Competition needs new economics

I. Eight billion people, a hundred stages

Competition is one of the largest creator economies on the planet. Esports was forecast to reach $48 billion by 2034 when I first wrote this, in February 2026; add hackathons, data-science challenges, art, music, writing, film and scientific prizes, and the sums run far higher.

And almost none of it works for the people competing.

Traditional competitions are bound by geography, gated by money, and judged behind closed doors. A filmmaker in Lagos cannot enter a festival in Cannes without the airfare. A developer in Dhaka cannot afford a hackathon with a $500 registration fee. A musician in Buenos Aires plays for local judges who have never heard anything like what she is building.

The stages exist. The talent exists. The infrastructure that connects them does not.

Web3 promised to fix this: decentralise the stages, open the gates, let merit win. Instead it produced a new set of broken promises: engagement farming dressed up as competition, tokens that dump on their own communities, platforms that extract more than they give back. What competition needs is infrastructure whose economics serve the competitors, and another platform will not supply it.

II. The dual failure

The competition landscape suffers from two distinct pathologies, and fixing either one alone has not been enough.

The centralisation pathology. Traditional platforms are vertical silos: one for data science, one for hackathons, festivals for film. Each serves a niche, each is centralised, each captures value from the talent it hosts. Judging is opaque: small panels decide in private, and participants cannot verify that the result was fair. Entry costs money, travel and connections. The audience watches, and its attention is spent without ever becoming participation.

These platforms work for the people they work for. They systematically exclude most of the world's talent.

The tokenomics pathology. Web3 tried to decentralise, and its dominant economic design undid the attempt. Time-scheduled token emissions create a zero-sum game between insiders and everyone else. Teams and early investors receive large allocations that vest on a fixed calendar; the market knows when the unlocks come; holders sell ahead of them; insiders have every reason to inflate the price beforehand. "Token unlock" became a bearish signal, a synonym for insider extraction. The people who build and use a platform end up opposed to the people who funded it, and projects die of economic self-destruction long before they could fail on product.

The "quest" platforms compound it. They gamify low-skill engagement, follow an account, join a server, and call it competition: marketing funnels presented as meritocratic stages.

A real fix has to address both at once: the access problem of centralisation and the economic problem of broken emissions.

III. Any competition, on one set of rails

The first half of the answer is general-purpose infrastructure for hosting any competition: an esports bracket, a film festival, a hackathon, a poetry slam, a research contest.

Today each vertical has its own platform, its own templates, and its own claim on your community and your data. The alternative is a no-code toolkit of formats an organiser composes freely: elimination brackets from sixteen players to a thousand entries, points-based challenges across rounds, seasonal leagues with promotion and relegation. Organisers set their rules and prizes and launch without developers, without lock-in, and in time without asking anyone's permission. A DAO hosts its own hackathon; a university runs its own research challenge; a gaming community organises its own league. Each new organiser brings a community with them.

IV. Judging you can audit

The most persistent problem in competition is trust. Who decides the winner, on what basis, and can anyone check?

Three evaluation channels, blended by the organiser, answer it. Critique: a panel of credentialed experts, essential where the judgement needs deep domain knowledge. Crowd: community voting, right where popular appeal is the point. Cognition: AI analysis for whatever can be measured, code quality, musical structure, statistical method, extending the automated judges that programming contests have used for years.

An esports event might use fully automated scoring; a film festival might blend 40% experts, 30% crowd and 30% AI; a hackathon might weight experts at 60% and technical analysis at 40%. The weights are published before the competition starts, and every vote, score and analysis is recorded so that anyone can audit the outcome.

The long-term idea I find most interesting: renowned experts training digital versions of their own judgement. A great chef's palate, a master pianist's ear, made available to competitions their owner could never attend, with their consent and under their name.

V. When spectators take part

Competitions draw enormous audiences, and almost all of that energy is wasted: spectators watch, react and leave, with nothing at stake and no reason to stay between events.

Prediction markets change that. Spectators forecast outcomes, and not only who wins: whether the leader holds first place after the third round, whether entries pass five hundred, which genre dominates this year. Beyond engagement, the markets do work: they price talent, aggregate quality judgements and read sentiment in real time. When many informed people stake something on an outcome, the signal is often more accurate than any single judge.

VI. Releasing value only when it is earned

The second half of the answer is the economics, and it is the part I consider the real contribution.

Reject time-based unlocks entirely. Release new governance tokens only in response to real demand, in a four-step cycle that runs in the open.

Demand triggers. Real platform activity, fees paid, stakes placed, generates the only signals that can trigger a release.

Release. A calculated amount of new tokens is distributed to every stakeholder in the predefined proportions: team, investors, community treasury. On its own, that would be inflation.

Buyback and burn. A significant part of the platform's external revenue, earned in stable assets from fees, sponsorships and markets, buys back exactly the amount released to team and investors, and burns it.

Value injection. What remains of that revenue deepens the token's liquidity.

The effect: new tokens enter circulation only when the platform is used, insider emissions are neutralised by revenue-funded buybacks, and each release strengthens the system instead of draining it. A fork of the code cannot copy this, because the mechanism runs on external revenue, and only a product people actually use generates it. Product quality and economic health become the same thing, and the relationship between team, investors and community moves from adversarial to aligned.

When I wrote this in February 2026 I attached a three-token design and a launch sequence of flagship events. I have left both out: they were a plan, and the mechanism stands without them. It has since become more common, too: by 2026 several large protocols had moved to buybacks funded by real revenue, which is the same idea arrived at from the other end.

VII. New economics

The venture of the Fucina Nexus Program working on this problem is Ribalta, whose function is multi-species online competitions with verifiable economic mechanisms. It is in the pipeline; its first functional proof is a configurable event with judging, challenge, finalisation, and a sandbox payout and refund. Nothing more about it belongs here until it has something to show.

The world has more talent than at any point in its history: more creators, developers, artists, scientists and players. The stages have not kept up. They are too small, too expensive, too opaque and too extractive, and Web3's first attempt replaced one extraction with another.

What should have existed from the beginning is infrastructure for any competition, with economics in which nobody can rig the game. Competition needs new economics far more than it needs new platforms.