Robotics and Edge AI Academic Grant Program

The NVIDIA Graduate Fellowship Program awards up to $60,000 to PhD students whose doctoral research advances accelerated computing and its applications, with the program explicitly encouraging submissions in artificial intelligence, robotics, autonomous vehicles, and embodied and physical AI. Beyond the funding, fellows are paired with NVIDIA researchers and complete a mandatory in-person summer internship at an NVIDIA office, giving recipients direct access to the company's research organization and hardware. Applications are reviewed competitively over roughly eight weeks by a committee of more than 80 senior NVIDIA technical staff led by Chief Scientist Bill Dally, evaluating student quality, research quality, and relevance to NVIDIA's research domains; NVIDIA notes that many qualified candidates go unselected because the number of awards is limited. The application window closes in mid-September each year for fellowships supporting the following academic year — the most recent published cycle closed 3:00 p.m. Pacific on September 15 for the 2026-2027 academic year — so applicants should confirm the current deadline on the program page before preparing materials.

The U.S. Special Operations Command (USSOCOM) Broad Agency Announcement USSOCOM-BAAST-2020, updated in July 2025, is a long-running open BAA accepting white papers on a rolling basis through December 31, 2030. A major 2025 amendment added a dedicated subsection for Advanced Autonomy and Artificial Intelligence, reflecting SOF's strategic priority to become early adopters of AI technologies as articulated in the SOF Renaissance strategic vision. The AI/autonomy subsection emphasizes modular, open integration of cutting-edge solutions and specifically seeks proposals addressing agentic AI and vision language action (VLA) models for sophisticated autonomous behaviors with adaptive learning, neural radiance fields (NeRFs) for 3D scene representation and navigation, generative AI for simulation and data augmentation, advanced automatic target recognition (ATR) algorithms with edge refinement, and machine learning operations (MLOps) for data management and model monitoring. Solutions must feature well-defined interfaces adhering to open standards to ensure interoperability with existing military architectures. The BAA also covers collaborative autonomy for multiple autonomous or remote-controlled systems sharing data in real-time, equipping small autonomous platforms including drones, ground robots, and uncrewed surface and underwater vessels with ISR and edge computing capabilities. SOFWERX, a military capabilities nonprofit partnering with SOCOM, hosts regular assessment events for selected participants to pitch and demonstrate solutions.