Network World reports on the most significant news for infrastructure and operations professionals.
Oracle’s Zettascale10 supercomputer connects 800,000 Nvidia GPUs for massive-context training and inference.
The Basque Government and IBM unveiled the first IBM Quantum System Two in Europe at the IBM-Euskadi Quantum Computational Center in San Sebastián, Spain.
The DGX Spark is already sold out at Nvidia’s online store, although retail chain Micro Center has a few hundred in stock at stores scattered around the US.
The Ethernet for Scale-Up Networking (ESUN) initiative includes AMD, Arista, ARM, Broadcom, Cisco, HPE Networking, Marvell, Meta, Microsoft, and Nvidia.
The 800G Thor Ultra NIC implements Ultra Ethernet Consortium specifications and introduces hardware-accelerated capabilities to modernize RDMA.
The collaboration combines custom silicon and open networking, potentially reducing reliance on Nvidia and reshaping next-generation AI data centers.
At the Open Compute Project's Global Summit this week, the semiconductor darling is announcing details of its newest superchip architecture.
Intel goes for cores over threads for new generation of Xeons.
The Cisco 8223 router is based on the 51.2Tb Silicon One P200 programmable chip.
AI agent and other AI capabilities can help IT teams make progress towards autonomous operational resilience.
Wide-ranging software updates from IBM are aimed at helping enterprises scale their AI operations faster and stay on top of emerging quantum risks.
The vendor's universal ZTNA solution aims to help enterprises secure diverse device ecosystems across hybrid environments.
It’s not a matter of being better than Nvidia, because Nvidia chips are becoming virtually impossible to get.
Nvidia and Fujitsu plan to build an AI infrastructure for healthcare, manufacturing, and robotics
IT leaders face rising downtime costs, fragmented tools, and cultural hurdles as they strive toward AI-driven observability, new research finds. Plans to consolidate onto unified observability platforms are in the works.
Intel casting a wide net to make deals government, tech giants and others to survive
IBM Network Intelligence uses AI to gather network telemetry and agents to help with reasoning, root-cause analysis and remediation.
Altera hardware and developer tools are launched as part of the company's newly established independence.
There’s no evidence of exploitation but sysadmins will want to check their environments.
Security improvements and confidential computing enhancements are also part of OpenStack 2025.2, known as Flamingo.
Netgear is embedding firewall software gained in its acquisition of Exium into its router platform to eliminate the need for separate firewall appliances or virtual machines at branch locations.
A new browser extension brings Cato Networks' zero-trust policies and enforcement capabilities to unmanaged and BYOD devices.
Newly launched MCP server provides structured, high-level access and orchestration capabilities, while Alkira’s NIA copilot guides network administrators through configuration and troubleshooting.
Zayo DynamicLink combines network ownership with software-defined delivery.
With services offline for a week, the incident may reshape safety standards, insurance premiums, and infrastructure design, forcing enterprises to balance efficiency with resilience and regulatory compliance.
HighPoint’s Rocket 7638D PCIe 5.0 switch card exploits Nvidia's GPUDirect interconnection to communicate with storage, bypassing CPU
Researchers uncover problems with firmware validation checks and Root of Trust.
Attackers have already exploited the SNMP vulnerability to execute remote code or deny service.
Equinix rolls out fabric intelligence for real-time awareness and automation for AI and multicloud workloads.
Cisco software includes a compiler that enables quantum workloads to run in distributed environments, including multivendor quantum compute nodes.
New features unveiled at Pure//Accelerate promise easier cloud migration, smarter Kubernetes ops, and built-in threat detection, helping CIOs manage data sprawl and AI risk across hybrid environments.
Two systems are better than one, say vendors. But what about vendor lock-in?
The microfluidic cooling system tested by Microsoft promises up to 3x better heat removal than cold plates, improved power usage, and reduced operational costs.
Investment in AI continues to grow, but just one in 10 AI projects is fully deployed.
Nvidia will invest up to $100 billion in OpenAI, and OpenAI will build and deploy at least 10 gigawatts of AI data centers with Nvidia systems.
Australia's telecommunications regulator is investigating an outage at Optus after an upgrade went wrong and network monitoring failed to spot the problem.
Nvidia's $5 billion Intel investment will have long-term implications for enterprise data center and networking products.
Nvidia intends to combine Enfabrica’s tech stack with its own hardware and software to offer more efficient training clusters for more powerful frontier models, analysts say.
Upscale aims to challenge Nvidia's dominance in AI networking infrastructure with an open solution that builds on SONiC and other standards-based initiatives.
LogicMonitor, Honeycomb separately update observability offerings with AI tools to reduce alert fatigue and speed developer processes.
The impact for enterprises in terms of chip availability, pricing and functionality is unclear, but it all-but kills worries about Intel’s near-term viability.
Once it gets rolling, the AI avalanche is unstoppable. In 2026, global AI spending is forecast to top $2 trillion, according to Gartner.
buzzword bingo, user errors top irritating problems IT pros face list, according to SolarWinds study
Power availability and infrastructure delivery timelines are making it extremely difficult for data center inventory to keep up with demand.
Analysts say enterprises may face higher GPU costs and delivery delays by late 2025 as Nvidia absorbs potential revenue losses.
Arista executives discuss liquid-cooled networking gear and support for Linear Pluggable Optics to achieve power and cost optimization for AI data centers.
The growing AI Processor market has seen over $13.5 billion in start-up funding, but rapid consolidation is expected in coming years
CalypsoAI’s platform to provide real-time threat defense against attacks and data security guardrails to help enterprises securely deploy artificial intelligence and generative AI.
Google's new Data Transfer Essentials is designed to enable enterprises to transfer data between clouds at no cost for supported Google services, aligning with the EU Data Act and aiming to reduce vendor lock-in.
Nebius deal promises to deliver additional resources to Azure customers.
