A two-sided market is a platform that connects two groups, such as operators and end users on a DePIN network, where each side's value depends on the other showing up first. Token incentives exist to break that circularity by subsidizing whichever side is the real binding constraint.
Subsidizing supply when demand is the bottleneck is not a strategy, it is a delay, and node count is the vanity metric that hides it.
How it works
In DePIN the two sides are infrastructure operators who supply compute, coverage, or storage, and buyers who pay for it. The protocol sits in the middle and sets the prices, rules, and incentive parameters that move value across both sides.
The cold-start problem bites hard here because operators commit real capital, often hundreds to thousands of dollars per node, before any buyer has confirmed the supply is worth purchasing. Neither side wants to be first.
Design consequence
Tokens subsidize the supply side long enough for demand to grow. The real decision is identifying which side is the binding constraint. In most early DePIN networks supply is easy to buy with rewards while demand is the genuine constraint, so successful designs pivot capital toward demand once coverage hits a threshold.
Rochet and Tirole showed that two-sided pricing is not single-sided pricing split in half: cross-group externalities force the platform to engineer a price structure that maximizes total participation. Emission schedules and fee structures encode that logic whether the team intends it or not.
Common mistake
The failure mode is persistent imbalance: the protocol keeps subsidizing supply because node count looks good, while demand never reaches critical mass. The token becomes a supply-side transfer payment with no commercial activity behind it, and emissions turn into dilution without purpose.
See DePIN Tokenomics Guide for how this applies in practice.
Know the terms but not sure how they apply to your project? That is what an engagement is for. We design, document, and stress-test the whole token economy inside the Tokenomics Data Room.
80+ projects advised. Complete tokenomics in 4 to 6 weeks.