Inaugural cycle · 2026–27
How it works
Seven weeks to build an AI startup, two required videos to submit, and a public rubric you can read before you start.
Eligibility
Who can compete
- Age
- 13–18 during the competition cycle
- Location
- Anywhere in the world
- Team size
- Solo, or teams of up to five
- Experience
- None required
- Format
- Fully online, including the finals — no travel
- Cost
- Free — no entry or submission fee
Calendar
The five phases
- 1
Aug 20 – Nov 1
Registration
Sign up free, individually or as a school group. Registered teams receive resources by email throughout the fall.
- The build-phase playbook
- Founder resources & templates
- Competition updates
- 2
Nov 2 – Dec 20
Build phase (7 weeks)
Teams spend seven weeks turning an idea into a startup. Optional milestones help you stay on track.
- Customer discovery
- Market validation
- MVP planning
- Pitch practice
- 3
Deadline Dec 20
Submission
The portal is open for the whole build phase — submit any time before the deadline.
- 3-minute pitch video
- 2-minute founder Q&A
- Pitch deck (recommended)
- Prototype (optional)
- 4
Dec 22 – Jan 20
Judging
Every submission is scored against the six published criteria in multiple independent passes, and every team receives a full breakdown of their scores.
- Multiple scoring passes, averaged
- Six weighted criteria
- Written feedback for all entrants
- 5
Jan 22 – Feb 14
Online finals
Finalists pitch live over video in individually scheduled sessions. You pick a slot that works in your time zone rather than staying up for someone else's — no travel, no cost.
- 5-minute pitch over video
- 5-minute Q&A with the organizing team
- You choose your session time
- Results announced February 14, 2027
Submission
What you submit
Judges primarily evaluate the quality of the idea, problem selection, business potential, and AI implementation. A polished prototype is not expected.
Required
3-minute pitch video
Every team submits a short presentation introducing their startup.
- Problem & solution
- Why AI?
- Target users
- Business model
- Market opportunity
- Long-term vision
Required
2-minute founder Q&A
An unscripted video answering three prompts published at the start of the build phase. This is how we verify you understand what you built.
- What did you build, and what did AI build?
- The hardest decision your team made
- What you'd do next with another month
Highly recommended
Pitch deck (8–10 slides)
Not required, but a strong deck often helps judges better understand the startup.
- Problem · Solution · AI technology
- Market · Product · Business model
- Competitive advantage · Go-to-market
- Vision · Team
Optional
Prototype / MVP
Extra evidence of feasibility — a polished product is not expected.
- Website or mobile app
- AI chatbot or ML model
- Figma prototype
- Interactive demo
Scoring
The rubric
These six criteria are published before the build phase opens. Nothing is scored on anything you cannot read here.
| Criterion | Weight |
|---|---|
| Problem importanceDoes the startup address a meaningful, real-world problem worth solving? | 20% |
| InnovationIs the approach original — a genuinely new angle rather than a copy? | 20% |
| AI implementationHow creatively and effectively is AI woven into the product itself? | 20% |
| Product qualityIs the product thoughtful, usable, and well executed for its stage? | 15% |
| Business potentialCould this become a real company — market, model, and traction path? | 15% |
| PresentationIs the pitch clear, confident, and compelling to judges? | 10% |
How scoring works
Scored by AI. Overseen by us.
Every submission is scored against the six published criteria by an AI evaluation system built for this competition and given the same rubric you can read on this page. Each entry is scored in multiple independent passes and the results averaged, so no single run decides a placement. The organizing team oversees the process and personally reviews every submission that reaches the top tier before finalists are named.
- The same rubric for everyone
- There is no panel whose taste you have to guess. Every entry is measured against the six weighted criteria published before the build phase opens.
- Multiple passes, averaged
- Each submission is scored several times independently and the scores averaged. A single unusual result cannot decide your placement.
- Humans review the top tier
- Scoring narrows the field. Before finalists are named, the organizing team reviews every submission in the top tier by hand — including your founder Q&A.
- You can ask why
- Every team receives a full breakdown of their scores against each criterion. If you think something was misread, you can request a human re-review within 14 days.
How we handle AI-generated submissions
Rules
Compete with integrity
- 01Projects must be created for this competition cycle by the registered team.
- 02AI tools are permitted and encouraged at every stage of development.
- 03Teams must understand and be able to explain every aspect of their project, regardless of which AI tools were used.
- 04All submissions must be the team's own work. Plagiarizing another team's startup is grounds for disqualification.
- 05Submissions must be in English and uploaded through the official portal by the deadline.
- 06Judges' decisions are final.
Permitted AI tools
- ChatGPT
- Claude
- Gemini
- Cursor
- GitHub Copilot
- Lovable
- Bolt
- Replit AI
- Midjourney
- Flux
- Runway
- ElevenLabs
…and any other AI system.
Problem areas
- Healthcare
- Education
- Finance
- Climate
- Accessibility
- Robotics
- Scientific research
- Productivity
- Public safety
- Agriculture