AI Training and Model Development (NVIDIA Academic Grant Program)

AI Training and Model Development (NVIDIA Academic Grant Program) is sponsored by NVIDIA Corporation. This program offers resources for academic researchers working on generative AI training and model development, focusing on foundation models for scientific domains and techniques for training, scaling, and customizing AI models.

Academic Grant Program for Researchers | NVIDIA Accelerating Innovation in Academia Submit your research proposal to the NVIDIA Academic Grant Program​. Advancing Academic Research The NVIDIA Academic Grant Program advances academic research by providing world-class computing access and resources to researchers. › Receive NVIDIA cloud, hardware, and/or software grants for research.

› Get access to NVIDIA models. Grant Application Support › Receive letters of support for grant applications related to your project. › Have the opportunity to present at GTC and network through NVIDIA’s global channels.

› Be considered for NVIDIA marketing support. To qualify for the Academic Grant Program, you must satisfy all of the requirements listed below. Submissions accepted worldwide.

Must be a full-time faculty member at an accredited academic institution that awards research degrees to PhD students. Previous winners must submit results through the program portal to be eligible for another award. Must adhere to the requirements included in the proposal template and relevant call for proposals (CFP).

Tip: check out our example proposal to see what a high-quality submission looks like. Must incorporate pretrained models from ai. nvidia.

com and/or make extensive use of NVIDIA software distributions. Timeline must align with the expected GPU hours dates and/or estimated hardware ship dates as described in FAQ . This is a competitive program.

Not all projects that meet the eligibility requirements will be accepted. NVIDIA is requesting proposals from academic researchers who are working in the following subject areas. Areas of interest include scientific simulation, quantum computing, and physics-informed machine learning.

Resources available : Up to 30,000 NVIDIA H100 80 GB hours or up to eight NVIDIA RTX PRO™ 6000 GPUs Submit Your Simulation and Modeling Project AI Training and Model Development Areas of interest include techniques for training, scaling, and customizing gen AI, and exploring new approaches and architectures for gen AI.

Resources available : Up to 30,000 NVIDIA H100 80 GB hours Submit Your AI Training and Model Development Project AI Inference, Agents, and Systems Software Areas of interest include techniques for customizing, enhancing, and operating gen AI, exploring compound AI systems, and exploring systems software for AI. Resources available : 2 NVIDIA DGX™ Sparks and up to 30,000 NVIDIA H100 80 GB hours.

Submit Your AI Inference, Agents, and Systems Software Project Please make sure you’ve read the relevant call for proposal (CFP) and understand all of the program requirements before submitting your application. Applications that don’t adhere to the proposal template won’t be reviewed. Before you craft your proposal, we highly recommend reviewing the resources below to ensure your submission meets the prerequisites.

All submissions must adhere to the requirements included in the proposal template . Read the entirety of the template carefully before getting started to ensure you understand what information is necessary to be considered. Level Up Your Proposal Webinar Watch this 30-minute webinar to get expert guidance to help you craft a strong submission.

Hosted by program leads and members of our reviewer team, this webinar covers insider tips and best practices, common mistakes, and frequently asked questions. Review the example proposal to see what a high-quality proposal looks like. Use this to guide you as you craft your own proposal, paying special attention to the overall level of detail and formatting.

Previously Funded Projects University of Michigan's VinuesaLab Advances Urban Flow Simulations With NVIDIA PhysicNeMo Rapid urbanization challenges cities to monitor and simulate turbulent wind patterns for air quality, climate resilience, and infrastructure planning.

To meet this need, researchers at the University of Michigan developed Diff-SPORT, which combines generative models with advanced inference to deliver scalable, interpretable urban flow simulations. The framework was built using NVIDIA PhysicsNeMo .

Read About Urban Flow Simulations Boston Children’s Researchers Deploy AI Across Their Hip Clinic to Support Patients and Doctors Hip disorders, prevalent in young people, cause symptoms like stiffness and pain and are hard to diagnose with just 2D imaging. Boston Children's Hospital uses an AI tool, VirtualHip, developed on the NVIDIA DGX™ platform, to create 3D hip models, assess issues, and guide treatment.

