Together AI partners with IBM and NVIDIA on dedicated B300 cluster for open-source inference
Together AI is scaling open-source model inference in collaboration with IBM and NVIDIA on a dedicated B300 cluster hosted on IBM Cloud.
#nvidia
View all AI newsNVIDIA researchers introduce PivotOPD to train AI agents on error recovery
NVIDIA researchers developed PivotOPD, a method designed to teach AI agents how to avoid early task mistakes and recover when errors occur. The training approach uses a teacher model to show agents better actions and guide them back on track over subsequent steps.
Jaguar Type 01 uses NVIDIA Hyperion and Halos
NVIDIA said the Jaguar Type 01 uses its Hyperion computer and sensor platform to make real-time decisions from what the car sees and senses, paired with the Halos safety system. The company said the software passed 150,000 tests over tens of thousands of hours before road use and will receive over-the-air updates.
NVIDIA applies SynthID watermarking to Cosmos models on build.nvidia.com
NVIDIA confirmed that content generated by NVIDIA Cosmos on build.nvidia.com includes SynthID watermarks, allowing users to upload and verify the media using Google DeepMind's publicly released SynthID Detector.
NVIDIA paper details VERA framework for co-evolving agent harnesses and models
An NVIDIA research paper presents VERA, a system that converts benchmark trajectories into over 9,000 restartable sandboxes with rubric scoring while co-evolving agent harnesses alongside model weights. The team open-sourced the environment corpus, reporting that a 27B agent scored 71.6 on AutoCoWorkBench to top Claude Opus 4.8.
SpaceX reportedly seeks $40 billion to buy Nvidia AI chips
According to the Financial Times, SpaceX is reportedly seeking $40 billion to fund purchases of Nvidia AI chips, with Apollo leading the financing.
CoreWeave introduces RL Rollouts for live model weight updates in reinforcement learning
CoreWeave has introduced RL Rollouts, an infrastructure feature that loads new model weights directly into live inference deployments during reinforcement learning without disrupting active requests. NVIDIA and You.com used the feature to post-train Nemotron 3.5 Lightning for web search.
Splunk expands work with awscloud and NVIDIA to bring local compute to Splunk AI
At splunkconf26, Splunk announced expanded collaboration with awscloud and NVIDIA to deliver more local compute power for Splunk AI, enabling customers to run governed AI on sensitive machine data in self-managed environments.
NVIDIA AI says ATT, SoftBank, and Indosat are adopting open models for telecom agents
ATT, SoftBank, and Indosat are deploying open models for autonomous network operations and local language adaptation using a full-stack framework provided by NVIDIA AI.
NVIDIA releases Nemotron-Labs-3-Competitive-Coding on Hugging Face
NVIDIA AI has released Nemotron-Labs-3-Competitive-Coding on Hugging Face. The competitive-programming specialist model is based on Nemotron-3-Ultra and powers the team's performance on the IOI 2026 problem set.
CoreWeave makes NVIDIA Vera Rubin NVL72 available as Cognition runs Devin in production
CoreWeave announced that NVIDIA Vera Rubin NVL72 is available on its platform. Cognition is running Devin in production on the hardware, achieving up to 4.8x the total token throughput of GB200 NVL72.
NVIDIA Mid-Harness raises TerminalBench-Lite Pass@1 from 50.0% to 68.0%
An NVIDIA paper on test-time compute for terminal agents says a GPT-5.6 Sol verifier choosing among eight sampled shell commands lifts TerminalBench-Lite Pass@1 from 50.0% to 68.0%. Extra samples add almost nothing when the verifier is weak.
Ramine Roane says AMD made 502 core vLLM commits in 90 days
Ramine Roane says AMD is currently the leading company contributing code to vLLM, with 502 core commits in 90 days, 13% of all organization contributions, and nearly four times Nvidia. He also credited Red Hat, IBM, Embedded LLM, and Inferact.
ClawHub runs two open-source scanners and a judge harness on agent-skill uploads
Patrick Erichsen says ClawHub runs two open-source security scanners on every upload and a judge harness over both results. He says the team worked with Tencent Hunyuan and NVIDIA AI on ClawScan, and he believes this launch makes ClawHub the most secure place to find and install agent skills.
NVIDIA OpenShell can govern agents with OpenClaw Enterprise
NVIDIA AI says organizations can use NVIDIA OpenShell as an open source option for governing agents with OpenClaw Enterprise.
Jure Leskovec shares NVIDIA Kumo Tabular as an open tabular model family
Jure Leskovec says NVIDIA Kumo Tabular is a new family of foundation models for tabular data, released with open weights, open-source software, and a permissive license for commercial use.
NVIDIA researchers release Physis-Lang for physics reasoning in video captions
Physis-Lang is an open, self-evolving framework that adds physics reasoning to video captions. NVIDIA AI says adding that reasoning to prompts alone raised NVIDIA Cosmos 3’s PhyGenBench score by 5.62 points without retraining.
clem says NVIDIA is acquiring Hugging Face
clem says NVIDIA’s acquisition lets Hugging Face hire people it could not as a small startup and give them a decade to make open-source AI win. He invited those people to message him directly.
Perplexity’s Secure Intelligence Institute works with university researchers and NVIDIA
Perplexity says its Secure Intelligence Institute works with researchers at Stanford, CMU, Duke, Columbia, Ohio State, and UVA, and is also working with NVIDIA and the Open Secure AI Alliance to share tools and findings.
Salesforce and NVIDIA build Slack Admin Bridge for OpenShell
Salesforce and NVIDIA have built Slack Admin Bridge for OpenShell. OpenShell is the secure runtime, and the bridge puts incremental human oversight of AI agents into Slack.