Ci risiamo, ennesimo caso di scheda grafica NVIDIA top di gamma danneggiata dal famigerato cavo di alimentazione. Questa volta sembra che sia successo con la variante aggiornata, quindi il cosiddetto 12V-2x6 al posto del precedente 12VHPWR, pensato proprio per risolvere i suoi problemi. L'ultimo episodio coinvolge una fonte particolarmente nota nella comunità hardware. Il sito specializzato Club386 ha documentato il danneggiamento di una RTX 5090 Founders Edition durante una sessione di test.
Le immagini pubblicate mostrano pin anneriti, plastica deformata e danni evidenti sia sul connettore della scheda video sia sul cavo di alimentazione. Secondo quanto riportato, il sistema era alimentato da un Be Quiet Dark Power 13 da 1000 watt collegato tramite un singolo cavo 12V-2×6 nativo, correttamente installato e conforme alle specifiche previste.
Il 12V-2×6 era stato progettato per migliorare la sicurezza del precedente 12VHPWR grazie a pin di rilevamento modificati e a una forma migliorata dei pin di potenza. Tuttavia, questo risolve solo parte del problema. Le GeForce RTX 40 e RTX 50 non hanno circuiti di monitoraggio del carico, quindi se per qualsiasi ragione uno dei sei pin di potenza non fa bene contatto tutta l'energia viene trasferita attraverso i restanti cinque. Il problema è che i fili elettrici hanno una sezione molto al limite calcolando che una RTX 5090 può arrivare ad assorbire anche 600 W nei carichi AI o nel gaming.
A indústria da beleza está a evoluir para um modelo mais preditivo, centrando-se na capacidade de antecipar as necessidades da pele através de dados, diagnóstico e IA. Ao TEK Notícias, Guive Balooch explica como esta mudança impacta o desenvolvimento de novos produtos e a relação com os consumidores.
A Nvidia continua a apostar no ecossistema de robótica e apresentou o primeiro sistema de segurança que prepara o ambiente para os robots agirem em ambientes dinâmicos, ao lado de trabalhadores humanos.
Nvidia and telecom partners including SoftBank Corp and NTT Data plan to demonstrate a stack of data, models, simulation tools and secure runtimes at DTW Ignite 2026, aimed at helping operators move from task-based automation to fully autonomous network operations.
The company argues while generative AI has delivered strong returns in network management, customer care and back-office automation, most of the impact has been limited to speeding up predetermined steps while humans still correlate insights and direct next actions.
Nvidia’s claims automation is now “the launchpad to autonomy” rather than the end goal, with AI agents expected to proactively monitor for problems and coordinate changes across network, IT and business systems while keeping humans in control of policy.
Nvidia cites figures showing 54% of telecom operators view data-related issues as their biggest obstacle to building telecom-specific AI models, largely because the most valuable network and customer data is too sensitive to use directly for training.
SoftBank Corp is tackling the problem by using Nvidia’s NeMo Safe Synthesizer and NeMo Anonymizer tools to generate privacy-preserving synthetic datasets which mirror the structure of real network performance and configuration datasets.
Those datasets are being used to fine-tune its Large Telecom Model (LTM) and build specialised network agents.
On the runtime side, Nvidia stated its NemoClaw blueprints and OpenShell secure runtime are designed to give long-running agents policy-based guardrails and sandboxed access to telecom systems, allowing operators to expand agent responsibilities while keeping behaviour auditable and governed.
NTT Data is applying Nvidia’s open Nemotron models alongside NemoClaw to build agents which track long-term network performance trends and escalate anomalies to specialised research agents for deeper telemetry analysis.
Other partners building on the same framework include: AdaptKey, which is piloting self-healing 5G agents with operators; ServiceNow, which is bringing its Project Arc incident-response tooling to telecom network operations centres; and Tata Consultancy Services, which is developing a multi-fidelity AI sensor architecture for faster anomaly detection.
Simulation is the third pillar of the effort, intended to let agents test recommendations in near-real-time digital environments before acting on live systems.
Software company Forsk stated its GPU-accelerated radio propagation model now achieves ray-tracing-level accuracy up to 200 times faster than CPU-only systems, while Viavi Solutions reports order-of-magnitude gains in RAN simulation throughput after shifting workloads to Nvidia’s RTX Pro 6000 Blackwell GPUs.
KDDI and KDDI Research, working with Nvidia, Keysight and Samsung Research America, are building a high-fidelity RAN digital twin aimed at the 6G era, allowing multiple autonomous agents to simulate scenarios for future radio conditions and traffic shifts.
The TM Forum’s DTW Ignite conference runs 23–25 June 2026 in Copenhagen.
Nvidia unveiled what it described as the industry’s first full-stack safety platform for physical AI designed to accelerate the deployment of humanoid robots in industrial environments.
