Challenge Center
The challenge center contains all necessary information to compete in AI Cup. Please visit this page for any updates about the challenge.
AI Cup Overview
The AI Cup 2026 is a nationwide, implementation-focused AI competition organised by Epoch and enabled by AIC4NL. The competition is designed to connect students, researchers, and industry partners around societal and scientific challenges where artificial intelligence can play a meaningful role. Unlike many AI competitions that solely evaluate model accuracy, the AI Cup also looks at the complete picture and evaluates how such AI technology can be implemented.
The 2026 edition centres on the challenge of developing an AI technology capable of identifying bird species using radar data to improve bird conservation around wind farms. Participants work in multidisciplinary teams over five weeks, competing in two complementary tracks: performance and implementation. The performance track aims to create the heart of the technology, proving that an AI model can be used for the proposed task. The implementation track is intended to identify how this AI technology can be effectively integrated into a broader operational system that addresses the underlying problem posed by the challenge provider.
Five finalist teams are nominated to present their work at the 2026 AIC4NL Congress. There, they present both their technical results and their implementation proposals to a professional audience and jury panel, competing for a total prize pool of €7,000. Through this structure, the AI Cup aims to:
- Bridge emerging AI talent and industry.
- Instigate participants to think beyond model accuracy and engage with the practical realities of deploying AI.
- Foster the development of a new AI technology for a sustainable and social cause.
- Become an AI talent benchmark for students in the Netherlands.
Tracks
Team Epoch, in alignment with the Delft AI Initiative, believes that truly strong AI solutions are not only high-performing, but also implementable and responsible. AI Cup 2026 embodies these principles through two complementary tracks: Performance and Implementation.
Future AI engineers must look beyond optimizing a single loss function. Industry increasingly values professionals who understand the limitations of their models and can integrate AI systems effectively within broader operational contexts. Strong results on a single metric are rarely enough to move a technology from experimentation to real-world deployment. For this reason, we encourage teams to be multidisciplinary, combining expertise in areas such as machine learning and data science, domain knowledge relevant to the challenge, software engineering, and system design.
Performance Track | 60% of overall score
This track is focused on the technology, maximizing the performance of the AI model for the core task provided by TNO. Assessment is based exclusively on quantitative metrics computed on the Kaggle leaderboard. This track rewards teams that demonstrate strong technical proficiency in model design, training, and evaluation, independent of deployment or application context.
Important note: You must use one of the blue performance track buttons from this page to get access to the Kaggle competition for the first time.
Implementation Track | 40% of overall score
This track is intended to identify how the AI model can be responsibly and effectively integrated into a broader operational system that addresses the underlying problem posed by TNO. The deliverable will be an implementation proposal that covers:
- System description ( Input / Model /Output)
- Responsible AI analysis
- Data collection & labeling methods
- Mitigation & conservation methods.
The implementation track rewards teams that demonstrate a strong ability to translate a high-performing model into a feasible, responsible, and impactful application.
Nomination Score
An overall score will be computed for all participants via a comparative system based on their ranking in the performance track and in the implementation track. It is a rank normalized score in which 60% is attributed to the performance track and 40% to the implementation track. This score will be used to nominate the top 5 teams
Structure & Prizes
Kick-off
(13/02/2026)
The kick-off marks the start of a five-week journey, in which you will develop end-to-end AI models, design practical implementation strategies, and compete for a share of the €7,000 prize pool. Participants who secure a ticket will receive first-hand insights from TNO and research experts involved in the challenge context. The event also offers team-matching opportunities and introductory workshops for beginners.
If you were unable to attend, don’t worry. All essential information required to start the challenge is available online. No challenge-relevant information is exclusive to kick-off attendees.
AI Futures Day: Career Fair & Nomination
(28/03/2026)
After 5 weeks of competing, 5 teams will be nominated for the next stage as per the following criteria:
- Team with the highest Performance Score
- Team with the highest Implementation Score
- 3* teams with highest Nomination Scores
This is to allow for recognition of teams that demonstrate strong technical proficiency even without application context, and vice versa. The evaluation of implementation proposals at this stage will be done by Team Epoch and TNO.
Attention: Presence of at least one team member is mandatory at the nomination day to be nominated and win a prize.
*If there are overlaps between the top rankers of the overall solution and 1st position for Performance or Implementation, the next ranking overall solution team(s) will be nominated to ensure a total of 5 teams at the next stage.
Award day at the Dutch AI Congress
(14/04/2026)
On the Award day, all members of the 5 teams will be able to attend the AIC4NL Dutch AI Congress. Each team will present their implementation proposals to a jury of industry experts from AIC4NL.
The final ranking will be done as per the combination of the AIC4NL evaluation of the teams’ implementation proposals and performance scores from the previous stage. Accordingly, the total prize pool of € 7.000 will be distributed as shown below.
🥇 1st place: €3,000 🏆
🥈 2nd place: €2,000
🥉 3rd place: €1,000
The runner up teams (4th and 5th) will receive €500 per team.
Kick-off day presentations
Nomination Leaderboards (Top 30 validated teams)
The top 5 teams on the Overall Leaderboard have been nominated and are set to present to a jury at the Dutch AI Congress by AIC4NL. The jury will decide on the final ranking.
