Nvidia Launches Open Agent Safety Platform with OpenShell and BlueField-4 Sentry to Prevent Rogue Autonomous AI
Nvidia has debuted an open-source runtime sandbox and hardware-isolated watchdog architecture designed to intercept errant autonomous AI agent behaviors before processes compromise enterprise networks.
NVIDIA Launches Open Agent Safety Platform With OpenShell Runtime and BlueField-4 Sentry Watchdog
NVIDIA officially introduced the Open Agent Safety Platform, combining the open-source OpenShell kernel sandbox with out-of-band BlueField-4 DPU telemetry to quarantine rogue autonomous agents within milliseconds.
NVIDIA Unveils RTX Spark Superchip and Open-Source PAIR Local AI Router to Power On-Device Agentic Computing
At IFA Berlin, NVIDIA introduced the RTX Spark personal superchip integrating a 20-core Grace CPU with a Blackwell GPU alongside PAIR, an open-source router distributing local AI inference.
NVIDIA NVLabs Introduces Native CUDA Rust Toolchains Bringing Memory Safety Directly to GPU Kernel Programming
NVIDIA NVLabs unveiled cuda-oxide and cutile-rs, launching open-source compiler toolchains that allow developers to write high-performance GPU kernels in memory-safe Rust without foreign function wrappers.
Mitsubishi Electric Unveils Chip to Grid DSX2 Power Architecture for NVIDIA Vera Rubin Supercomputing Centers
Mitsubishi Electric announced its DSX2 Chip-to-Grid reference architecture, synchronizing utility power substations and microsecond liquid cooling for NVIDIA Vera Rubin NVL72 data centers.
Fujitsu Open-Sources OpenQARP Framework to Standardize Modular Quantum Application Development
Licensed under Apache 2.0, the modular C++ and Python framework provides over 100 quantum algorithms and circuit primitives compatible with NVIDIA CUDA-Q and high-performance supercomputer simulators.
Nvidia Announces CUDA Rust to Enable Native GPU Kernel Compilation with Strict Compile-Time Memory Safety
The new official compiler frontend allows systems engineers to write high-performance compute shaders and distributed tensor kernels directly in Rust without unmanaged C++ pointers.
Cohere Launches Model Vault with Confidential Computing for Zero-Exposure Enterprise AI Inference
Enterprise AI provider Cohere debuts hardware-isolated inference environments combining Intel TDX, AMD SEV-SNP, and Nvidia confidential computing to guarantee zero cloud provider visibility into sensitive enterprise prompts.
Apple Explores Enterprise Server Return with M8 Ultra Chips and Nvidia NVLink Interconnects
Reports reveal Apple is prototyping custom AI data center server racks combining multi-chip Apple silicon with high-speed Nvidia networking.
AVerMedia Unveils Jetson-Powered Robotics Platforms for Intelligent Drones and Humanoid Automation
AVerMedia has introduced three purpose-built Physical AI robotics platforms powered by NVIDIA Jetson modules, tailored for autonomous drones, mobile rovers, and humanoids.
Qualcomm and Cornelis Networks Launch Active Compute Fabric to Challenge Nvidia AI Interconnects
Qualcomm Technologies and Cornelis Networks have unveiled a strategic partnership to commercialize Active Compute Fabric alongside a $205 million funding round, delivering open-standard scale-up interconnects to challenge Nvidia's NVLink dominance.
Agility Robotics Unveils Digit 5 Humanoid Robot with Cage-Free Safety and $300M in Orders
Agility Robotics has unveiled Digit 5, a cage-free humanoid robot featuring NVIDIA Thor edge AI, a 50lb payload, and $300M in orders ahead of a $2.5B public listing.
Trump Calls Jensen Huang Onstage at All-In Summit to Denounce AI Doomerism and Regulations
In an unscripted speakerphone call during Nvidia's CEO panel at the All-In Summit, Donald Trump pledged federal deregulation and compared datacenter compute clusters to strategic oil reserves.
NVIDIA Unveils RTX PRO 5500 Blackwell Workstation GPU with 84GB GDDR7 for Enterprise Agentic AI
NVIDIA announces the RTX PRO 5500 Blackwell workstation graphics card featuring 84GB of high-speed GDDR7 memory and 192 streaming multiprocessors, engineered for local LLM fine-tuning and agentic AI pipelines.
NVIDIA Expands AI for Media Platform at IBC 2026 With Real-Time Synthetic Video Detection and Broadcast NIMs
At IBC 2026 in Amsterdam, NVIDIA expanded its Holoscan for Media platform with microservices for real-time synthetic media verification, multilingual live speech dubbing, and automated sports intelligence playbooks.
Figure AI Secures $3.5B Compute Partnership with Nscale to Power Helix Humanoid Brain
Humanoid robotics pioneer Figure AI partnered with Nscale in a $3.5 billion infrastructure agreement, securing 100,000 next-generation NVIDIA GPUs to train its Helix physical AI model.
NVIDIA Introduces CUDA Rust at RustConf 2026 to Bring Memory Safety to GPU Kernels
NVIDIA unveiled CUDA Rust at RustConf 2026, launching two compiler backends that bring compile-time memory safety and zero-aliasing guarantees directly to high-performance GPU execution.
DOJ Probes Nvidia's $20B Groq Licensing Deal Over Potential Antitrust Evasion
Federal antitrust enforcers opened an inquiry into Nvidia’s $20 billion non-exclusive licensing transaction with Groq to determine whether the pact bypassed merger reporting requirements.
Nvidia Acquires Hugging Face for $12.93 Billion in Landmark Open Source AI Deal
Nvidia has entered into a definitive merger agreement to acquire open source AI platform Hugging Face for $12.93 billion, pairing dominant silicon hardware with community model distribution.
Nvidia Agrees to Acquire Hugging Face for $12.9 Billion in Landmark AI Development Expansion
Nvidia reached a definitive agreement on September 3, 2026, to acquire open machine learning hub Hugging Face for $12.93 billion, uniting the premier open model repository with its accelerated computing stack.
Caterpillar and FieldAI partner to bring physical AI foundation models and digital twins to heavy jobsites
Caterpillar partnered with robotics leader FieldAI on September 2, 2026, combining heavy machinery telemetry with NVIDIA Omniverse digital twins and robot-agnostic physical AI models for autonomous jobsite operations.
NVIDIA and MediaTek bring custom AI accelerators into the NVLink ecosystem
MediaTek will use NVLink Fusion for custom AI infrastructure as NVIDIA backs the partnership with a $3.5 billion investment, widening the market for non-NVIDIA compute around NVIDIA networking.
NVIDIA and MediaTek extend their partnership from local AI to cloud infrastructure
The companies are joining MediaTek's edge and high-volume silicon experience with NVIDIA networking, GPUs and NVLink Fusion, backed by a $3.5 billion NVIDIA investment.
NVIDIA's Vera CPU says agentic AI needs a new hardware baseline
Vera is shipping now, and NVIDIA is treating that as proof that agentic AI needs more than GPUs. It needs a CPU designed around orchestration, tool use, and fast single-threaded work.