Faucets

Where the word comes from, and what transferred

The faucet and sink vocabulary was not invented in crypto. It comes from virtual-economy research and game design, where a faucet is a mechanism that generates new wealth out of nothing and a sink is a mechanism that permanently removes wealth from circulation. Academic work on MMORPG economies traces the terminology through Lehdonvirta and Castronova's treatment of virtual economies, and uses it to analyse currency health as a running balance of the two.¹

Game studios have been operating these systems at scale for two decades and their language is blunter than ours. Designers talk about MIMO, money in and money out, and treat a currency's condition as the arithmetic between them.² The crypto adaptation came later, when token designers borrowed the same framework wholesale.

One thing did not transfer, and it matters more than the vocabulary. A studio can patch its economy in a Tuesday build. You ship your faucets as a contract and sometimes as a consensus rule, so the design has to be right at launch in a way a game economy never does.

The four taps almost every token opens

Work rewards pay participants for something the network needs: block production, storage, bandwidth, compute, market making. Bitcoin's block subsidy is the reference implementation, issuing new coins to whoever produced the block on a schedule that halves every 210,000 blocks.³ Usage incentives are different in kind. They pay people to behave as though demand exists before it does, and they are a marketing budget denominated in equity.

Insider vesting is the third tap, and the one founders least like to call a faucet. Tokens held by the team, the treasury and early investors are not in circulation until they unlock, at which point they are indistinguishable from any other new supply. Airdrops and treasury grants are the fourth, and they land in the wallets least attached to the outcome.

Only the first category has a coherent defence, and only when it is genuinely paid for. A work reward funded from fees the protocol earned is a cost of goods sold. A work reward funded by issuance is a subsidy that token holders are underwriting, whether or not the deck says so. That distinction sits on the business rather than on the mechanism, which is why we open every emission conversation with what the protocol actually sells.

Every faucet has a receiver, and the receiver has a business

The emission schedule tells you how many tokens appear. It says nothing about what happens in the following hour, and that is the part that moves the market. A validator with electricity, hardware and staff sells a portion of rewards to cover costs. That is not sentiment. It is accounts payable. A liquidity mining participant with no other relationship to the protocol typically sells close to everything, because the position was underwritten as a yield trade and exits when the yield falls.

Across the 100+ projects the firm has worked on, the emission model almost always exists as a supply curve and almost never as a receiver ledger. Those are two different documents. The supply curve answers how much. The receiver ledger answers who gets it, what business they are running, and what they will do on the day it lands.

Be careful with any specific sell-through number, including ours. No source we would cite establishes a constant for what share of emissions is sold on receipt. What you can do is estimate per receiver class, write the assumption down, and check it against onchain behaviour once the tap is running. An assumption you wrote down is falsifiable. An assumption you did not is just a mood.

The taps you can close and the taps you cannot

Faucets differ enormously in how reversible they are, and teams routinely learn this in the wrong order. A vesting release is contractual. It sits in a signed agreement, and renegotiating it costs goodwill you may not have. Consensus-level issuance is harder still, since changing it means a client release and a social process to match.

Governance-set incentive rates are the flexible tap. An emissions parameter a token holder vote can turn down is a lever you will want, and the time to install it is before you need it. Our guidance is to concentrate your uncertainty in the adjustable tap and keep the irreversible ones conservative. You will be wrong about demand. Design so that being wrong is survivable rather than permanent.

Why faucet design is the easy half

Anyone can write an emission schedule. It is a decay curve and a set of buckets, and it produces a chart that looks like a plan. Sink design is the hard half, because a sink has to be something people genuinely want to do, and you cannot schedule that into existence. Which is why the two belong on the same page of a model even though they sit on different pages of a glossary.

The rule we apply is that a faucet ships with a named sink capable of absorbing it, and an unmatched faucet is treated as a defect in review. That is a firm design rule and we will say plainly that it is a practitioner heuristic rather than an established result. No study we are aware of fixes a threshold ratio at which a token model becomes unsound. What we can say is that in review the unmatched faucet is where the model breaks first, and it breaks in the direction you would expect.

What we make a team write down before approving an emission

Five lines per faucet, before it goes into the model. The receiver class, named specifically enough to be countable. What the protocol buys with the emission, in the protocol's terms rather than in narrative terms. The expected sell-through for that receiver class, with the reasoning attached. The sink the emission is matched against. And the condition under which the tap closes, expressed as a metric rather than a date, so that closing it is a measurement instead of an argument.

A faucet that cannot fill all five lines does not go in. That sounds strict until you price the alternative: an emission you cannot justify, cannot switch off, and cannot explain to a holder watching supply grow while the product does not. The token is infrastructure for a business. If the business is not generating the demand that absorbs the emission, no schedule fixes it, and the schedule was never the problem.