Read about 3D generated analyses UC Santa Cruz Maps Coastal Flooding With NVIDIA Accelerated Computing At UC Santa Cruz, Professor Michael Beck and the Center for Coastal Climate Resilience use NVIDIA RTX™ GPUs and CUDA-X™ software to accelerate flood risk modeling and produce detailed visualizations.

Their research helps government agencies and NGOs assess flood risks and demonstrates how nature-based solutions, like coral reefs and mangroves, can help to reduce damage. Read About GPU-Accelerated Visualizations Frequently Asked Questions I don’t see a CFP that aligns with the work I’m doing. Should I still apply?

Projects that don’t align with the CFP selected will be declined. To be notified of future CFP announcements for academics, join the NVIDIA Developer Program . When is the submission deadline, and when will I find out if I’m selected for a grant?

Please refer to the dates below to determine when you can expect a decision on your application. Applications may be reviewed immediately upon submission - incomplete applications will be declined. Table description text placeholder Estimated Hardware Ship Date When are benefits available to successful applicants?

Please refer to the dates below to determine when you can expect to receive access to your cloud compute resources and/or hardware. Note, hardware shipping timelines vary greatly based on product and geography. Table description text placeholder Estimated Hardware Ship Date The ability to request letters of support for grant applications is available immediately.

What is the selection criteria used to evaluate projects? NVIDIA evaluates proposals based on the following factors: Demonstrable understanding of how to use NVIDIA technology, especially software and models Viability of project timeline and plan Compute request justification Additional factors that may contribute to award decisions include: Innovative or novel use case My proposal was declined. Can you tell me why?

Due to the volume of proposals we receive, we can’t provide individual feedback. Proposals may be declined for a variety of reasons, including but not limited to lack of details, failure to adhere to the CFP requirements or proposal template instructions, misalignment with topic area, lack of experience with NVIDIA technology, etc. What are the commitments if I’m selected? Applicants agree to comply with the program agreement .

Awards provided are unrestricted gifts, and awardees are asked to: Acknowledge NVIDIA Corporation in any resulting publications and include mention of the NVIDIA hardware, software, and/or models utilized Provide an update every six months via a follow-up email questionnaire that asks for information about publications and presentations that NVIDIA’s support enabled Submit proposals to present research outcomes (if any) at GTC or other NVIDIA events to promote the project If our research lab has multiple projects, can we apply more than once?

One proposal per PI or research group is allowed per quarter, a maximum of four proposals annually. For example, one submission allowed in the first quarter, January–March. If my proposal was declined, can I apply again?

Only one submission per person per quarter is allowed. Please be sure to carefully review and adhere to the required proposal template to be sure your proposal is considered. I received a grant.

Can I apply again? Each individual applicant is eligible to receive only one award per calendar year using the expected decision date. Previous winners must submit results through the program portal to be eligible for another award.

I received a grant. Where can I go for technical support? For technology- and platform-related questions, please visit our Developer Forums , which are moderated by NVIDIA experts and community members.

For hardware-related information, please visit our customer support page for answers to common questions or to chat with live support. What type of grant options are offered through this program? Each CFP describes the grant options available to applicants.

Does NVIDIA require any intellectual (IP) rights in the project? No. NVIDIA has no intention of owning the IP developed under the NVIDIA Academic Grant Program. The goal of this program is to support innovative research projects.

Still have questions? Reach out to our team. Get the Latest From NVIDIA on Higher Education and Research back.

Not you?

Clear form Organization / University Name Architecture / Engineering / Construction Automotive / Transportation Healthcare & Life Sciences Higher Education / Academia Bonaire, Sint Eustatius and Saba British Indian Ocean Territory Congo, Democratic Republic of the Falkland Islands (Malvinas) French Southern Territories Heard Island and Mcdonald Islands Korea Democratic People's Republic of Lao People's Democratic Republic Micronesia, Federated States of Saint Pierre and Miquelon Saint Vincent and the Grenadines Saint-Martin (French part) South Georgia and the South Sandwich Islands United States Minor Outlying Islands Venture Capital (NVentures) Professional Services for Data Science