News from manufacturers

Open Models, Closed Environments: Palantir Brings Secure AI to US Agencies With NVIDIA Nemotron

Palantir has introduced a new intelligent engine that uses NVIDIA Nemotron open models to serve U.S. government agencies. The engine allows agencies to run customized models on their own infrastructure, train on their own data, and retain full ownership of the resulting models. This deployment is supported by Palantir's Sovereign AI Operating System and NVIDIA accelerated computing.

From host node to heterogeneous rack: Rethinking the AI CPU

Arm describes a shift in AI infrastructure from accelerator-centric design to rack-scale heterogeneous systems optimized for agentic AI workflows. CPUs play multiple specialized roles in these racks, including prefill and decode host functions, as inference becomes more complex and memory-intensive.

“The CPU is not a commodity, but it is not a single prize either.” — Arm

The Ultimate Summer Sale Pairing: Steam Sale Meets GeForce NOW Discounts

NVIDIA announces a promotion pairing the Steam Summer Sale with GeForce NOW membership discounts, offering $70 off a 12-month Ultimate membership and $35 off a 12-month Performance membership. Six new games join the GeForce NOW library, including Dark Scrolls and The Adventures of Elliot: The Millennium Tales.

“Get $70 off a 12-month Ultimate membership or $35 off a 12-month Performance membership” — NVIDIA

Related: NVIDIA GeForce RTX 4080 , NVIDIA GeForce RTX 4080 Super , NVIDIA GeForce RTX 4080 Mobile/Laptop , NVIDIA GeForce RTX 5080

NVIDIA and AWS Collaborate to Bring AI to Production at Scale

NVIDIA and AWS are collaborating to bring AI to production scale. New EC2 G7 instances with NVIDIA RTX PRO 4500 Blackwell GPUs offer up to 4.6x AI inference performance over G6. NVIDIA cuVS enables GPU-accelerated vector search as default in Amazon OpenSearch Serverless, and AWS achieved NVIDIA Exemplar Cloud status for GB300 training workloads.

“Compared with G6 instances, G7 delivers up to 4.6x AI inference performance, up to 2.1x graphics performance and significantly faster GPU-accelerated data analytics on Amazon EMR using the NVIDIA cuDF library for Apache Spark workloads.” — NVIDIA

How Businesses Are Building Specialized AI They Can Trust

NVIDIA introduced the Agent Toolkit, an open, modular foundation for building specialized AI agents. The toolkit includes Nemotron open models, NemoClaw blueprints, and the OpenShell runtime to enable customization, safety, and scalability. It is already being used in life sciences, healthcare, cybersecurity, and other industries.

“CrowdStrike is running specialized security agents that triage alerts with 98.5% accuracy.” — NVIDIA

NVIDIA Powers Over 400 of the World’s 500 Fastest Supercomputers

NVIDIA technology powers over 400 of the world's 500 fastest supercomputers, representing 81% of the TOP500 list. The company's GPUs accelerate a record 238 systems, and its networking connects 376 systems. NVIDIA Grace CPU adoption has grown to 26 systems, and the top eight Green500 systems use NVIDIA GPUs.

“NVIDIA technology runs 81% of the TOP500 and 90% of the systems new to the list.” — NVIDIA

Related: NVIDIA B200 GPU accelerator

NVIDIA Brings Trusted, 24/7 AI Agents to Telecom Operations

NVIDIA announced AI agent technologies for telecom operators at TM Forum's DTW Ignite 2026. The company highlighted synthetic data generation and secure agent runtimes to enable autonomous network operations. Partners including SoftBank, AdaptKey, Amdocs, and NTT DATA are piloting these solutions for tasks like self-healing networks and proactive customer care.

“Automation is no longer the finish line — it's the launchpad to autonomy.” — NVIDIA

At ISC, JUPITER Shows What Exascale Science Looks Like

JUPITER, Europe's first exascale supercomputer at Forschungszentrum Jülich, uses NVIDIA Grace Hopper Superchips and Quantum-X800 InfiniBand networking. Four projects running on JUPITER demonstrate exascale capabilities: mapping the human brain at cellular scale, simulating Earth's climate at 1-kilometer resolution, building AI for wireless networks, and simulating a 50-qubit quantum computer. The CytoNet foundation model for brain microarchitecture trained on JUPITER in under five days using 6.5

“With JUPITER, Europe doesn't just join the exascale era — it leads it, across the widest range of science and AI of any system worldwide.” — NVIDIA

NAIRR Science Program Reshapes Scientific Research, Powered by NVIDIA AI Infrastructure

The NAIRR pilot program, supported by NVIDIA, has enabled over 700 research projects in areas like protein prediction and disease management. NVIDIA provides cloud-based DGX nodes and technical support, helping researchers accelerate workflows. Notable projects include Polymathic AI's fluid simulation dataset, University of Michigan's MIST molecular models for energy storage, and Boston University's BEACON pipeline for infectious disease detection.

From Materials Simulation to Experimental Astronomy, New NVIDIA AI Software Unlocks Scientific Discoveries

NVIDIA announced new software at the ISC conference, including the DAQIRI library, ALCHEMI NIM microservices, and cuPhoton reference code, to accelerate AI for science. These tools, part of NVIDIA CUDA-X, enable GPU-accelerated pipelines for tasks such as materials simulation and astronomical data analysis. cuPhoton demonstrated significant speedups in processing FITS data from the Rubin Observatory's LSST camera.

“cuPhoton accelerated loading and reading of FITS images collected by the Rubin Observatory’s Legacy Survey of Space and Time (LSST) by 14,900x.” — NVIDIA