single support model.

AI scale — Targeted at AI infrastructure providers scaling from 8,000 to 32,000+ GPU clusters.

Open networking — Cisco continues to support DIY community SONiC on Cisco 8000 while positioning validated SONiC on Nexus N9000 for customers seeking a supported operating model.

Cisco describes the distinction this way: build-your-own SONiC remains available on Cisco 8000 systems for hyperscalers and operators that maintain their own distributions, while Cisco-validated SONiC on Nexus N9000 is aimed at organizations seeking a production-ready binary and vendor accountability. Cisco also notes that it is a primary maintainer of FRR, participates in the SONiC Technical Steering Committee, and contributes to SAI development.

The significant development here is not simply that Nexus hardware can run SONiC. Cisco had already introduced the N9100 with the ability to run NX-OS or SONiC when it expanded its NVIDIA Spectrum-X portfolio in October 2025. Today’s announcement defines a broader operational model around SONiC: Cisco is validating and hardening the software, integrating it with Nexus Dashboard, supporting it through TAC, and extending the approach across multiple ASIC families.

That makes the announcement relevant to a larger debate around open Ethernet for AI infrastructure. SONiC emerged largely from hyperscale environments where operators possessed the engineering resources to integrate and maintain their own networking software. As neoclouds, enterprises and sovereign AI operators build increasingly large GPU clusters, the question becomes whether SONiC’s disaggregated model can be retained while shifting more integration and lifecycle responsibility back to the equipment supplier.

Cisco is effectively offering both models. Customers with substantial internal network engineering resources can continue using community SONiC and Cisco 8000 hardware, while Nexus customers can use a Cisco-supported distribution spanning Cloud Scale and NVIDIA Spectrum-X, with Silicon One support planned next. The latter is particularly notable because it gives Cisco a common open NOS strategy across multiple switching-silicon architectures rather than tying SONiC to one ASIC family. Cisco has not specified an availability date for the forthcoming Silicon One N9000 support.

Networking — Nexus • Silicon One • Cisco 8000 • SONiC

AI Infrastructure — Ethernet AI fabrics, Spectrum-X systems, Silicon One switching, optics, security and infrastructure management.

October 28, 2025 · AI Networking Cisco Expands AI Infrastructure Portfolio with NVIDIA The original N9100 launch introduced Spectrum-X silicon and NX-OS/SONiC software flexibility to the Nexus portfolio.

Original · Converge Digest

Front Method

Technology · United States

Vast uses tiered storage to ease AI agent memory demands

Credited to SiliconANGLE · siliconangle.com · 20h ago

Photograph · SiliconANGLE

AI agent memory is creating new demands on infrastructure as agents run longer sessions and spread across the enterprise. Retaining that context and making it available when needed puts pressure on memory capacity and data movement.

Those demands extend beyond the context held during an individual interaction. Enterprise agents also need shared knowledge that persists across sessions, according to Alon Horev, (pictured) co-founder and chief technology officer of Vast.

“Memory for agents is a bit different,” Horev said. “First of all, there are multiple types of memory. There’s long-term memory where an agent can see past conversations and past interactions and look back and learn from its past experiences.”

Horev spoke with theCUBE Research’s John Furrier and Dave Vellante at Fully Connected 2026, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed AI agent memory, key value cache offloading and the shift toward data movement as the next bottleneck. (* Disclosure below.)

Why AI agent memory is moving beyond the GPU

The pressure shows up first in inference. Each long-running session holds its KV cache in graphics processing unit memory, and a session of half a million tokens can take up one-tenth to one-twentieth of a GPU’s memory, Horev explained.

“It’s also possible the agent would stop talking to the [large language model] because it’s compiling code, it’s testing software, or, as a human, I want to have a cup of coffee,” he said. “What you see is that if you could stretch that memory wall and basically offload those sessions to storage, you can avoid that repeat recalculation.”

Vast’s approach uses memory in tiers. GPU memory is used first, then central processing unit memory on the same machine, then persistent media that can hold petabytes of KV cache, with Nvidia Corp.’s Dynamo software orchestrating the process, Horev noted.

“You can move a session from one busy GPU to one less busy GPU and move KV cache either over the network or read it from Vast,” he said. “Once you look at inference as a distributed problem where you have the opportunity to use GPU memory, CPU memory and Vast across a fleet of machines, you have more optionality and you have more optimized scheduling.”

The stakes rise as companies deploy thousands of agents that handle sensitive data and act on customers’ behalf. Those enterprises need to record everything their agents do and retain it for a set period, which makes AI agent memory both a governance and performance asset, according to Horev. Vast has also launched a confidential computing service for sensitive workloads.

“These conversations that the agent is doing, it’s also gold,” he said. “It’s the same information that it can use for fine-tuning or training or creating purpose-built models.”

Here’s the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of Fully Connected 2026:

(* Disclosure: TheCUBE is a paid media partner for the Fully Connected event. Neither CoreWeave, the sponsor of theCUBE’s event coverage, nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Support our mission to keep content open and free by engaging with theCUBE community. Join theCUBE’s Alumni Trust Network, where technology leaders connect, share intelligence and create opportunities.

Are you an AWS customer? Support SiliconANGLE financially by buying your AWS services from our Marketplace portal page and links: https://siliconangle.com/aws-marketplace/

Founded by tech visionaries John Furrier and Dave Vellante, SiliconANGLE Media has built a dynamic ecosystem of industry-leading digital media brands that reach 15+ million elite tech professionals. Our new proprietary theCUBE AI Video Cloud is breaking ground in audience interaction, leveraging theCUBEai.com neural network to help technology companies make data-driven decisions and stay at the forefront of industry conversations.

Atlassian lays groundwork for humans and AI agents to work side by side

Caterpillar and CoreWeave shorten the learning loop for physical AI

Google goes nuclear to power its AI data centers, committing to buy 890 megawatts of energy from Constellation

Meta teams up with Bret Taylor's Sierra Technologies on new standards for AI agent commerce

Multiply Labs raises $75M to optimize pharmaceutical manufacturing with robots

In the agentic era, standing privilege must go: Key thoughts from the SailPoint Navigate keynote

INFRA - BY MARIA DEUTSCHER. 10 HOURS AGO

SECURITY - BY ZEUS KERRAVALA. 10 HOURS AGO

Join our community