The company said its expanded Halos platform combines AI compute, system software, sensors and inspection tools into a single safety architecture aimed at helping robotics companies build and certify autonomous systems operating alongside human workers.
The chipmaker explained as robots become more autonomous and increasingly rely on AI foundation models, distributed sensors and accelerated computing, “scaling these systems requires a full-stack safety architecture” before they can be trusted to work alongside humans in factories, warehouses and logistics operations.
The stack includes Nvidia’s IGX Thor industrial computing platform, Holoscan Sensor Bridge for sensor connectivity and Halos OS software stack.
Nvidia also opened its Halos AI Systems Inspection Lab to robotics partners, aiming to help them complete pre-certification testing and validate safety systems before approaching regulators.
Digit Humanoid robotics company Agility Robotics is the first partner to adopt elements of the platform into the proprietary safety system for its Digit humanoid robot, which is already being deployed for industrial applications with customers including Amazon, Schaeffler and Toyota Motor Manufacturing Canada.
The robotics player will also use Nvidia’s Inspection Lab to validate Digit’s safety-related software, AI components and cybersecurity protections against strict standards.
Agility Robotics CEO Peggy Johnson said: “For humanoids to deliver value at scale, safety has to be built into the robot and validated across the entire system.”
She added the partnership with Nvidia “extends our leadership in responsible automation”, describing it as “a non-negotiable requirement for bringing humanoids safely into industrial workflows”.
Deepu Talla, VP of robotics and edge AI at Nvidia said: “Physical AI is transforming how factories, warehouses and logistics operations work. He added developers will be able to “harness Nvidia’s proven autonomous vehicle safety foundation to develop safer robots faster” and deploy systems “alongside workers with greater confidence”.
Does a thought-experiment about US ascendancy in the technology say as much about AI jitters as it does about the reality?
It’s 2031 and the US and China are about to tear Europe into pieces.
The US ploughed vast sums into datacentres and the EU did not. China built robots and Europe did not. American companies “restructured” their workflows around AI and fired people, while EU workers went on long lunch breaks and handed over administrative tasks to the AI model Claude.
La scarsa disponibilità e l’elevatissimo valore delle schede video gaming di ultima generazione continuano a rappresentare un terreno fertilissimo per i truffatori che non si fanno scrupoli a sfruttare gli utenti più sprovveduti: in Cina il noto tecnico e rivenditore conosciuto come Brother Zang si è trovato per le mani una presunta GeForce RTX 4090 acquistata per l’equivalente di circa 200 dollari che si è rivelata essere un falso particolarmente elaborato.
A una prima occhiata la scheda sembrava in tutto e per tutto autentica: il PCB era molto credibile e le serigrafie del codice modello sulla GPU erano corrette (sigla AD102-300-A1). Tuttavia, un’analisi più approfondita ha fatto emergere diversi dettagli che non quadravano. Tanto per cominciare, il seriale della data di produzione indicava che la GPU era stata prodotta nel 2030, che, al netto di essere in possesso di un flusso canalizzatore, è ragionevolmente impossibile. Mancavano inoltre alcuni dettagli identificativi normalmente presenti sui chip originali, come specifiche incisioni e codici QR di verifica.
creating a fake 4090 by repurposing 3080/3090 die is nothing new. whats new, however, is that this specific example uses plastic to imitate the silicon die. there is no "glue"/"adhesive" surrounding the silicon.
Nvidia ha mostrato una nuova generazione di robot capaci di apprendere autonomamente compiti complessi nel mondo reale, aprendo la strada a sistemi sempre più indipendenti nell’esecuzione di attività manuali di precisione, come per esempio l’installazione di una scheda grafica su una motherboard. La dimostrazione arriva dal progetto ENPIRE, una piattaforma sperimentale che combina intelligenza artificiale agentica e robotica fisica per consentire alle macchine di migliorare le proprie capacità direttamente attraverso l’esperienza.
Nel video condiviso su LinkedIn da Jim Fan, direttore dell’AI Research di Nvidia, si possono osservare diversi robot impegnati in operazioni che richiedono coordinazione, precisione e adattamento. Tra gli esempi mostrati abbiamo, oltre all’inserimento di una scheda grafica in una scheda madre, l’organizzazione di piccoli pin metallici in appositi contenitori e l’installazione di fascette da elettricista, incluso il taglio del materiale in eccesso.
Alla base del progetto c’è un approccio che Nvidia definisce AutoResearch. In pratica, un gruppo di agenti AI riceve un obiettivo da raggiungere e dispone di una vasta flotta di robot, GPU e token per trovare autonomamente la soluzione migliore. Gli agenti osservano l’ambiente, sperimentano, correggono gli errori, consultano documentazione tecnica, analizzano risultati precedenti e affinano progressivamente le strategie operative. Secondo Jim Fan, direttore dell’AI Research di Nvidia, il progetto rappresenta uno dei primi esempi concreti di auto-addestramento, strumento classico del mondo AI, applicato direttamente al mondo fisico.
