Entry timing, commitment, and installed capacity
A market is already running when a new firm has to decide whether to step into it — and that decision isn't just yes or no, it's a question of how large a position to commit to. Meanwhile, the incumbents already inside face their own version of the same tension: how much capacity to commit to before knowing whether a new entrant will show up and reshape the market they're building for. Incumbents aren't purely reactive, either — their own behavior can deter entry altogether, keeping a new competitor out before it ever arrives. Students feel both sides of that trade-off, often in the same session.
The simulation supports several different configurations of that tension, without switching tools:
- Sunk vs. flexible installed capacity. Whether installed capacity is a one-way ratchet — once committed, it can never be reduced, only added to — or fully flexible, freely raised or lowered each round with cost based only on whatever's currently installed, can be configured. The same market can teach very different lessons about commitment depending on which is chosen.
- One-shot vs. repeated entry. Whether a firm that declines to enter is locked out for good, or free to reconsider in any later round, can be configured — isolating how much of the "first-mover" story depends on entry actually being irreversible.
- Single vs. multiple entrants. More than one prospective entrant can compete for the same market at once, turning the entry decision into a race rather than a solo call — and giving incumbents more than one rival's behavior to react to.
The Capacity Race
Each team runs a GPU compute provider, renting out processing power to AI companies that need it. Every round, incumbent teams decide how much GPU capacity to install; price falls as total installed capacity across every active provider rises. Starting from a configurable round, a new entrant can choose whether to step in and compete too — reshaping the market for everyone already there.
Installed capacity can be configured as a sunk investment — once committed, it can only go up, never down. Build too much, and you're stuck paying for it whether or not the market ends up rewarding it, exactly like the real over-build risk playing out in this industry today.
