The Fee Market

Every ten minutes or so, the Bitcoin network produces one new block, and that block can hold only so much. That hard limit turns block space into a scarce good — and where there’s a scarce good and many buyers, there’s a market. The fee market is how Bitcoin decides whose transactions get into the next block when not everyone can fit.

The scarce good is block space

A block isn’t measured in “number of transactions.” It’s measured in weight units, capped at 4,000,000 weight units per block (often described as a “block size” of up to ~4 MB in the best case, ~1–2 MB in typical practice). Every transaction consumes some of that budget depending on how much data it takes to express.

The key facts:

• Supply is fixed and inelastic. No matter how many people want to transact, each block offers ~4M weight units, and blocks arrive at a roughly steady rate set by difficulty. You cannot conjure more space by paying more — you can only out-bid others for the space that exists.
• Demand is variable. Sometimes the network is quiet; sometimes a surge of activity wants in all at once.

When demand for space in the next block exceeds supply, something has to ration it. Bitcoin rations it with price.

Users bid in sat/vB

Transactions don’t pay a flat fee — they pay a fee rate, measured in satoshis per virtual byte (sat/vB).

Why “per byte,” not “per transaction”

A miner choosing what to include wants to maximize fees earned for the limited space available. A transaction that pays 10,000 sats but eats 1,000 vB (10 sat/vB) is less attractive than one paying 5,000 sats in 250 vB (20 sat/vB), even though it pays more in absolute terms. Miners are buying revenue per unit of space. So the price that matters is the rate, sat/vB — and that’s what your wallet is really setting.

This is a continuous, blind auction. Each user attaches the fee rate they’re willing to pay; miners, acting purely in self-interest, sort the waiting transactions by fee rate and pack the highest payers into the next block until it’s full. The result is price discovery: there’s no posted price list, just thousands of independent bids resolving into a clearing rate, block after block.

   The mempool, sorted by fee rate (high → low)
   ┌──────────────────────────────────────────────┐
   │  60 sat/vB  ┐                                 │
   │  45 sat/vB  ├─ fit into the NEXT block (top   │  ← miners take
   │  30 sat/vB  ┘   ~4M weight units of bids)     │     the top bids
   │ ───────────  clearing rate (the cutoff) ───── │
   │  12 sat/vB  ┐                                 │
   │   5 sat/vB  ├─ wait for a future, cheaper     │  ← low bids
   │   1 sat/vB  ┘   block                         │     wait
   └──────────────────────────────────────────────┘

That waiting room of unconfirmed transactions is the mempool — each node’s view of pending bids.

Congestion dynamics

Because supply is fixed and demand swings, fees are naturally volatile:

• Quiet periods: if all waiting transactions fit in the next block, there’s nothing to bid against. Fee rates fall toward the network’s tiny minimum — you can transact for almost nothing.
• Surges: when far more transactions arrive than fit, users start out-bidding each other for the scarce slots. The clearing rate spikes; low-fee transactions can wait hours, days, or get dropped from mempools entirely.
• Self-correction: high fees suppress non-urgent demand (people wait), which eventually clears the backlog and pulls fees back down. It’s a thermostat made of incentives.

Your wallet’s “fast / medium / slow” options are just estimates of what rate is likely to clear in the next few blocks versus over the coming hours. Because a transaction’s fee can become stuck if conditions change, Bitcoin offers escape hatches: RBF (broadcast a replacement version of the transaction paying a higher fee) and CPFP (a child transaction pays to drag a stuck parent along). Those mechanics, and the question of when an unconfirmed transaction can be trusted, are covered in RBF/CPFP & the mempool. The arithmetic of how fees are actually deducted — inputs − outputs = fee, with change handled explicitly — lives in fees & change.

How the fee market helps strangers agree on one ledger

It’s worth pausing on the textbook’s recurring question here, because the fee market answers it in a way that’s easy to miss. Block space is the one truly scarce resource inside the system, and an open auction is a neutral, permissionless way to allocate it. No one decides whose payment matters more; no committee approves transactions. Anyone, anywhere, can bid, and the same rule (highest fee rate wins the scarce space) applies to all. That neutrality is part of why untrusting strangers can share one ledger: the rationing rule plays no favorites and requires no trusted gatekeeper.

Why fees matter long term

In Bitcoin’s early decades, fees are a rounding error next to the block subsidy. But as covered in miner incentives, the subsidy halves toward zero, and eventually fees must carry the entire security budget. That makes the fee market not just a congestion mechanism but the long-run economic foundation of the network’s security.

