What Comes After the Last Halving? Fees as the Incentive Model

In short: After the last halving around 2140, miners no longer receive new bitcoin, only the fees of the transactions in their block. These fees then have to pay for the entire security of the network. As of November 2026 they make up less than 1% of miner revenue, so the gap is large. But it does not have to be closed in 2140; it closes step by step over the coming decades, because the subsidy shrinks with every halving.

Among bitcoiners the question is called the "security budget": how much money flows to miners each year so that they provide computing power? The more computing power, the more expensive an attack on the network would be. This article works out what the budget looks like today, when fees will have to carry the main load and which scenarios support that. What the end of issuance means for the Power Law model is covered in the neighbouring article What happens to the Power Law after the last halving?.

What miners live on today

Whoever finds a new block may pay themselves two things in it:

1. The subsidy (block subsidy): new bitcoin that did not exist before. Since the halving in April 2024 it has been 3.125 BTC per block.
2. The fees: the sum of all fees the senders of the transactions in the block offered.

Together they are the block reward. How the halving reduces the subsidy is explained in The Bitcoin halving explained simply.

Satoshi planned the transition from the start. The whitepaper states in the section "Incentive" that once a predetermined number of coins have entered circulation, the incentive for miners can transition entirely to transaction fees. The system is then completely free of inflation.

How small fees are right now

Data from mempool.space shows what share of miner revenue came from fees. The figures compare the yearly average of fees per block with the subsidy:

Two things stand out. Fees swing widely: in phases of heavy demand, such as late 2017 or 2023, when inscriptions (Ordinals) took up a lot of block space, they rose markedly. And 2025 and 2026 were exceptionally quiet years. In the 144 blocks up to 13.11.2026, 4.27 BTC came from fees and 450 BTC from the subsidy.

Individual days show what is possible. On 20.04.2024, the day of the fourth halving, fees brought in about 9.7 BTC per block on the daily average, more than the new subsidy of 3.125 BTC. On 22.12.2017 it was about 9.7 BTC per block against a subsidy of 12.5 BTC at the time. Such spikes never last long, though. What lies behind these fee waves is covered in Mempool analysis: what transaction fees reveal about the market.

The subsidy, epoch by epoch

An epoch lasts 210’000 blocks, about four years at ten minutes per block. In each epoch the subsidy is half as large as in the one before. The years assume that from today a block arrives every ten minutes on average. Only the block numbers are fixed.

So the switch does not happen all at once in 2140. Around 2140 only the very last sat disappears. The real shift from subsidy to fees takes place from the 2030s to the 2050s.

The chart shows how many bitcoin are created in total. After 2035 the curve is practically flat, because 99% of the supply has been mined by then:

When fees overtake the subsidy

Put today's fees next to the table. In 2026 fees brought in an average of 0.0201 BTC per block. In epoch 11 (from 2052) the subsidy of 0.0244 BTC is still slightly larger; in epoch 12 (from 2056), at 0.0122 BTC, it is smaller. Even if fees never rise in bitcoin terms, they will be miners' larger income around 2056.

At the 2023 fee level, 0.44 BTC per block, parity would already come in epoch 7 (from 2036, 0.39 BTC subsidy). The answer to "when will fees carry the network?" therefore depends heavily on which year you consider normal.

What fees mean for the individual

A more tangible question is: how much would each single transaction have to pay for fees to replace the subsidy? A block currently holds about 4’000 transactions. In the 144 blocks up to 13.11.2026 the average was 3’961.

The table shows the core of the debate. Replacing today's subsidy with fees would mean about €48 per transaction. Nobody pays that today for an everyday payment. From the 2040s on, however, fees of one to two euros are enough to keep the budget level in bitcoin terms. How your own fee is made up and how to avoid overpaying is explained in Bitcoin transaction fees and the mempool.

Three scenarios in which fees are enough

1. The price rises

Fees are paid in sats, while miners' costs (electricity, machines, staff) are in euros and dollars. What counts is the value of the budget, not the number of bitcoin. If a bitcoin is worth ten times as much, the same fee in sats buys ten times as much electricity.

