Lock Reduction Interval
Lock reduction interval is a network-wide parameter, listed in the documentation as
LockReductionInterval, that sets how many blocks must pass for the network lock cost to halve. The
reference describes it as the number of blocks needed for the network lock cost to fall by half, set
at root permission
(Subnet Hyperparameters,
Create a Subnet).
What It Controls
The cost to create a subnet is dynamic. The documentation explains that this cost doubles every time a subnet is created and then lowers gradually over time, so the lock reduction interval is what paces how quickly that decay happens (Create a Subnet: Burn cost).
Specifically, it is the block span over which the elevated cost halves. A shorter interval makes the creation cost fall faster after each new subnet, while a longer interval keeps the cost elevated for longer before it eases back down (Subnet Hyperparameters, Create a Subnet).
Because the span is counted in blocks, the real-time pace depends on block time. The default interval is 14 times 7,200 blocks, which is 100,800 blocks or roughly 14 days, so the elevated cost halves over about two weeks before easing further toward the floor (Subnet Hyperparameters, Create a Subnet).
The underlying decay is linear, not compounding. On chain the creation cost is read as the last paid
lock doubled, minus a fixed amount per block equal to that last lock divided by the lock reduction
interval, multiplied by the blocks elapsed since the last creation. After one full interval the
doubled cost has therefore fallen by exactly the last lock — back to the previous value — and it
keeps descending at that same per-block rate until it reaches the minimum lock floor
(get_network_lock_cost).
How the Decay Bottoms Out
The decay does not continue to zero. As the creation cost falls across the lock reduction interval it approaches the minimum lock cost floor rather than dropping away entirely, so the interval governs the speed of the descent while the floor governs where it stops (Subnet Hyperparameters).
- Lock reduction interval — how fast the creation cost decays.
- Network min lock cost — the floor the decay approaches.
Why It Is Network-Wide
Subnet creation is a network-level action, so the lock reduction interval is a single network parameter rather than per-subnet chain state. It applies to the cost of registering any new subnet, which is why it is not keyed to an existing netuid the way per-subnet hyperparameters are (Create a Subnet, Bittensor Networks).
The current creation cost that this interval acts on can be checked through the
btcli subnets burn_cost command, which reports the live TAO required to register a new subnet at
that moment (Create a Subnet).
Documented Type and Setter
The Subnet Hyperparameters
reference lists LockReductionInterval as a u64 measured in blocks, changed through the
sudo_set_lock_reduction_interval extrinsic at root permission. Because it governs network-wide
subnet creation economics, it is set at the root level rather than by individual subnet owners
(Subnet Hyperparameters).
The constants reference gives the startup default as
api.consts.subtensorModule.initialNetworkLockReductionInterval, listed as 100800. That constant
names the initial two-week block span, while the live LockReductionInterval storage value is the
root-settable value the network uses after launch
(Subtensor Constants: initialNetworkLockReductionInterval,
Subnet Hyperparameters).
Distinction from Network Rate Limit
Network rate limit is the minimum block wait between subnet registrations, while lock reduction interval is the span over which the creation cost halves. One paces how often subnets can be created, the other paces how fast the cost falls (Subnet Hyperparameters).
- Lock reduction interval — how fast the creation cost decays.
- Network rate limit — minimum wait between new subnet registrations.