NVIDIA
-
Infrastructure & Chips
Meta opens up Prometheus, its gigawatt-scale Ohio supercluster, and says it is scaling toward about 1 million NVIDIA GPUs
Meta published a detailed engineering account of Prometheus on October 19, 2026. The New Albany, Ohio supercluster's first phase, now hundreds of thousands of NVIDIA GPUs across several buildings, trained the Muse models…
-
Infrastructure & Chips
Modal will run AI agent sandboxes on NVIDIA Vera CPUs after its tests showed up to twice the throughput
AI cloud company Modal said on October 19, 2026 that it will add NVIDIA's Arm-based Vera CPU to the fleet behind its Sandboxes product, after in-house benchmarks showed up to double the task throughput of an equivalent x…
-
Industry
Germany signs a declaration of intent with NVIDIA to build open physical-AI models for its industry
Germany's economy and research ministries signed a joint declaration of intent with NVIDIA on October 19, 2026 to develop specialized, open AI models for robots, factories and other physical systems, organized through th…
-
Open Source
NVIDIA and AI4Bharat open a 2.8T-token Indic pretraining corpus and pipeline
On October 16, 2026 NVIDIA and AI4Bharat at IIT Madras announced Nemotron-pretraining-Indic-v1, about 2.8 trillion tokens across 22 Indian languages, plus an open NeMo Data Designer generation recipe releasing soon.
-
New Models
NVIDIA releases Nemotron 3.5 Super VL, an open 120B multimodal model for always-on agents
NVIDIA on October 15, 2026 released open Nemotron 3.5 Super VL, a 120-billion-parameter multimodal model for specialised always-on agents that organisations can post-train and host themselves, with Bolt edge variants lis…
-
Safety & Ethics
NVIDIA pledges $250 million in UK-hosted compute over five years for agentic AI safety research
NVIDIA said it will give about $50 million a year of UK-hosted compute for five years to fund agentic safety work, with part of the envelope going to Oxford's BOLD lab, UCL's SOFAIR and Imperial College London.
-
Infrastructure & Chips
Scaleway plans an AMD MI355X AI supercomputer in Sweden and an NVIDIA B300 GPU Zone in Copenhagen
French cloud provider Scaleway announced at its ai-PULSE event two Nordic AI compute sites: its first supercomputer outside France, built on AMD Instinct MI355X in Sweden, and a Copenhagen cluster of several thousand NVI…
-
New Models
NVIDIA unveils Nemotron 3.5 Bolt 74B, an open 2-bit model built to run AI agents locally on RTX Spark laptops
NVIDIA announced Nemotron 3.5 Bolt 74B-A9B, a fully open mixture-of-experts model compressed to 24 GB for local agents on RTX Spark, due on October 15 alongside agent tooling updates before the laptops launch on October …
-
Infrastructure & Chips
NVIDIA opens Vera Rubin rack designs to OCP and expands MGX AI factory standards
At the OCP Global Summit, NVIDIA contributed Vera Rubin rack and compute tray designs through open MGX, published MGX 1.2 liquid-cooled rack advances, and released open-source DSX AI factory software components with hype…
-
Business
Nvidia in early talks to deepen its stake in Reflection AI or buy the open-weight model startup
The Financial Times and Bloomberg report that Nvidia is discussing a larger investment, a computing deal or an acquisition of Reflection AI, days after the startup unveiled its Beam open-weight model. No terms have been …