Hackathons are not pitch competitions. The sooner you internalize that, the better your chances.
Every year teams show up with decks. They have TAM slides, revenue projections, go-to-market strategies. They have thought deeply about their customer segment. They lose.
Judges and voters don't care. Nobody is writing a check. Nobody is acquiring your weekend project. The question being asked is not "would this make money?" or "What business problem does this solve?" The question is "is this cool?"
The voting criteria that actually move the needle are innovation, difficulty, uniqueness, and the pure wow factor. That's it. If your project doesn't make someone say "how did they build that in 24 hours" you are not winning.
Business thinking is not just neutral at a hackathon — it is actively harmful. It crowds out the creative risk-taking that wins. The team that builds something technically audacious and half-broken will beat the team with a polished pitch every time.
Something interesting is happening. Product managers and designers are showing up to hackathons without a single engineer on their team and competing seriously.
AI has made this possible. What used to require a backend developer, a frontend developer, and probably a weekend of debugging can now be assembled through prompts. Non-engineers are shipping working demos in hours.
This is genuinely new. Hackathons used to filter hard on technical ability by default. You needed people who could actually build things. That barrier is collapsing. The question of who gets to participate is being rewritten in real time.
Whether this is good or bad for the format is worth thinking about. The playing field is flatter. But the ceiling for what a strong engineering team can do is also higher, because they can move faster too.
The other shift is in demos. Recorded videos with AI voiceovers are now the default submission format at many hackathons. A synthesized voice walks through a screen recording. No live demo. No real person presenting.
This makes sense from a risk management perspective. Live demos fail. Recorded demos don't. But something is lost when there's no human in the room making the case for what they built.
The best hackathon presentations have always been a little chaotic. Someone is nervous. The demo almost breaks. They recover. You feel the 24 hours in the room with them. That tension is part of why it works.
A polished AI narrated video removes all of that. It also makes it easier to obscure how much was actually built versus mocked up. When a real person demos live, you learn pretty quickly what's real.
Hackathons are adapting to a world where AI changes who can build and how fast. The format will keep shifting. But the core of what makes a winning project hasn't changed: build something surprising, make it work well enough to be convincing, and make people feel something when they see it.
Business plans won't do that. A polished deck won't do that. Show something nobody expected and you have a shot.