An honest tension

A healthy long-run fee market needs sustained, robust demand for block space — yet high fees also push everyday payments off-chain (onto layers like Lightning, or onto exchanges’ internal ledgers), which can reduce on-chain fee demand. Bitcoin must thread the needle: enough fee pressure to fund security, not so much that it strangles the very usage that generates fees. Whether on-chain demand will reliably clear at a high-enough rate, a century out, is unsettled. It’s the same open question from the previous page, viewed from the demand side.

It actually happened — when block space got genuinely expensive (2017, 2023–24)

The fee auction is usually cheap, but real demand surges have shown what scarcity pricing feels like. In late December 2017, a wave of activity overwhelmed block space and the average on-chain fee climbed above roughly $50 — briefly making small payments uneconomical and driving home that you’re bidding for a fixed resource. In 2023, the Ordinals / BRC-20 craze (see ordinals & data) packed blocks with inscription and token-mint data, keeping the mempool congested for months and repeatedly spiking fee rates. The sharpest preview came at the April 2024 halving block (height 840,000): the launch of the Runes protocol set off such a fee frenzy that the block’s fee revenue set a record and far exceeded its brand-new 3.125 BTC subsidy — a live glimpse of the fee-dominated future the previous page described.

The architect’s lens

The fee market is a deliberate rationing mechanism — the five questions sharpen why an auction:

• Why does it exist? Because block space is a fixed, inelastic supply (~4M weight units per ~10 minutes) with many buyers — and a scarce good with many buyers needs rationing. Bitcoin rations it with price, via a continuous blind auction in sat/vB.
• What problem does it solve? Allocating scarce blockspace neutrally and permissionlessly — no committee approves transactions, no one decides whose payment matters; highest fee rate wins, the same rule for everyone. It also becomes the long-run security foundation as the subsidy fades.
• What are the trade-offs? Fees are naturally volatile — near-free when quiet, spiking to hours-long backlogs in surges (Dec 2017’s ~$50 average, the 2023 Ordinals/BRC-20 waves) — and there’s an honest tension: high fees push everyday payments off-chain, which can reduce the very demand security relies on.
• When should I avoid it? Auction-rationed blockspace is the wrong fit for high-volume, low-value everyday payments — those belong on Lightning or other layers; the base auction is for settlement, not coffee.
• What breaks if I remove it? With no price signal, fixed blockspace must be rationed some other way (first-come, or a gatekeeper) — reintroducing a trusted allocator — and once the subsidy is gone, nothing would fund the security budget at all.

Check your understanding

1. What exactly is the scarce good in the fee market, and what fixes its supply per block?
2. Why do miners and wallets care about fee rate (sat/vB) rather than the total fee paid?
3. Describe what happens to the clearing fee rate during a demand surge, and why it self-corrects.
4. In what sense is the fee auction “neutral,” and how does that neutrality support a shared ledger among strangers?
5. Why does the fee market become more important to Bitcoin’s security over time, not less?

Show answers

1. The scarce good is block space, measured in weight units and capped at 4,000,000 weight units per block. Supply is fixed and inelastic — blocks arrive at a roughly steady rate set by difficulty, and you can’t conjure more space by paying more, only out-bid others for what exists.
2. Because miners are buying revenue per unit of space, not absolute fees. A transaction paying 10,000 sats over 1,000 vB (10 sat/vB) is less attractive than one paying 5,000 sats in 250 vB (20 sat/vB), so the rate sat/vB is what decides inclusion — and what your wallet is really setting.
3. During a surge far more transactions arrive than fit, so users out-bid each other and the clearing rate spikes, leaving low-fee transactions to wait hours/days or get dropped. It self-corrects because high fees suppress non-urgent demand (people wait), which clears the backlog and pulls fees back down — a thermostat made of incentives.
4. It’s neutral/permissionless: no committee approves transactions and no one decides whose payment matters more — anyone, anywhere can bid, and the same rule (highest fee rate wins the scarce space) applies to all. That gatekeeper-free, plays-no-favorites rationing is part of why untrusting strangers can share one ledger.
5. Because the block subsidy halves toward zero, and eventually fees must carry the entire security budget. That makes the fee market not just a congestion mechanism but the long-run economic foundation of the network’s security — though it must thread the needle: enough fee pressure to fund security without strangling the on-chain usage that generates fees.