AI workloads are driving new hardware and storage requirements in OpenStack, while AI-assisted security research is driving up the number of vulnerability reports. Those are two of the prevailing them...

Supercapacitors help data centers manage AI power spikes with ms response for peak shaving, voltage stabilization, and brief backup – complementing UPS.

How optical frequency combs boost DWDM in AI data centers, reducing power and raising channel density for next-gen co-packaged optics.

IP Fabric builds a read-only digital twin of enterprise network infrastructure, and version 8.1 of the platform adds application infrastructure mapping and a redesigned cloud-native data model.

New Selector Foundry runtime environment lets network operations teams build, test, version, and govern their own AI agents inside the company's NetOps platform.

Cohere plans on making the full serving stack used in Model Vault open source so that independent auditors will be able to validate that it doesn't log, export, or leak data.

KeewanoDB gives every user, device and agent its own table, so AI agents can read event history in sequence instead of rebuilding it at query time.

Cornelis takes aim at AI bottlenecks with programmable network fabric.

Work on 1.6 terabit network connectivity began in 2020 and became a formal IEEE study group in 2021, before AI was the industry's dominant discussion point. That work became the IEEE P802.3dj Task For...

OpenAI's new Data agent for ChatGPT Work has no published benchmark, while rival Databricks released one the same week — a gap enterprise buyers should weigh.

Databricks' Adaptive Instructed-Retriever lets AI agents decide how many times to search per question, matching frontier model quality in half the time.

At VMware Explore 2026, Broadcom detailed what’s next for networking in VCF 9.1 as well as how both AI for networking and networking for AI will have an impact. The updates focus on transit gateway fl...

Security teams face multiple roadblocks to cyber resilience. Learn five obstacles to overcome to help prepare for and withstand a cyberattack.

Tailscale got its start back in 2019 as a VPN provider based on the open-source WireGuard technology. As it turns out, enabling private connectivity in the AI era has a lot of practical utility.

Cohere's Parse 5, a 2.3-billion-parameter model, offers cost-effective document parsing at $1.50 per 1,000 pages, prioritizing structure over raw accuracy.

In the Kubernetes 1.37 release, codenamed Garhwal, new scheduling capabilities were built for AI and machine-learning training jobs that need groups of pods to start and scale together.

A person who forwards AI-generated text, code, or other output without reading, understanding, or validating it. The person acts only as a relay between the AI system and the intended recipient.

Snowflake's new Cortex AI Gateway feature routes AI tasks to smaller models automatically, using the same access controls that already govern enterprise data.

Nvidia's Nemotron 3.5 Lightning pairs with its NeMo Switchyard router, which reassigns models mid-task and cuts task costs to a third in Nvidia's own tests.

Millions of people rely on a network of unpaid volunteers to stay anonymous online, and for 25 years that network has run on altruism rather than profit. Today, the Tor network's biggest threat is cen...

Alongside the disclosure, runZero released OOBscan, an open-source tool that scans for vulnerable BMCs and other out-of-band management devices.

Tencent's Team Memory gives AI agents shared access to chat history, code, and docs. Practitioners are asking how it handles a wrong or conflicting fact.

NatJack attack class exposes design flaw across decades of network infrastructure.

New specs from the LoRa Alliance are aimed at cutting manual work from low-power network onboarding and reducing coverage gaps in hard-to-reach locations.

Among 120 million newly registered domains, more than 22% were weaponized or showed threat characteristics, according to Infoblox's 2026 Threat Landscape Report.

Defenders need more bandwidth to keep up with AI-coordinated botnets, and Netscout is boosting its Arbor Cloud DDoS mitigation capacity to 33 Tbps.

A VB Pulse survey of 101 enterprises found most AI agent errors trace to missing context. Credible raised $10M to build a portable, open-source alternative.

'Many organizations already use Cloudflare for their public DNS. Internal DNS extends that same platform to private DNS, so public and private are managed from the same global network and control plan...

IPv4 address exhaustion is a key driver of the 1.58 release of NetworkManager, the open-source Linux network configuration tool suite.

As context windows and multi-turn interactions grow, so does the GPU compute wasted recalculating work a model has already done. Augmented Memory Grid, a NeuralMesh 6 feature built specifically for this, is Weka's answer.

LangChain, Conviva and CoreWeave leaders say scoring AI agent conversations one at a time hides broken products — and explain the shift to cohort analysis.

Zillow's SVP of Engineering says its AI ROI claims are credible only because of a DORA metrics baseline built years before the company's AI rollout began.

Intuit's AI VP explains why natural-language handoffs broke its agent orchestration layer, and why rebuilding twice in four months was the fast path.

LinkedIn rebuilt Kubernetes for agent speed, Walmart tackled duplicate agents, Zendesk its data pipelines: three fixes for enterprise AI agent infrastructure.

Meta infrastructure VP Barak Yagour says agentic queries grew 30x in six months, and enterprises have roughly 20 months to rebuild for AI agents.

TOP500 crowns LineShine while hyperscalers stand up multi-GW AI campuses. Why FLOPS and megawatts measure different victories in 2026.

Infoblox acquires Kentik to pair DNS authority with real-time traffic visibility.

Couchbase AI Data Plane combines persistent agent memory, vector search and an enterprise MCP server that runs on-device when cloud connectivity drops.

Real environments can't inject edge cases on demand. Alibaba's Qwen-AgentWorld simulates them — and outperformed real-environment RL across seven benchmarks.

Upscale AI raised $190M in new funding to support AI networking scale up and scale out architecture. Skyhammer is its switch silicon designed to connect GPUs and XPUs at very high speed.

NetBox Labs designed its newly launched platform to span the full infrastructure lifecycle, from initial procurement through decommission.

F5 adds AI-powered web application firewall (WAF) to guard networks against pre-exploit attacks.

Research from Infoblox finds the traffic routing technology is widespread across corporate, government, and financial networks, and defenders often cannot see it.

Cohere's North Mini Code ranks 8th of 127 open-weight models on output speed — but generates 3x the output tokens of comparable models in independent testing.

Google's open-source diffusion language model generates 256 tokens in parallel and self-corrects, hitting 4x speed on one GPU at a cost to quality.

'Quantum networking has many practical and commercial use cases in the classical world today,' said Ramana Kompella, head of Cisco Research and Cisco Fellow, in a session at Cisco Live 2026.

Snowflake's Horizon Context and Cortex Sense give agents a shared definition of business logic, targeting hybrid retrieval failures at the catalog layer.

As Cisco Live kicks off in Las Vegas, Tom Gillis shares how AI has changed networking, including inside Cisco, where AI coding tools have accelerated software development across its entire portfolio. ...

Cisco aims to become the infrastructure layer for the agentic enterprise. It launched a flurry of new products and features at its Cisco Live event taking place this week in Las Vegas.

Data center foundations and high-density cores still rely on concrete, while low-carbon mixes, mass timber, and retrofits move the needle at the edges.