AI compute infrastructure spanning accelerated hardware, systems and deployed platforms for training and inference.
Research candidate, not an investment rating. Identity and classification reflect a bounded evidence snapshot observed Jul 26, 2026.
Period ended Apr 26, 2026 · filed May 20, 2026
Period ended Jan 25, 2026 · filed Feb 25, 2026
Values are issuer-filed US-GAAP XBRL facts retrieved from SEC EDGAR. Non-GAAP measures and market prices are intentionally excluded from this canonical snapshot.
NVIDIA spans accelerated compute, networking, systems, inference software and open model families. Its AI exposure is operational and material rather than inferred from branding.
Blackwell and Rubin GPU systems for training and inference.
NVLink, Spectrum-X, ConnectX and BlueField infrastructure.
CUDA, Dynamo, TensorRT-LLM and enterprise AI tooling.
Nemotron, Cosmos, Alpamayo, Isaac GR00T and BioNeMo families.
GPU architecture and rack-scale systems designed for generative, reasoning and agentic AI workloads.
Official Blackwell architecture overviewNext-generation platform combining Vera CPU, Rubin GPU, networking, storage and rack-scale infrastructure.
Vera Rubin production announcementOpen-source inference operating software for orchestrating generative and agentic AI at scale.
Dynamo 1.0 production announcementModel families across agentic, physical, automotive, robotics and biomedical AI.
Open model families announcementNVIDIA and Microsoft presented a new Windows PC class centered on local personal AI agents.
NVIDIA reported production scaling across server makers and the wider AI infrastructure supply chain.
Issuer-reported Data Center revenue was $75.2B and NVIDIA announced a new reporting framework.
NVIDIA added model families for agentic, physical, automotive, robotics and healthcare AI.
The annual results reported 65% revenue growth and $193.7B in full-year Data Center revenue.
The financial projection can be rebuilt from the recorded SEC endpoints and filing accessions without an LLM.
AI roles are presented as source-anchored research interpretations. The document stream is regulator-primary, but a filing’s existence does not prove the claims contained inside it.
Role evidence, not a proprietary investment score
The issuer supplies compute, cloud or accelerated infrastructure used for AI workloads.
NVIDIA identifies itself as Nasdaq: NVDA and reports Data Center growth while attributing demand to accelerated computing and AI infrastructure.
NVIDIA identifies itself as Nasdaq: NVDA and reports Data Center growth while attributing demand to accelerated computing and AI infrastructure.
Its fiscal-year results describe deployed platforms for AI training, inference, storage and AI PCs.
These sources support the issuer/listing and AI-exposure assessment. First-party evidence proves attribution—not independent truth.