Latest News - Page 5 - DataCenterNews US

Nvidia says software cuts DeepSeek V4 token costs fivefold

Software improvements have slashed the cost of serving DeepSeek V4 on Blackwell, underscoring the race to make AI deployments economical.

By Mark Tarre • 4 min read • This month

Hybrid Cloud

Google Cloud expands Spanner for AI & multi-model data

Businesses can now run transactional, graph and vector workloads together as Google broadens Spanner for AI applications across clouds and on-premises.

By Mark Tarre • 5 min read • This month

Smart Cities

Nvidia launches vision AI agent blueprints for industry

Shortages of training data and engineering effort are slowing industrial vision AI projects, prompting Nvidia to package reusable blueprints for developers.

By Mark Tarre • 5 min read • This month

Data centers

AI demand to push memory fab spending past USD $50bn

Memory makers are set to spend more than USD $50 billion on 300mm fab equipment in 2026 as AI demand drives HBM and DDR5 capacity.

By Sean Mitchell • 3 min read • This month

DevOps

Google launches remote MCP server for Gemini Enterprise

External coding tools can now reach Google Cloud models and notebooks via an open protocol, while IAM Deny policies keep access tightly governed.

By Mark Tarre • 4 min read • This month

Edge Computing

AMD launches Versal Premium Gen 2 with on-package memory

The chips aim to cut board space and design complexity for defence, testing and video systems that need high bandwidth in compact form factors.

By Joseph Gabriel Lagonsin • 4 min read • Last month

Robotics

Claude models go live on Microsoft Foundry via Azure

Azure customers can now deploy Claude for governed enterprise agents as Microsoft Foundry widens access to Anthropic's models on NVIDIA GB300 hardware.

By Mark Tarre • 3 min read • Last month

Hybrid Cloud

Palantir & NVIDIA launch air-gapped AI for agencies

U.S. agencies can now train and keep control of AI models on isolated systems, with Palantir and NVIDIA targeting sensitive government work.

By Sean Mitchell • 3 min read • Last month

Semiconductors

Gimlet Labs joins MLCommons to shape agentic inference

The partnership could shape how buyers compare AI hardware as agentic workloads demand more realistic, vendor-agnostic performance tests.

By Mark Tarre • 4 min read • Last month

Data centers

Wasabi launches Impact Circle to track cloud emissions

Partners can now measure and offset storage-related emissions as cloud providers face mounting scrutiny over AI-driven infrastructure use.

By Joseph Gabriel Lagonsin • 4 min read • Last month

Edge Computing

FAR Labs opens access to cheaper AI inference platform

Developers facing rising AI bills can now register for early access to FAR Labs' platform, which claims lower inference costs on some models.

By Joseph Gabriel Lagonsin • 4 min read • Last month

DevOps

AI infrastructure incidents hit 93%, Spacelift warns

Most firms are unprepared for AI-driven infrastructure risk, as Spacelift found only 19% have the governance needed to curb incidents.

By Joseph Gabriel Lagonsin • 4 min read • Last month

DevOps

Minimus opens full image catalogue without registration

Developers can now pull thousands of hardened container images for free, as the company drops registration and expands access across its library.

By Sean Mitchell • 4 min read • Last month

Data centers

SiTime launches Chorus 2 clock generators for AI systems

Timing in AI and networking boards is getting simpler, as the new devices can replace up to 12 oscillators and cut power use.

By Sofiah Nichole Salivio • 3 min read • Last month

Network Infrastructure

Indigo appoints Nick Barton as Chief Revenue Officer

The hire is aimed at sharpening Indigo's push into hyperscaler and network accounts as competition intensifies across international connectivity markets.

By Sofiah Nichole Salivio • 3 min read • Last month