A higher price therefore helps twice: it keeps the dollar value of the shrinking subsidy up, and it makes every fee in sats worth more. How the subsidy develops in dollars if the price follows the Power Law is worked out in the neighbouring article on the last halving. Its result: until the end of the 2030s, the subsidy would even gain value in dollars despite the halvings. That pushes the problem back, but does not make it disappear.

2. More demand for block space

Block space is scarce. A block holds at most 4 million weight units, which is 1 million vbytes. At about 4’000 transactions per block, about 210 million transactions fit into a year.

If only a few million people use Bitcoin directly on the blockchain, there is room enough and fees are low, as in 2025 and 2026. If many more people and companies want to hold and move their own bitcoin, they bid against each other for the same space. Then fees rise without anyone changing the rules.

3. Layer 2 settles on the blockchain

Networks like the Lightning Network move payments off the blockchain. That sounds like fewer fees for miners, but it is only half the story. Every Lightning channel has to be opened and closed on the blockchain. A single transaction on the blockchain can thus secure thousands of payments.

This changes who pays fees: fewer coffee buyers, more settlements by exchanges, payment providers and Lightning nodes with a lot of money at stake. A settlement of a million euros can easily bear a fee of €20 or €50. The blockchain thus becomes more of a settlement layer, similar to how banks settle with each other through large, infrequent payments.

The debate: will fees be enough?

Whether this transition works is a genuinely open question. Serious developers and researchers see it differently.

The concerns

• Volatile income: fees come in waves. A network that lives only on fees has less protection in quiet phases. The figures for 2025 and 2026 show how low fees can fall.
• Budget relative to value: the more value a network secures, the more an attack pays off. If the value of all bitcoin grows faster than the budget, an attack becomes cheaper in relative terms. The developer Paul Sztorc put this concern sharply early on.
• Fee sniping: researchers at Princeton University (Carlsten, Kalodner, Weinberg, Narayanan, 2016) showed that mining without a subsidy can become unstable. If a block holds a lot of fees, it can pay to re-mine it and collect the fees yourself instead of building on top of it.

There is already a simple defence against fee sniping: Bitcoin Core and other wallets set the current block height as the earliest valid block (nLockTime) for new transactions. Such a transaction cannot move into a re-mined older block. A full mempool with many waiting transactions dampens the incentive further, because the next block is well paid anyway.

Tail emission: the proposal that is not catching on

An obvious solution would be to never let the subsidy fall all the way to zero, but to issue a small amount of new bitcoin per block forever. This is called tail emission. The developer Peter Todd argued for it in 2022, on the grounds that it roughly offsets lost coins. The cryptocurrency Monero has had such a permanent issuance since 2022.

In Bitcoin this proposal has found no broad support, and there is no serious attempt to implement it. The reasons:

• For most users the 21 million limit is the core of the promise. Changing it once shows that it can be changed.
• A higher subsidy is a hard rule change (hard fork). Blocks with more new bitcoin are invalid under today's rules; every node would have to install new software.
• The question is not urgent. The subsidy is large enough for decades to come, and by then it will be clear how the fee market develops.

Most bitcoiners therefore expect: the rules stay, and fees have to fill the gap.

Why the question is decades away

For you as a holder there are three reassuring points.

First, the network keeps working with fewer miners. If revenue falls and miners switch machines off, mining becomes easier after at most 2’016 blocks. Blocks keep arriving about every ten minutes. How that works is explained in Difficulty adjustment: the heart of Bitcoin. What is open is not whether Bitcoin keeps running, but how much computing power protects it.

Second, the subsidy stays large for a long time. Until 2036 it is at least 0.78 BTC per block, more than fees in any past year except 2017. Until then the price above all decides the budget.

Third, there is time to learn. Every halving is a test run: when the subsidy falls, it shows whether hashrate and fees keep up. So far, computing power has kept growing across every epoch, even if it can dip briefly right after a halving. Miners, developers and researchers are watching closely whether that continues. Solutions that do not change the rules, such as better use of block space or new Layer 2 networks, can arrive step by step.