AI Academic Grant Program provided by NVIDIA

Recently, many PhD students have inquired about accessing high-performance computing resources, such as the NVIDIA DGX system and top-tier GPUs like the H100 and A100, at no cost. While students themselves are not eligible to apply directly, they may participate by collaborating with their academic supervisors, who can submit proposals on their behalf. The NVIDIA AI Academic Grant Program offers these advanced computing resources to support innovative research in artificial intelligence and related fields. The HK PolyU is among the institutions eligible to benefit from this program initiative.

Applicant Eligibility

• Must be a full-time academic staff at an accredited institution
• Postdoctoral researcher and PhD Students must apply through their academic supervisor

Available Resources
Successful applicants may receive access to:
• NVIDIA DGX H100 systems
• A100 and RTX Pro 6000 GPU cards
• Jetson Developer kits, including AGX Orin Nano, AGX Thor, and Drive AGX Thor
• Cloud-based GPU compute instances (Up to 32,000 A100 hours)

Program Topic	Generative AI Training and Model Development (GenAI-TMD)	Generative AI Alignment and Inference (GenAI-AI)	Robotics and Edge AI (REA)
Focus Area	Techniques for training, scaling, and customizing generative AI; foundation models for science; novel architectures	Multi-model systems; techniques for customizing, enhancing, and operating generative AI	Robotics and autonomous vehicles; 5G/6G; smart spaces; federated learning
Resource Award	Up to 32,000 A100 80GB hours or equivalent cloud compute (max. of 8 concurrent GPUs)	Up to 2x NVIDIA DGX Spark System* Up to 32,000 A100 80GB hours or equivalent cloud compute (max. of 8 concurrent GPUs)	Up to 2x NVIDIA Jetson developer kits* Up to 4x NVIDIA RTX PRO 6000 GPUs* Up to 32,000 A100 80GB hours or equivalent cloud compute (max. of 8 concurrent GPUs)
Download	GenAI-TMD Brochure	GenAI-AI Brochure	REA Brochure

* Physical hardware will be shipped to the PolyU

Proposals should adhere the official Proposal Template and should not exceed three pages, not including appendices.

Submission is structured around fixed quarterly periods:

Q1: 1 January – 31 March
Q2: 1 April – 30 June
Q3: 1 July – 30 September
Q4: 1 October – 31 December

Each applicant may submit one proposal per quarter, with a maximum of four proposals annually. However, only one award per applicant per year is permitted. Award decisions are typically announced within 3 months following the submission deadline of each quarter. Awarded GPU hours must be used within 6 months, unused hours will be forfeited. Hardware shipping timelines vary depending on the product model.

For full program detail, visit the NVIDIA AI Academic Grant Program Page.

For proposal writing support, please contact Dr. Vincent NG (cstyng@polyu.edu.hk)