En 2026, la RTX 5070 est la carte de référence pour les joueurs exigeants : elle fait tourner tous les grands titres en 1440p Ultra avec le ray tracing activé, sans compromis sur la fluidité. En dessous, la RTX 5060 couvre largement les jeux compétitifs et les AAA modérés en 1080p-1440p....
What happens when you give AI coding agents a lab full of robotic arms, some compute resources, and a “generous token budget” for teaching the robots various tasks? The agents can apparently figure out a training regimen that teaches the robots to successfully cut zip ties and even insert GPUs into thin sockets on motherboards.
That glimpse into how AI can act in a fully autonomous way to automate robot training was made possible by a new agent harness framework—software that wraps around AI models to enable their use of various tools while also providing capabilities such as memory, context, constraint, and feedback loops. That agentic harness, called ENPIRE, was developed by robotics researchers at the Nvidia GEAR (Generalist Embodied Agent Research) lab alongside collaborators from Carnegie Mellon University in Pittsburgh and the University of California, Berkeley.
“A part of our NVIDIA GEAR lab now self-improves tirelessly overnight,” wrote Jim Fan, director of AI at NVIDIA, in a LinkedIn post. “We just read the reports in the morning.”
Nvidia reportedly told Chinese customers its new Vera CPU for AI data centres could be available in August, in a move to regain ground in a market where its advanced GPU business has been stalled by US export curbs.
Reuters reported the chipmaker is pitching Vera as an alternative route into China after shipments of its H200 AI chip stalled for months. CEO Jensen Huang (pictured) previously said Nvidia’s market share in China had effectively fallen to zero, reflecting tighter US restrictions and Beijing’s push to support domestic suppliers.
The news site noted Vera is Nvidia’s first standalone CPU for agentic AI workloads, with the company claiming it delivers up to 1.8 times the performance of rival processors.
Huang unveiled Vera in March as part of Nvidia’s Vera Rubin architecture and cast the CPU as a major new growth engine, as the company looks to expand beyond GPUs and into the server processor market.
Sources told Reuters one major Chinese cloud provider is planning to order more than 300 servers containing two Vera CPUs each for testing, with the chips set to be deployed first in overseas data centres before any wider commitment.
The move also sharpens Nvidia’s challenge to Intel and AMD, which dominate the server CPU market.
Because Vera is based on Arm architecture rather than x86, Nvidia is entering the segment as AI infrastructure demand shifts further toward inference and agentic computing, where CPUs are becoming increasingly important alongside accelerators.
Reuters added selling CPUs into China may prove easier for Nvidia than shipping high-end GPUs, which remain subject to stricter US controls.
Broader adoption, however, is likely to hinge on software compatibility and the willingness of Chinese operators to move workloads away from domestic AI silicon.
O próximo processador de IA da Google terá uma arquitetura de fabrico dividida. A TSMC deverá produzir o núcleo computacional, cabendo à Samsung o fabrico do componente de interligação à memória. A Intel poderá também entrar na equação com a produção da TPU.
Jensen Huang fechou em Seul uma série de acordos estratégicos com SK Hynix, SK Telecom, Naver, Doosan, LG e Hyundai, reforçando a aposta da Nvidia na Coreia do Sul como pilar central da sua expansão em IA e robótica.
Au Computex, Jensen Huang n’est pas seulement venu présenter une nouvelle puce ARM pour Windows. À travers RTX Spark, le patron de Nvidia dessine surtout une vision beaucoup plus ambitieuse de l’avenir du PC, qui serait destiné à devenir comme R2-D2.
Microsoft's Build developer conference kicked off today, and as with almost everything the company has done in the last few years, Microsoft's opening keynote focused overwhelmingly on AI and other closely related technologies. There's Microsoft Scout, an OpenClaw-based "Autopilot" agent that can hook into Microsoft 365 data to perform tasks for users; several new AI models; an expanded preview of "Codename MDASH," which is a "multi-model agentic scanning system" meant to detect and fix software vulnerabilities.
A few of those announcements stood out to us as particularly interesting, either for esoteric technical reasons or because they seem like they may have some utility for those who aren't spending their every waking moment using generative AI tools. (Microsoft's recent efforts to make its flagship operating system faster, more reliable, more useful, and less annoying didn't really come up, but there have been plenty of other announcements on that front lately.)
On the hardware front, we didn't get any updates for existing Surface devices (not counting yesterday's Surface Laptop Ultra announcement), but we did get something new: the Surface RTX Spark Dev Box is "a compact developer PC" built around Nvidia's new RTX Spark chip with up to 128GB of built-in memory.
Login
