Nebius x NVIDIA Global AI Hackathon

Overview

The Nebius x NVIDIA Global AI Hackathon is an online event designed for AI builders, ML engineers, and founders to develop practical, high-performance AI solutions. Participants will utilize NVIDIA open source models served on Nebius through Nebius Token Factory, leveraging high-performance GPU infrastructure. The hackathon encourages experimentation with new ideas and pushing the boundaries of open source AI. Entrants are tasked with building a working AI system that runs on either Nebius Token Factory or Nebius AI Cloud and incorporates at least one NVIDIA open source model. The event offers opportunities for technical onboarding, workshops, team formation, and feedback. Projects must fit into one of four tracks: Coding and Agentic Engineering, Best Apps and Agents, Personal AI, or Physical AI. Submissions require a working project, a chosen category, a detailed project description, a working demo (or equivalent for Physical AI), a demo video, and a public code repository with an open source license. Feedback on Nebius and NVIDIA tools is also requested. With over $50,000 in cash and other prizes, the hackathon seeks to recognize innovative and well-implemented AI solutions. Judging criteria include technological implementation, design, potential impact, and the quality of the idea. Participants also have the chance to win city-specific awards if they attended IRL Builders & Brews events.

What to build

Participants must build a working AI system that runs on either Nebius Token Factory or Nebius AI Cloud and uses at least one NVIDIA open source model. Projects should fit into one of the following tracks: - Coding and Agentic Engineering Track: Build coding agents and developer tools, specifically agents that write, run, and test code in Token Factory Sandboxes. - Best Apps and Agents Track: Create any app or agent that is practical for users, such as productivity tools, copilots, or automated workflows. Power these with Nemotron models on Nebius through Token Factory, using Nemotron 3 Ultra for reasoning and Nano/Super for everyday calls. Deployment with Nebius Serverless Endpoints or use of Nebius Serverless Jobs is encouraged but not required. - Personal AI Track: Develop an always-on, private assistant with persistent memory, reusable skills, access to chosen tools and information, and the ability to execute tasks across daily workflows. Use at least one NVIDIA open source model and tools like NVIDIA NemoClaw, OpenShell, Hermes Agent, and Nebius Serverless. - Physical AI Track: Build embodied and edge agents that interact with the real world, including robotics, IoT, and on-device intelligence. These should be powered by Nemotron, GROOT, Cosmos, and Sonic models and coordinated through an agent runtime. Nebius Serverless Jobs can be used for simulations, data generation, policy evaluation, or sensor data processing, and Nebius Serverless Endpoints for real-time inference.

Prizes

Key dates

• Deadline2026-10-30T10:00:00-07:00

Eligibility

Participants must be above the legal age of majority in their country of residence. All countries/territories are eligible, excluding standard exceptions.

How to enter

1. Join the hackathon through the Devpost platform.
2. Build a working project that meets the challenge requirements: runs on Nebius Token Factory or Nebius AI Cloud and uses at least one NVIDIA open source model.
3. Select the track that best fits your project.
4. Prepare a project description detailing what was created, why, and how it works.
5. Provide a URL to a working demo, hosted application, or test build (not required for Physical AI submissions).
6. Upload a public YouTube video (3 minutes or shorter) demonstrating the project, with audio explaining the use of Nebius Token Factory and NVIDIA Nemotron or other NVIDIA open source models. Physical AI submissions must include at least 1 minute of physical hardware/robot operation footage (or key application modules if no hardware).
7. Submit a publicly accessible code repository URL (GitHub, GitLab, or Bitbucket) with an open source license (e.g., Apache 2.0, MIT, MPL 2.0) visible at the top, including a README with setup and running instructions. Highlight the use of NVIDIA Nemotron, Token Factory, and other Nebius tools/services.
8. Provide feedback on Nebius Token Factory, AI Cloud, and any NVIDIA tools, models, or technologies used.
9. If the project existed before the Submission Period, include a written explanation of significant updates made during the Submission Period.
10. If you attended an IRL Builders & Brews event, choose the city you attended for a chance to win the City Winner Award.

Submission requirements

• All submissions must run on either Nebius Token Factory or Nebius AI Cloud.
• All submissions must use at least one NVIDIA open source model.
• Project must fit into one of the specified tracks: Coding and Agentic Engineering, Best Apps and Agents, Personal AI, or Physical AI.
• A working project (app, agent, or new creation).
• A clear category selection for the project.
• A project description explaining creation, purpose, and functionality.
• A working demo (URL to hosted app or test build); not required for Physical AI.
• A demo video (3 minutes or less, public YouTube) showing the project in action, with audio explaining usage of Nebius Token Factory and NVIDIA models. Physical AI videos must include at least 1 minute of physical hardware/robot operation or key application modules in action.
• A public code repository URL (GitHub, GitLab, or Bitbucket) with an open source license (e.g., Apache 2.0, MIT, MPL 2.0) and a README including setup/running instructions. Highlight use of NVIDIA Nemotron, Token Factory, and other Nebius tools/services.
• Feedback on Nebius Token Factory, AI Cloud, and any NVIDIA tools, models, or technologies used.
• If a pre-existing project, provide a written explanation of updates made during the submission period.
• For Physical AI projects without physical hardware, the demo video must show key application modules in action for at least 1 minute.

Judging criteria

• Technological Implementation: How well is the project built, and how effectively does it use Nebius Token Factory or AI Cloud model(s), and NVIDIA Nemotron as part of the solution?
• Design: Does the project deliver a complete, coherent product experience, not just a technical proof of concept?
• Potential Impact: Does the project make a credible, specific case for solving a real problem for a real audience and does the solution actually address it based on what's demonstrated?
• Quality of the Idea: Is this a creative, non-obvious use of Nebius Token Factory or AI Cloud model(s), and NVIDIA Nemotron, and does the team show genuine understanding of the problem space?

Sponsors & partners