Halving
A pre-programmed cut, usually of 50%, in the new supply a network pays out per block. It fires on a protocol trigger rather than a date, usually a block height and on networks like Bittensor a cumulative issuance threshold, so the emission curve steps down rather than tapering. Bitcoin's 2024 halving took the block reward from 6.25 BTC to 3.125 BTC at block 840,000, and Bittensor's first halving in December 2025 cut daily issuance from 7,200 TAO to 3,600.
TradFi parallel: Closest to a dividend policy written into the charter rather than voted each year: the payout rate is cut on a published trigger, nobody decides it at the time, and the whole market knows it is coming years ahead.
Key Takeaways
- 01A halving cuts per-block issuance by a fixed factor on a published schedule. It is emission, not an unlock: nothing moves from locked supply to unlocked supply
- 02The trigger is a protocol rule, not a date: usually block height, and on Bittensor cumulative issuance. Bitcoin's 2024 halving was set at block 840,000 and its estimated date shifted with block time, taking the reward from 6.25 BTC to 3.125 BTC
- 03Zcash halves every 1,680,000 blocks, roughly four years: 6.25 to 3.125 ZEC in November 2020, 3.125 to 1.5625 ZEC in November 2024, with the next expected around late 2028
- 04Halving the reward is not always halving issuance. Zcash's Blossom upgrade halved the block reward and cut block time from 150 to 75 seconds at once, holding the issuance rate flat, and ZIP 208 frames it as a re-denomination that did not advance the halving counter
- 05Cutting the source cuts every claim on it: Bittensor's December 2025 halving took daily issuance from 7,200 TAO to 3,600 and shrank the emission pool subnets compete over
- 06The label is used loosely. Cosmos governance called proposal #848 the ATOM Halving, but it capped the maximum inflation rate at 10% rather than halving a block reward, and it passed narrowly, on about 51.7% of non-abstaining votes
How It Works
A halving reduces the per-block reward paid to miners or validators by a fixed factor at a fixed interval. It is an issuance mechanism, not a vesting one: no allocation moves from locked to unlocked, no lockup expires, and the tokens involved do not exist until the block that pays them is produced. That is why halvings sit on the emission side of a supply model alongside block rewards and staking issuance, rather than on the unlock calendar. The trigger is usually block height rather than a calendar date, so the date cannot be fixed in advance, only estimated. Not every schedule counts blocks: Bittensor halves on cumulative issuance instead, and an issuance trigger is estimable only to a season rather than to a few days. Bitcoin's 2024 halving was scheduled at block number 840,000 and its expected arrival moved with block time, which is why Tokenomist's own coverage named a month rather than a day. Close to the event a block-height trigger tightens: the 2028 halving currently projects to 11 to 13 April 2028 at observed 9.957 minute intervals, though that is a projection carrying about two days of block-timing noise before any hashrate drift, not a narrowing.
Zcash is the cleanest worked example because it publishes the whole ladder. Block rewards fall 50% every 1,680,000 blocks, about four years. The reward was 12.5 ZEC per block at the October 2016 launch, the first halving in November 2020 at block 1,046,400 took it from 6.25 to 3.125 ZEC, the second in November 2024 at block 2,726,400 took it to 1.5625 ZEC, and the third is expected around late 2028 at block 4,406,400. The interesting step is the one that does not count. The Blossom upgrade in December 2019 cut block time from 150 seconds to 75 seconds and halved the block reward at the same time, which held the issuance rate constant. ZIP 208 frames it as a proportional re-denomination that did not advance the halving counter, so it is a technical halving rather than an official one. Halving the reward and halving issuance are only the same thing when block cadence is unchanged.
A halving also moves whatever sits downstream of issuance. Zcash's second halving in November 2024 activated ZIP 1015, restructuring the 20% non-miner share into Community Grants at 8% and a Lockbox at 12%, so a supply event doubled as a funding-policy event. The Lockbox no longer merely accumulates: ZIP 271 executed a one-time disbursement of 78,750 ZEC to a 2-of-3 multisig held by the Zcash Foundation, the Electric Coin Company and Shielded Labs. Bittensor shows the same propagation in a different shape: its first halving in December 2025 cut daily issuance from 7,200 TAO to 3,600 and thereby reduced the total emission pool that every subnet competes over, with the next halving projected somewhere between late 2029 and 2030 depending on how much TAO is recycled, against a 21M hard cap and about 66% of the roughly 11.23 million TAO issued staked as of August 2026. Issued supply is the basis to use, because the circulating-supply field third-party trackers publish for TAO is frozen at 9,597,491 across April, June and August 2026. Cutting the source cuts every claim on it.
The word travels further than the mechanism. Cosmos governance branded proposal #848 the "ATOM Halving", but it set a maximum inflation rate of 10% rather than halving a block reward, and it passed narrowly, on about 51.7% of non-abstaining votes. The cut in annualised staking yield from roughly 19% to 13.4% is the proposal's own projection rather than a measured outcome. A follow-up proposal, #868, went after the 7% minimum inflation floor to take it to 0% and failed, so the floor still stands and live mint parameters still show inflation_min at 0.07. That is a governance parameter change, reversible by another vote, not a schedule compiled into the protocol. Dogecoin sits at the other extreme. Its last true halving was block 500,000 on 2014-12-14, taking the reward from 31,250 to 15,625 DOGE, and the move to flat perpetual issuance of 10,000 DOGE per block came later, at block 600,000 on 2015-02-25, a cut of about 36% rather than a halving. No supply cap is defined. Where the schedule is genuinely fixed, the effect is anticipated rather than discovered: Tokenomist's 2024 Arbitrum analysis explicitly modelled a large ARB unlock landing near the Bitcoin halving, on the argument that supply arriving into an anticipated demand cycle need not read the same as supply arriving into a flat market.
