NVIDIA Academic Grant Program
Supports in-kind GPU compute and hardware awards for university researchers worldwide.
The NVIDIA Academic Grant Program awards GPU cloud hours and/or physical GPU hardware (in-kind, no cash) to faculty-led research projects at PhD-granting institutions. Awards are unrestricted gifts; NVIDIA retains no IP. Three active 2026 call-for-proposal tracks cover Simulation and Modeling, AI Training and Model Development, and AI Inference / Agents / Systems Software — each offering up to 30,000 H100 hours and/or physical GPU hardware. Applications run quarterly: proposals submitted in a given quarter receive decisions the following quarter and GPU access or hardware shipment shortly after. One proposal per PI or research group per quarter; maximum four per year. Prior awardees must submit results before applying again. The selection portal is hosted on Salesforce (academicgrants.nvidia.com) and requires a login to submit. This program replaced the legacy Applied Research Accelerator Program (which had cash tiers); awards are now exclusively in-kind compute resources.
Each grant below is a distinct funding opportunity with its own eligibility, scope, and deliverables.