NVIDIA Cosmos Data Factory speeds AI dataset creation up to 60×
NVIDIA's Cosmos Data Factory, using eight RTX PRO 6000 GPUs, offers five blueprints and six skills that cut dataset build time by up to 60× and cost by 45×.
NVIDIA announced its Cosmos Data Factory, an open platform that bundles five cloud‑native blueprints and six composable skills for curating and synthesizing physical AI data.
The system runs on a cloud server equipped with eight NVIDIA RTX PRO 6000 Blackwell Series GPUs. NVIDIA reports "up to 60× faster dataset assembly" and a 45× reduction in cost for autonomous‑vehicle scenario generation compared with traditional pipelines.
The five reference blueprints cover generating driving scenarios, augmenting image attributes, weather‑and‑lighting video augmentation, creating rare event videos, and producing defect images for visual inspection.
The platform is powered by NVIDIA Nemotron and aims to simplify prompt‑to‑data pipelines for autonomous vehicles, robotics and vision AI agents.
NVIDIA invites developers to explore the blueprints and watch a livestream on building training‑ready datasets with only two prompts.
Fontes
- Build Training-Ready Physical AI Datasets with Two Prompts in NVIDIA Cosmos, NVIDIA Developer (official), 18 Nov 2026
Escrito automaticamente a partir das fontes acima, depois verificado pelo nosso código: todo número e nome deve aparecer nelas, e nenhuma frase pode copiar uma fonte. Encontrou um erro? Reporte‑o.
Qual IA você pode usar grátis hoje?
Todo modelo de IA gratuito, classificado por votos humanos cegos, com a memória e velocidade que ele precisa na sua própria GPU.