About
Conceived by Anduril founder Palmer Luckey, the Al Grand Prix is an open challenge to the boldest engineers from around the globe, in or out of universities.
No human pilots or hardware modifications will be allowed. The competitive edge is gained entirely by optimizing the best code for the race.
The Stakes
$500K
Prize Pool
Win a Job
at Anduril
Operations
& Hardware
Operated By
Drone Champions League
Hardware By
Neros
Competitors will race fully autonomous, identical drones built by Neros Technologies, incorporating DCL’s Al vector module.
The Path To The AI—Grand Prix
The Path To
The AI—Grand Prix
Supported by
Join Us
Do you have what it takes?
The Al Grand Prix is open to university & independent teams worldwide.
FAQ
What is AI Grand Prix?
The AI Grand Prix, founded by Anduril in partnership with the Drone Champions League (DCL), Neros Technologies, and JobsOhio, is a global autonomous drone racing competition challenging the world’s best engineers to prove their autonomy software under real-world flight conditions. Teams and individuals from around the globe are invited to develop AI systems that pilot high-speed racing drones through dynamic, professional-grade courses with zero human control. Individuals or teams of up to 8 will compete for a $500,000 prize pool and a job at Anduril. All drones are identical and supplied by Neros. More competition details to come.
When and where will it take place?
Who can participate and how to register?
Registration is open to individuals, university teams, research organizations and those with a passion for AI programming. No professional credentials or certifications are required.
To get started, register through this page. You’ll be the first to receive updates on next steps.
Is there an age restriction to sign up?
All ages are allowed, but parents/guardians must register for any interested parties below the age of 14. Written parental consent and age verification will be required for all minors.
Does AI Grand Prix cover flights, accommodations, or other expenses?
No. Each participating team is responsible for covering its own expenses related to the AI Grand Prix, including travel, accommodations, and any additional costs incurred. The Top 10 performing teams at the AI Grand Prix Ohio will be guaranteed a cash prize of at least $5,000.
Which drones will be used? Can teams use their own drone?
Competitors will race fully autonomous, identical drones built by Neros Technologies, incorporating DCL’s AI vector module. All drones will be provided by the AI Grand Prix. Technical specifications will be shared at a later stage.
What does “Win a Job at Anduril” mean?
Winning a job implies one role at Anduril for the 2026 competition season. Standard work eligibility requirements apply. Winning participants can select an alternative cash bonus if they do not meet the necessary eligibility requirements, choose not to pursue a job at Anduril, or if another team member from the winning team is hired. Additionally, top university performers that reach the physical qualifier stage in Southern California will be screened in person for potential internship and full-time entry-level roles at Anduril.
Is this a one-time event?
Are teams allowed to apply?
Participants can compete as individuals or as teams of up to 8 people. For teams, the team leader should register at this stage. More information on additional team members to be collected post registration.
Is there a registration fee?
No. There is no registration fee.
What happens after registration?
When will the simulator and interface specs be released?
Interface specifications will be released in the second half of March. The simulator package (including the course) releases in May. Updates, parameters, and previews will also be shared via weekly newsletters and website FAQs.
What sensors will be available in the virtual and physical environments?
Will there be access to coordinates or absolute positions?
What vision hardware is available (camera / LiDAR)?
Do sensor models or drone models vary in the virtual qualifiers?
No. There is one standardized virtual drone for the virtual qualifiers. The interface remains consistent between both rounds; the environment becomes more complex in the second round.
How do Virtual Qualifier Round 1 and Round 2 differ?
Round 1 is intentionally simple, with a small number of gates and a desaturated environment with visually highlighted gates designed to help teams get started. Round 2 is significantly more realistic and visually complex, including a real 3D-scanned environment and a more challenging course.
Will there be wind or external disturbances in the virtual qualifiers?
No. There will be no wind in the virtual qualifiers. Difficulty increases primarily through realism and visual complexity.
Is qualifying head-to-head or time-trial?
Will teams get the map/course ahead of time?
Will Python code run on the physical drone too?
Yes. The intention is that your Python-based approach transfers from the virtual environment to the physical drone, although some adaptation should be expected.
Can flight data be exported from the simulator?
Yes. Teams should have access to flight data logs for analysis and iteration.
How is scoring handled (time, missing gates, penalties)?
Scoring is primarily time-based but runs must successfully pass gates to count. Teams may need to balance speed versus reliability (faster control can increase missed-gate risk).
Does the drone know the track/map? Is SLAM required?
The drone will not “know” the track. Teams must detect gates and navigate using onboard sensing (primarily vision). Gate position details may be provided only at a rough level; flight-path optimization is the team’s responsibility.
What onboard compute will be allowed (CPU/GPU/TOPS)?
What libraries are allowed? Python only or also C/Cython?
Teams will broadly be able to use the libraries they need, with potential restrictions communicated later. Python is the primary interface, and compiled extensions (e.g., C/Cython) are generally expected to be possible.
What is included in the AI compute module?
At a high level, the module is expected to include onboard AI compute with RAM/SSD, an FPV camera feed, IMU sensors (gyros/accelerometers), and connectivity modules such as Wi-Fi/Bluetooth. Using connectivity to pilot the drone would result in disqualification (the flight must be autonomous).
Can teams have sponsors, and can sponsor logos be placed on the drone?
Yes, more on this to be shared later.
What’s the rough timeline (virtual qualifier, physical qualifier, finals)?
The virtual qualification phase runs May – July, with a cutoff for Round 2 targeted around the end of July (approx.). Selected teams will be invited to an in-person physical qualifier in September 2026 in Southern California. The final event – the AI Grand Prix Ohio – is planned for November. More competition details to come.
Is the physical qualifier indoor or outdoor?
Indoor, with consistent lighting for fairness. Expect obstacles and visual “distractions.” The finals will also be indoors and may include additional disturbances due to spectators (e.g. cameras/flash).
