News from manufacturers

How Nations Are Deploying AI for Strategic Priorities

Nations are investing in AI infrastructure, including AI factories, to develop models tailored to local needs. These efforts span physical and data infrastructure, with a focus on foundation models trained on local datasets. Key components of national AI strategies include workforce development, ecosystem building, and public-private partnerships.

“The AI factory will become the bedrock of modern economies across the world” — NVIDIA

Joyride Through July With 12 Games Coming to GeForce NOW

NVIDIA announced 12 new games coming to GeForce NOW in July, including Monopoly: Star Wars Heroes vs. Villains. The GeForce NOW Summer Sale offers discounts on Performance and Ultimate memberships. The service allows cloud gaming across various devices.

NVIDIA Unlocks AI Compute at Scale, Inviting Partners to Power the AI Infrastructure Buildout

NVIDIA introduces a new business model for AI clouds to procure its infrastructure through revenue-sharing and credit-support, enabling faster access to accelerated computing for AI startups and enterprises. Sharon AI and Firmus are among the first partners, with Sharon AI deploying up to 40,000 NVIDIA Grace Blackwell GB300 GPUs and Firmus building a 360-megawatt AI factory campus in Indonesia. The initiative aims to meet growing demand for token-scale AI inference and training capacity.

“Sharon AI is deploying up to 40,000 NVIDIA Grace Blackwell GB300 GPUs.” — NVIDIA

NVIDIA and Partners Build in America, for America

NVIDIA and its partners are manufacturing products within the United States to serve the American market. The initiative focuses on domestic production and supply chain resilience.

Top Arm-based innovations advancing AI, cloud, edge and mobile computing in June 2026

Arm's June 2026 newsletter highlights innovations including Microsoft Azure Cobalt 200 VMs, low-light AI vision, and Unreal Engine graphics optimizations for mobile. The Cobalt 200 CPU offers performance gains for cloud-native and agentic AI workloads. Arm researchers developed a generative AI solution for low-light image enhancement, and Unreal Engine MegaLights and Nanite optimizations improve mobile graphics.

“Cobalt 200 delivers: 16% faster agentic sandbox creation 42% higher agentic sandbox memory bandwidth 50% better agentic sandbox tool…” — Arm

Intel to Report Second-Quarter 2026 Financial Results

Intel will report its second-quarter 2026 financial results on July 23, 2026, after market close. The company will hold an earnings conference call at 2 p.m. PDT that day to discuss the results.

NVIDIA BioNeMo Agent Toolkit Brings Accelerated AI to Life Sciences Researchers in Claude Science

NVIDIA's BioNeMo Agent Toolkit integrates with Anthropic's Claude Science, an AI workbench for scientific research. The toolkit provides accelerated models and microservices as callable skills, enabling natural language-driven workflows in drug discovery, genomics, and other life sciences. It is used by 18 of the top 20 pharmaceutical companies.

“18 of the top 20 pharmaceutical companies use NVIDIA BioNeMo” — NVIDIA

Apple Creator Studio gets smarter, faster, and more connected

Apple announced updates to its Creator Studio suite of creative applications, enhancing performance and connectivity.

How NVIDIA’s Inference Software Stack Powers the Lowest Token Cost

NVIDIA's inference software stack, codesigned with its hardware and supported by an open-source ecosystem, reduces token costs by optimizing performance across GPU, CPU, networking, and system layers. On the Blackwell platform, the stack achieved up to 5x cost reduction on DeepSeek V4 in one month. Companies like Baseten, Cognition, Deep Infra, and Together AI use NVIDIA's tools to improve throughput and manage inference workloads.

“the software stack has already reduced token costs by up to 5x on the DeepSeek V4 model in just one month” — NVIDIA

How Jaiveer Singh Is Helping Robots — and Developers — Move Faster

NVIDIA Isaac ROS is a robotics software platform built on ROS 2, providing CUDA-accelerated libraries and AI models for autonomous mobile robots, manipulation systems, and humanoids. The platform is modular, allowing developers to combine packages with existing ROS code. It runs on workstations, NVIDIA DGX Spark, and Jetson edge systems.

“Compared with the original Isaac SDK, Isaac ROS is completely modular. We ship the software like a bunch of LEGO bricks — you get to assemble them however you want, and you can easily combine our packages with existing ROS code written by y” — NVIDIA