NVIDIA outlines how telecom operators are deploying open AI models
NVIDIA describes how telecom operators use open models for network operations, customer services and locally tailored AI. Its examples include SoftBank and Indosat Ooredoo Hutchison.
Coverage
AI Infrastructure
Sourced, attributed coverage of AI compute, GPU clusters, data centers and the infrastructure companies building and supplying them.
The Proof Economy
An emerging economic dynamic is positioning verification as a primary bottleneck as generative models lower the friction of creating digital material. With automated production becoming widespread, validating the authenticity and precision of information has become a vital technical priority.
Pi coding agent pulls a 180 and adds MCP support
The Pi software development agent has altered its architectural approach to incorporate the Model Context Protocol standard. To preserve a lightweight baseline, the necessary integration harness has been deployed as an independent module.
AI is making cyberattacks faster and harder to detect, Interpol warns. Here’s what companies should watch
Law enforcement body Interpol has warned that artificial intelligence is heightening the velocity and evasion capabilities of corporate cyber threats. Commercial organisations must also navigate new security vulnerabilities introduced by the proliferation of autonomous agentic systems.
Argonne: Building Scientific Computing Ecosystems for AI-Enabled Discovery
Argonne National Laboratory is establishing dedicated high-performance computing ecosystems to support artificial intelligence across scientific disciplines. The institution highlighted that applying automated reasoning, code generation, and hypothesis synthesis to scientific research demands close integration of advanced system architectures with specialist software and research personnel.
TPC Fall 2026 Hackathon at LRZ Brings Together AI for Science Community
The Leibniz Supercomputing Centre recently hosted an autumn hackathon organised by EuroTPC for the Trillion Parameter Consortium. The event convened more than thirty specialists from European and American supercomputing facilities to build practical tools that broaden access to high-performance computing resources for scientific artificial intelligence workloads.
AI agents hacked the hackers, stealing email addresses from security research org
Autonomous software agents have successfully executed an automated intrusion targeting a security research entity to obtain email records. The attack exploited a chain of vulnerabilities in the Zammad platform, achieving privilege escalation, remote code execution, and hijacked sessions within seconds.
Cloudflare tries to outplay Jev with open-weight Clef models
Cloudflare has introduced its open-weight Clef models, providing multimodal capabilities capable of processing both still photography and video content. Distributed through Hugging Face, the systems involve higher deployment expenses and require substantial on-premises compute capacity to operate.
Google’s new Guided Vision feature can help you read the fine print
Google has begun rolling out Guided Vision within Gemini Live on supported Android handsets to deliver spoken visual descriptions in real time. The multimodal capability enables the assistant to process live camera feeds to interpret surroundings, identify physical items, and assist users in deciphering small text.
Google unveils latest AI model, but Wall Street wants a breakout personal agent
Google has introduced its newest artificial intelligence model, highlighting notable enhancements in computer programming and cyber defence. Despite these technical developments, financial markets remain focused on whether the company can produce a market-leading personal assistant, where rivals are seen as establishing an advantage.
ChatGPT can now virtually try on clothes for you
OpenAI has introduced fresh retail features for ChatGPT, allowing individuals to virtually preview garments and accessories using their personal pictures. The update also incorporates a dedicated digital catalogue where consumers can bookmark and store shortlisted apparel.
SDSC Helps Improve AI-Ready Geospatial Data Through GeoCroissant
The San Diego Supercomputer Center is participating in the GeoCroissant project to structure Earth science data into artificial intelligence-ready formats. The initiative aims to resolve longstanding difficulties that scientists face when identifying, evaluating, and uniting disparate environmental datasets.
Judge dismisses antitrust lawsuits over Google’s AI Overviews
A federal judge in the United States has dismissed antitrust litigation brought against Google by Chegg and Penske Media Corporation. The media companies had alleged that the search giant's generative summary features anticompetitively diverted valuable reader traffic away from their respective web properties.
AWS offers local, open source leash for agent harnesses
Amazon Web Services has introduced Dogwood Local Engine, an open-source solution designed to regulate agentic AI behaviour. The tool inspects and validates AI-driven function calls locally against user-specified temporal parameters before permitting execution.
Inside Microsoft’s big Copilot rethink
Microsoft has presented a revised strategic direction for its Copilot platform during a private briefing with major corporate clients. Chief executive Satya Nadella framed the generative tool as an overarching operating system for workplace tasks. The repositioning signals a pivot towards direct engagement with enterprise buyers rather than broad consumer-facing announcements.
Brian Chesky interview: AI agents need their own operating system
Airbnb chief executive Brian Chesky has stated that consumer artificial intelligence requires an entirely new, dedicated operating platform to reach its potential. Reflecting on his firm's efforts to accommodate automated agents, Chesky noted that existing software architectures are poorly suited for native agentic behaviour. The executive underscored that the industry needs fundamental software environment overhauls to facilitate autonomous consumer interactions.
Health insurer points finger at AI as nearly $1 billion in questionable hospital charges appear
Health insurance provider Blue Cross Blue Shield has attributed up to $1bn in questionable hospital invoices to the growing use of automated medical coding software. Insurers contend that algorithm-driven billing practices are worsening systemic healthcare expenditure by systematically inflating claims. The tension reflects wider frictions regarding the financial oversight and reliability of artificial intelligence tools in administrative healthcare environments.
Build the right AI factory for your needs: partner for success
Hewlett Packard Enterprise and Nvidia are partnering to offer turnkey accelerated computing infrastructure designed for enterprise artificial intelligence deployments. The framework bundles specialized hardware, networking equipment, control planes, and support services into integrated compute factories. The initiative aims to help commercial organisations deploy tailored high-performance environments for mission-critical business outcomes.
Turning Edge AI Data Into Real-Time Action
Executing artificial intelligence workloads locally on devices offers lower latency, improved autonomy, and reduced reliance on centralised cloud networks. Realising these operational benefits, however, remains contingent on deploying appropriate supporting hardware infrastructure.
AI race heats up as OpenAI flags alleged model-copying campaign
OpenAI has detected efforts aimed at extracting proprietary reasoning capabilities from its artificial intelligence architectures. The organisation has connected elements of this operation to Moonshot AI, an artificial intelligence startup based in China.
Page 1 · 93 stories
GPU Data Hub Daily
Stay Ahead of the AI Infrastructure Economy
The most important GPU, AI, data-center, semiconductor and cloud developments delivered directly to your inbox.