Overall Ranks
| Rank | Team Name |
|---|---|
| 1 Nominated | Stochastic Gulls Descent |
| 2 Nominated | Geto-Dacians |
| 3 Nominated | Determined |
| 4 Nominated | Echo |
| 5 Nominated | Teamgazo |
| 6 | Chicken Jockey |
| 7 | Neural Nexus |
| 7 | Trained On Rakia |
| 9 | Earth & Moon |
| 10 | Gluck Gluck |
| 11 | Bonis Avibus |
| 12 | Egg First |
| 13 | Gloria Buzau |
| 14 | Kmr |
| 15 | Birdwatchers |
| 16 | What The Flock |
| 17 | Kaiming Kings |
| 18 | Sofya Grishkova |
| 19 | Did You Bring The Brain? |
| 20 | Spear |
| 21 | Stupidbirds |
| 22 | Cbu |
| 23 | The Cormorant Problem |
| 24 | Whoami |
| 25 | Chungus |
| 26 | Montana Management |
| 26 | Verdant Phantom |
| 28 | Polska Gurom |
| 29 | Subarctic1712 |
| 30 | Thought Too Hard |
Performance Rank (60%)
| Rank | Team Name |
|---|---|
| 1 Nominated | Stochastic Gulls Descent |
| 2 Nominated | Geto-Dacians |
| 3 | Trained on Rakia |
| 4 | Chicken Jockey |
| 5 Nominated | Echo |
| 6 | Birdwatchers |
| 7 Nominated | TeamGazo |
| 8 Nominated | Determined |
| 9 | Earth & Moon |
| 10 | gluck gluck |
| 11 | Neural Nexus |
| 12 | CHAIKA |
| 13 | Kaiming Kings |
| 14 | Bird |
| 15 | Egg First |
| 16 | Did you bring the brain? |
| 17 | Gloria Buzau |
| 18 | Bonis Avibus |
| 19 | What the flock |
| 20 | Verdant Phantom |
| 21 | StupidBirds |
| 22 | Montana Management |
| 23 | chungus |
| 24 | Spear |
| 25 | KMR |
| 26 | Rambler |
| 27 | Last Minute |
| 28 | Polska Gurom |
| 29 | CBU |
| 30 | MadEye |
Implementation Rank (40%)
| Rank | Team Name |
|---|---|
| 1 Nominated | Stochastic Gulls Descent |
| 2 Nominated | Determined |
| 3 Nominated | Geto-Dacians |
| 4 | Sofya Grishkova |
| 5 | Kmr |
| 6 Nominated | Teamgazo |
| 7 Nominated | Echo |
| 8 | The Cormorant Problem |
| 9 | Bonis Avibus |
| 10 | Lev & Mitch |
| 11 | Subarctic1712 |
| 12 | Gloria Buzau |
| 13 | Egg First |
| 14 | Neural Nexus |
| 15 | Segmentation Fault |
| 16 | J_Py |
| 17 | What The Flock |
| 18 | Gluck Gluck |
| 19 | Earth & Moon |
| 20 | Whoami |
| 21 | Cbu |
| 22 | Chicken Jockey |
| 23 | Spear |
| 24 | Thought Too Hard |
| 25 | Grain |
| 26 | Trained On Rakia |
| 27 | Did You Bring The Brain? |
| 28 | Kaiming Kings |
| 29 | Stupidbirds |
| 30 | Dreemteem |
TIMELINE
Date
Milestone
Friday, February 13
Event | Kick off day and start of the challenge
Saturday, March 14
Deadline for Team Validation
Tuesday, March 17
Deadline for Implementation track submission
Saturday, March 21
Deadline for getting your AI Futures day ticket
Tuesday, March 24
Deadline for Performance track submission
Saturday, March 28
Event | AI Futures Day: Career Fair + Nomination and Private leaderboard reveal
Tuesday, April 14
Event | Award Day at AIC4NL Congress
* All deadlines are at 11:59 PM Central European Time (CET), unless stated otherwise.
Team Validation
To verify eligibility, participants are expected to complete the validation form before the deadline. If you fail to do so, your solution will not be considered. Every week we will update the Challenge Center with the validated teams, use it to check if your team validation was processed correctly.
To be eligible for prizes the entire team must:
- Consist of BSc, MSc, or PhD students, or recent graduates
- Fill in the validation form
- Be present at the nomination day (at least 1 person per team)
FAQ
Only students (Bachelors, Masters or PhD) and recent graduates (graduated in 2025) may officially participate in AI Cup 2026. We may also welcome a limited amount of guests to compete for “fun” without any money prize involved.
To be eligible for prizes:
- The entire team must consist of BSc, MSc, or PhD students, or recent graduates (graduated in 2025).
- Your entire team must have filled the validation form individually before the team validation deadline.
- At least one member of you team must be physically present at the nomination day.
Make sure to bring an official document with a picture to the nomination day. If no one from your team will be able to attend the Nomination day under any circumstances, let us know in advance via the email [email protected]!
We strongly advise that every member of the team fill out the form because this helps us with the organization of the in person events and it will allow your team to stay up to date with the latest developments by receiving update emails with the next steps.
Yes! To be able to receive a prize, you must complete the team validation form before the end validation deadline.
The teams are managed in Kaggle, which is the platform used for the submission of the performance track. Even if your team is not making submissions in the performance track, all your teammates must be in your Kaggle team. You are not allowed to change the name of your team name after submitting your team validation.
At the Kick-off we will be offering a team matching exercise to help you find your teammates. If the Kick-off has passed and you are still looking for a teammate, use the “Team” tab on Kaggle to turn on the notification that you are looking for a team (this will show up on the leaderboard, so make a submission). You may also use the discussion forum to create a new topic to look for teammates.
No, we encourage you to use your existing Kaggle account, or register with the email you prefer, so you can build your AI Cup experience within your chosen account.
To join the competition in Kaggle for the first time the participant must use the performance track buttons here at the challenge center. We also make use of a validation form, which must be filled by the participants before the validation deadline. After the deadline participants that are not verified won’t be considered in further evaluations and will be hidden in the Kaggle leaderboard.