1,000+ Opportunities
Find the right grant
Search federal, foundation, and corporate grants with AI — or browse by agency, topic, and state.
NVIDIA's Academic Hardware Grant Program donates GPU hardware to qualifying academic researchers, with a specific track focused on AI Safety and Trustworthy AI research (including interpretability, alignment, evaluation, red-teaming, and bias/fairness). Awardees receive workstation- or datacenter-class NVIDIA GPUs delivered to their institutions for use in research projects.
Applications are accepted on a rolling basis through the NVIDIA Higher Education and Research portal. Projects must produce a publishable research output or open-source artifact and acknowledge NVIDIA's hardware contribution. The AI Safety track is specifically highlighted by NVIDIA's collaboration with the AI Safety Directory and prioritizes researchers working on interpretability, evaluations, robustness, and trustworthy ML.
Strong fit for graduate students, postdocs, and early-career faculty doing AI safety, alignment, or interpretability research at universities without existing GPU access, as well as nonprofit AI safety research organizations.
Get a weekly digest of new grants like this
A free weekly digest of new foundation and federal funding opportunities as they're added to Granted. Unsubscribe anytime.
Or search similar grants →According to the current listing, eligibility includes: Faculty, postdoctoral researchers, and graduate students at accredited universities and research institutions worldwide. AI Safety Directory track requires AI-safety-focused project scope. Recipients must publish or open-source research outputs and acknowledge NVIDIA's hardware contribution. Confirm the full requirements in the official notice before applying.
The current listing shows donated NVIDIA hardware (workstation-class GPUs such as RTX 6000 Ada, RTX A6000, or datacenter-class cards including H100 PCIe and L40S for higher-impact projects). Single hardware grants typically valued at $5,000 to $30,000 per approved project; larger awards possible for high-impact AI safety and trustworthy AI research. Verify award ceilings, matching requirements, and allowable costs in the official notice.
NVIDIA Academic Hardware Grant Program for AI Safety and Trustworthy AI Research with Donated GPU Workstations and Datacenter Cards is funded by NVIDIA. Verify program details on the funder's official page before applying.
Start from the official opportunity page linked in this listing — it carries the sponsor's submission instructions.
Amazon Research Awards gives unrestricted gifts to academic researchers working on topics Amazon cares about. The Fall 2026 call covers nine topics: Agentic AI; AI for Information Security; Amazon Security (using AI on hard security problems); Automated Reasoning for systems assurance by mathematical proof; AWS Cryptography; Build on Trainium (agentic recursive improvement of kernels and frameworks); Multilingual and Contextual AI Evaluation; On-Device Multimodal Perception; and Synthetic Data Generation for Multimodal Perception. Awards go to the PI's academic institution or organization. Recipients are encouraged to publish their results and open-source related code, and each is assigned an Amazon research contact. Amazon extended the Fall 2026 submission deadline to November 11, 2026. Each topic has its own CFP page with amounts and eligibility details.
Google's Gemma 4 Developer Agent Competition on Kaggle, launched 23 September 2026, challenges entrants to post-train Gemma 4 open models into autonomous coding agents. The agents must run offline on consumer hardware, navigate large codebases and draft fixes for real software issues. Fine-tuning, including LoRA, and reinforcement learning are encouraged, and one submission is allowed per day. The $110,000 prize pool covers a main leaderboard track paying $37,000, $18,000 and $10,000 for the top three. A separate paper track carries $35,000 with its own 12 November 2026 deadline. A $10,000 bonus goes to the best agent built on the smaller Gemma E2B or E4B models. Winning solutions are released under Apache 2.0, and top entries may be featured at a Google-hosted event such as a NeurIPS expo workshop. The entry deadline is 25 November 2026, and final submissions are due 2 December 2026 at 11:59 PM UTC.
The Enveda CASMI 2026 competition on Kaggle is the 2026 Critical Assessment of Small Molecule Identification challenge. It launched on 14 September 2026 and is sponsored by Enveda Therapeutics. Entrants build machine-learning models that predict 2D chemical structures, as SMILES strings, from LC-MS/MS tandem mass spectra of natural products in complex biological extracts. Models must handle both known and novel molecules, going beyond spectral library matching. The stated applications are drug discovery and finding disease biomarkers. The $50,000 prize pool pays $16,000, $12,000, $9,000, $7,000 and $6,000 to the top five teams. Winning solutions are released under the MIT licence, and the data is limited to competition and non-commercial academic use under CC BY-NC 4.0. The entry deadline is 7 December 2026, and final submissions are due 14 December 2026 at 11:59 PM UTC.
On October 8, 2026, eleven companies pledged $2.4 billion in AI tools and compute credits to the Genesis Mission Consortium. Nvidia's $1 billion runs over five years, no allocation rules exist yet, and the credits flow to 15+ federal agencies, not to principal investigators. What that means for proposals.
Read articleNVIDIA committed $250 million over five years to UK agentic safety and security research, delivered as compute access through UK-based Cloud Partners. No open call, no cash, and the commercial terms are not yet agreed. A breakdown of compute-as-grant and how to budget for it.
Read articleNSF 26-513 makes roughly $100 million available for up to 10 State and Regional AI Infrastructure Hubs at $4M to $12M each over five years. One award per state or multi-state region. One proposal per organization. And NSF is not buying you GPUs — it funds the coordination, the workforce and the faculty training, while the compute has to come from partners you have to already have.
Read article