Real World Examples
Bitcoin (BTC): the 2024 halving at block 840,000
View →The block reward was cut in half from 6.25 BTC to 3.125 BTC at block number 840,000. Because the trigger is block height rather than a date, the expected arrival was estimated from roughly ten-minute block times and shifted between forecasts. Tokenomist covered it alongside the January 2024 spot ETF approvals as the two supply and demand narratives defining that year.
Zcash (ZEC): a published ladder, and one step that does not count
View →ZEC halves every 1,680,000 blocks, about four years: 6.25 to 3.125 ZEC at block 1,046,400 in November 2020, 3.125 to 1.5625 ZEC at block 2,726,400 in November 2024, and 1.5625 to 0.78125 ZEC expected at block 4,406,400 in late 2028. The December 2019 Blossom upgrade also halved the reward but cut block time from 150 to 75 seconds simultaneously, holding issuance flat. ZIP 208 frames that as a proportional re-denomination that did not advance the halving counter, so it is not counted as an official halving.
Bittensor (TAO): a halving that reshapes an internal contest
View →TAO's first halving in December 2025 cut daily issuance from 7,200 TAO to 3,600, with the next projected between late 2029 and 2030 depending on how much TAO is recycled, against a 21M hard cap that the Opentensor Foundation says carries no pre-mine. Because every subnet competes for a share of block emissions, weighted since dTAO by the subnet's alpha price moving average, the halving reduced the total pool each subnet is competing over, not just the headline issuance number. About 66% of the roughly 11.23 million TAO issued is staked as of August 2026, measured on issued supply rather than the frozen circulating-supply field third-party trackers publish.
Cosmos (ATOM): a governance vote wearing the name
View →Proposal #848, titled ATOM Halving, set the maximum inflation rate to 10% and passed narrowly, on about 51.7% of non-abstaining votes. Its cut in annualised staking yield from roughly 19% to 13.4% is the proposal's own projection rather than a measured outcome. It is a parameter change decided by governance, not a schedule compiled into consensus, and a follow-up proposal, #868, targeted the 7% minimum inflation floor to take it to 0% and failed, so the floor still stands. Useful as the counterexample: a halving you can vote on is a different risk object from one you cannot.
Dogecoin (DOGE): the chain that stopped halving
View →Dogecoin's last true halving was block 500,000 on 2014-12-14, which took the reward from 31,250 to 15,625 DOGE. The move to flat perpetual issuance of 10,000 DOGE per block came at block 600,000 on 2015-02-25 and was a cut of about 36%, not a halving. There is no fixed supply cap. It is the useful contrast to Bitcoin: the same proof-of-work block reward structure, but with the step-down schedule terminated, which turns a decaying emission curve into a permanent one.
Frequently Asked Questions
Is a halving the same as a token unlock?
No, and they belong on different sides of a supply model. An unlock moves tokens that already exist from a locked allocation into available supply on a vesting schedule. A halving changes the rate at which brand new tokens are minted per block to miners or validators, and the tokens involved do not exist until the block is produced. A halving reduces future issuance; an unlock adds to present float.
Why can nobody give the exact date of a halving?
Because the trigger is a protocol rule rather than a calendar entry. Bitcoin's 2024 halving was scheduled at block 840,000 and Zcash's at blocks 1,046,400, 2,726,400 and 4,406,400, and the date those blocks arrive depends on realised block times, which vary with hashrate, so published estimates drift. Close to the event a block-height estimate can tighten: Bitcoin's 2028 halving currently projects to 11 to 13 April 2028, and even that spread is pure block-timing noise before any hashrate drift. Where the trigger is cumulative issuance rather than block height, as on Bittensor, the estimate is only good to a season.
Does every halving actually halve issuance?
Not necessarily. Zcash's December 2019 Blossom upgrade halved the block reward and simultaneously cut block time from 150 seconds to 75 seconds, which doubled the number of blocks and left the issuance rate unchanged. ZIP 208 frames it as a proportional re-denomination that did not advance the halving counter, so it counts as a technical halving rather than an official one. The number to check is tokens issued per unit of time, not tokens issued per block.
Why do markets react before a halving rather than on the day?
Because the schedule is public and deterministic, so there is nothing to discover on the day. That is the same reason Tokenomist's 2024 Arbitrum work modelled a large ARB unlock against the Bitcoin halving date in advance: when both the supply cut and the supply release are known well ahead, positioning happens across the run-up rather than at the event. It is worth separating that pricing behaviour from the popular claim that halvings cause bull runs, which is a historical pattern across a small number of events rather than a mechanism.
Related Terms
Track on Tokenomist
Supply-side analysis for educational purposes. Not financial advice. Verify assumption and precision labels on the relevant token page.
Tokenomist.ai provides a complete solution for supply-side tokenomics data. Analyze future token emissions, track vesting schedules, and compare standardized tokenomics and allocation across projects to gain actionable insights