Network World reports on the most significant news for infrastructure and operations professionals.
Neutral atom company QuEra has run a simulation showing a dramatic reduction in the number of physical qubits needed to create one logical qubit, but for memory, not computation.
The new EL2000 chassis and DL145 Gen11 are purpose-built to push distributed AI inference, computer vision, and IoT processing to factory floors, power plants, and remote outposts.
Latest Cisco SD-WAN software adds network protection services and more AI awareness.
EMA research shows chatbot-based AIOps is losing ground to agentic AI, with collaborative, proactive models delivering faster operations, fewer errors, and stronger results.
Auvik is bringing agentic AI to network and IT operations with its Aurora platform.
Admins are being warned by cyber experts from the US and UK that this is part of continuing campaign to crack Cisco firewalls.
AI-related traffic moves differently, especially during training cycles. To cope with sudden spikes, network managers have to rethink their infrastructure.
The company is assembling a multi-architecture stack spanning AWS, Nvidia, AMD, Arm, and its own silicon. In the agentic era, no single chip wins, so it’s betting on all of them.
Cirrascale customers can run Gemini inferencing locally, preserving data sovereignty.
We're keeping tabs on efforts to launch AI into orbit, including new customers for the largest orbital compute cluster, the first commercial space data storage service, and a proposed marketplace for AI inference capacity in space.
Cisco's Universal Quantum Switch runs at room temperature and links resources over standard telecom fiber.
Major data center projects from Microsoft, Oracle, OpenAI, and others are expected to miss target opening dates.
The new eighth‑generation TPUs mark a shift away from one‑size‑fits‑all accelerators, targeting distinct cost, memory, and networking needs for training versus inference.
SUSE and Cloudbase promise zero-downtime migrations off VMware, but whether enterprises are ready to move is another question.
Zero Networks extends microsegmentation to govern AI agents and block AI-driven attacks.
A multi-tenant authentication gap in Microsoft’s AI operations agent exposed live command streams, internal reasoning, and credentials to any Entra ID account, researchers said.
The investment underscores a shift from pure funding to securing long-term compute, as Anthropic looks to ease capacity constraints and lock in infrastructure amid intensifying demand for AI workloads, analysts say.
A slew of updates from Cloudflare includes its Mesh private networking fabric for AI agents and its Registrar API interface, which allows domain registrations to be handled by AI agents.
Cisco report finds 86% of organizations struggle to hire wireless talent, increasing security risks and costs as AI draws skills away from networking roles.
Logs in wireless access point flash memory grow by 5MB a day in certain IOS XE devices until space runs out.
Not so fast, say Cloudflare and APNIC.
Broadcom will provide the networking backbone to support Meta’s in-house AI chips.
With the cost of data centers and the computing equipment that fills them on the rise, chasing tax breaks could be a way for enterprises to optimize their costs.
Equinix Fabric Intelligence helps enterprises design, deploy, and manage their infrastructure
SolarWinds report finds 80% of IT professionals say AI is shifting roles from operator to orchestrator, while challenges around trust, governance, and skills persist.
New IBM security services aim to help enterprises identify risks introduced by frontier AI models that can discover vulnerabilities and launch autonomous attacks.
Analysts and consultants see it as taking a more conservative fiscal approach as OpenAI prepares for an imminent IPO.
The proposed statewide moratorium would pause new approvals through 2027, challenging hyperscalers and enterprises betting on rapid infrastructure buildouts.
Amazon’s $11.6 billion acquisition of Globalstar is set to bring D2D services to Leo, its low Earth orbit satellite network.
By bulking up AI observability with Galileo and identity with Astrix, Cisco is repositioning its secure network as the control plane for the agentic era. It's a calculated bet that the company best known for building the internet's infrastructure can
Data center construction has never been more challenging, between local revolts and component shortages.
According to the latest Google research, it could take as few as 1,200 logical qubits for a quantum computer to break elliptic curve cryptography. That means enterprises are up against the difficult challenge of securing their authentication mechanis
Why Linux 7's quiet TCP change could be the best thing to happen to your network.
Follow Intel news as the company continues to focus on foundry leadership and regaining market momentum in the processor sector.
AWS is betting that its Trainium chips can reshape the economics of training and inference, even as it retains a strong partnership with semiconductor giant Nvidia.
Intel and Google extended their existing partnership that includes smart networking controllers as well as CPUs.
Concerns about how local regulations will affect AI are also a factor.
Broadcom will help develop TPUs and provide them to Anthropic.
More than 60% of global AI compute capacity sits with hyperscalers, led by Google, raising new questions about control, pricing power, and access.
Galileo's technology will slide into Cisco’s Splunk observability portfolio for improved AI agent visibility and protection.
Neocloud revenues are rapidly scaling -- reaching $9 billion in Q4, according to Synergy Research Group
Threat research from Lumen Technologies focuses on telemetry from the internet backbone, where the largest ISPs interconnect, to reveal attack patterns that traditional endpoint security deployments cannot surface.
Likely delay in Nvidia’s Rubin GPUs may leave enterprises facing slower access to next-gen infrastructure and extended reliance on current systems.
Startup Aria Networks launches Deep Networking platform for AI factories, targets token efficiency.
The $25B initiative could redefine chip manufacturing at scale, but supply constraints and long timelines likely mean IT buyers won’t feel its effects for years.
Nvidia still holds the upper hand, but UALink ‘remains desirable’ as a way to enable multi-vendor environments, says an analyst.
Anthropic’s Project Glasswing will bring together a group of vendors to help define how AI resources will be protected from cyber threats.
With Nvidia now controlling Slurm’s roadmap, enterprises running mixed-vendor GPU clusters are asking whether open-source guarantees are enough.
Wireless infrastructure teams are under pressure to keep up with AI networking and security challenges.
Data center capital expenditures grew 57% last year and hit a record high, according to Dell'Oro Group. Backlog figures from Amazon and Google suggest AI-fueled data center investments will continue to surpass expectations.
