Network World reports on the most significant news for infrastructure and operations professionals.
Calling it the highest performance chip of any custom cloud accelerator, the company says Maia is optimized for AI inference on multiple models.
Neurophos expects the funding to speed up manufacturing development of its chips, which could be used to accelerate AI.
Intel expects CPU supply will improve after the first quarter
After changing its requirements for high-bandwidth memory chips, Nvidia is close to certifying a new source for supplies.
Enterprises are facing storage shortages, long lead times, and dramatic price increases, which may cause some to defer on-prem AI projects.
The tech giant is pivoting its strategy to focus on processors to support AI workloads, meaning less capacity for chips for PCs.
Analysts and AWS watchers agree that the roughly 15,000 cuts are caused by ‘culture’, but there is disagreement about what that actually means.
SATA is being replaced in high-performance storage situations but is not disappearing anytime soon
New Rust release brings major improvements to musl’s DNS resolver and tweaks the standard library to aid global allocators.
The distributed cloud networking market is forecast to hit $21 billion by 2029 as enterprises abandon fragmented WAN architectures.
After months of uncertainty, Intel comes back with a big hire and reaffirms its commitment to GPU technology.
New Relic’s latest observability capability provides enterprises with visibility into the performance, user experience, and reliability of applications embedded within ChatGPT.
The move underscores how satellite networks are increasingly being viewed as a resilience layer for enterprise WANs and data center connectivity.
Infrinia AI Cloud OS automates Kubernetes and inference services on GPU infrastructure.
The company plans to build and pay for dedicated energy infrastructure to reduce the risk of AI data centers straining local grids and slowing expansion.
Cisco is integrating its Nexus 9000 Series AI networking switches with Intel's Gaudi 3 AI accelerators.
The tie-in expands processing options for AI data centers, but leaves rival interconnect UALink out in the cold.
F5's Guardrails blocks prompts that attempt jailbreaks or injection attacks, and its AI Red Team automates vulnerability discovery in AI systems.
AWS’s new sovereign cloud for Europe boosts compliance controls, but analysts say its US ownership raises unresolved questions about legal authority and service continuity.
The latest version of IP Fabric's network assurance platform models cloud-native networking and security constructs as first-class objects in the digital twin.
Chinese hackers are behind attacks on AsyncOS Software that allow complete takeover of the appliance, company says.
A previous virtually identical zero day DoS vulnerability was targeted in the wild, and there's already a PoC for this one.
Uptime Institute warns that the gap between data center construction timelines and power infrastructure availability will exacerbate industry challenges.
Dell PowerStore adds storage density and promises to reduce total cost of ownership by 15%.
With utilities quoting four‑to‑ten‑year wait times — and one offering a 12‑year study period — Google says grid constraints are becoming the defining limit on AI‑era data center growth.
The multibillion-dollar deal shows how the growing importance of inference is changing the way AI data centers are designed and operated.
By decoupling sovereignty from dedicated data center regions, IBM's Sovereign Core aims to help CIOs meet rising regulatory scrutiny, automate continuous compliance, and deploy sensitive AI workloads into production under strict data‑residency condit
Just as the US approves exports to the country, Chinese customs officers have reportedly been told not to allow the chips into the country. What now?
Former Akamai engineers are behind startup IO River, which is developing an edge platform that creates a virtualization layer across multiple CDN providers.
The enterprise remains a key target for Cisco, which is keen to deliver the technologies to enable AI transformation, digital resilience, and workforce transformation.
Zhipu’s GLM-Image demonstrates the viability of domestic alternatives amid US export restrictions.
The company has unveiled a community-first AI infrastructure strategy promising no increases in local power or water costs, highlighting the mounting pressure on tech companies to balance AI growth with environmental and social impact.
A Texas energy company is proposing to use nuclear reactors from decommissioned Navy vessels to power data centers. Repurposing the reactors will be prohibitively difficult, critics warn.
A new top-level effort unifies data center and network oversight as the company targets multi-gigawatt AI scale.
Cloudflare argues that the requirement is technically unreasonable and would impair DNS performance globally.
Announcements include new HPE Aruba Networking CX switches, integration with Juniper Mist AIOps to optimize connectivity across distributed store environments, and upgrades to HPE Nonstop Compute to improve system availability.
Economic uncertainty, elevated interest rates, and AI adoption have driven workforce reductions across tech companies worldwide, according to a RationalFX report.
The Cloudflare hiccup involved a sequencing change that confused some systems, and Cloudflare quickly walked it back, but the incident showcases enterprise network fragility in an uncomfortably concrete way.
A successful attacker could view sensitive information that even admins can’t access.
The vendor has issued a patch to close four holes in its flagship Backup & Replication suite; version 13 users are advised to audit their backup config files and closely monitor backup jobs.
New Lenovo servers cover high-performance tasks and edge computing.
In the past, AWS has almost always lowered prices or changed pricing models instead of raising them directly.
With capacity shifting to high-margin HBM for AI data centers, traditional DRAM supply is collapsing, pushing enterprise IT costs sharply higher and eroding procurement leverage.
Lenovo’s “AI cloud gigafactory” model promises speed and scalability for enterprises, but experts caution that utility, cooling, and fiber bottlenecks could slow adoption.
As AI networking requirements grow, the Ethernet Alliance will take up a 400G/lane project while the Ultra Ethernet Consortium will prioritize flexible congestion management and small-message performance.
ThousandEyes tracks internet and cloud traffic and provides Network World with weekly updates on the performance of ISPs, cloud service providers, and UCaaS providers.
AMD announced new products, acquisitions and partnerships in 2025 as it worked to boost AI performance across platforms ranging from desktops and mobile devices to data centers and supercomputers.
The Nvidia Rubin platform consists of six new chips, including four networking processors.
The move could reshape cost planning for enterprises running large-scale AI workloads.
Announcements included the MI440X GPU for on-prem enterprise deployments and a longer-term roadmap that includes rack-scale AI systems and next-generation accelerators.
