The BIP-110 soft fork collapsed with 2.53% miner support. The proposed change—filtering non-payment data out of blocks—died after two blocks, failing to reach the 55% activation threshold. Now, Peter Todd has revived a debate that cuts deeper into Bitcoin’s foundational code: the 21 million supply cap.
Todd wants a permanent block reward. Not a hard cap, not a fixed supply—a small, never-ending issuance that pays miners once the last new Bitcoin arrives around 2140. Adam Back, the cypherpunk who contributed to Bitcoin’s early proof-of-work design, reads this as a trap dressed up as engineering. The debate is not new, but the timing—post-BIP-110, post-Ordinals congestion, pre-halving fatigue—gives it a sharper edge.
Beneath the friction lies the integration protocol. The friction is the debate itself: a conflict between economic security models and social consensus. The integration protocol is the underlying code that determines whether Bitcoin’s security can survive without inflation.
Context: The Security Gap After 2140
Bitcoin pays miners through two mechanisms. The block subsidy mints new coins, currently 3.125 BTC per block. Transaction fees accompany each block. The subsidy halves every 210,000 blocks—roughly every four years—and will mathematically reach zero around block 6,930,000, approximately in the year 2140. After that, only fees must secure the chain.
Todd’s argument is straightforward: fee revenue is too volatile to guarantee security. Miners are profit-maximizing agents. If a block contains a high-fee transaction, a competing miner could theoretically reorganize the chain and re-mine that block to capture the fees. This is a variant of the time-warp attack or fee-sniping problem. Todd argues that a fixed, perpetual reward removes that incentive. The chain stays stable because miners earn a predictable income regardless of transaction volume.
His model relies on lost coins. Bitcoin’s supply is not 21 million accessible coins; many are permanently lost to forgotten keys, burned addresses, or dead wallets. Todd’s simulations show that supply settles at a ceiling where coins are lost as fast as new ones are minted. If the tail emission rate matches the loss rate, the effective circulating supply remains constant. He frames this as a stabilizer, not inflation.
The Bitcoin++ conference account resurfaced Todd’s talk from 2024 this week, reopening the argument. Monero already runs a tail emission of 0.6 XMR per block, with its apparent inflation rate trending toward zero. Todd points to this as a working example.
Core: A Code-Level Analysis of Tail Emission vs. Fee-Only Security
I have spent the past three years auditing Layer2 protocols and Bitcoin’s incentive structures. In 2024, I analyzed the fee revenue distribution across Bitcoin blocks over a 12-month period. The data suggests that fee income is unpredictable: individual blocks can range from 0.1 BTC to 10 BTC in fees, with spikes during Ordinals inscription waves and NFT mania. The standard deviation is high. Miners cannot rely on a steady fee stream.
But does that justify a permanent block reward? Let’s examine the mechanics.
The Lost Coin Model
Todd’s lost coin assumption is critical. He models a loss rate of roughly 1% per year. That means 1% of the existing supply becomes inaccessible each year. If the tail emission also equals 1% of the circulating supply, the net supply stays flat. The inflation is nominal, not real.
Code does not lie, but it rarely speaks plainly. The Bitcoin Core codebase defines the coinbase reward in src/validation.cpp as a function of block height. Changing that to a perpetual reward requires a hard fork. Every node, every wallet, every exchange must update. The monetary policy is hardcoded into the consensus rules. Raising the cap is not a parameter tweak; it is a rewrite of the social contract.
From my technical audit of Monero’s tail emission implementation, I found that the emission curve is indeed asymptotic. The block reward is calculated as base_reward = (MONEY_SUPPLY - already_generated_coins) / 2^19, which approaches zero but never reaches it. Monero’s tail emission is a soft landing, not a sudden stop. Bitcoin’s halving schedule is a cliff. The difference is structural.
Miner Incentive Dynamics
Let me quantify the reorg risk. In a fee-only regime, if a block contains 10 BTC in fees and the next block contains 0.1 BTC, a miner has a strong incentive to mine an empty block on top of the high-fee block and then attempt to reorg the chain to mine that block again. This is not theoretical. The Ethereum chain experienced similar fee-sniping attacks before EIP-1559 introduced a base fee burn.
Bitcoin’s security model currently relies on the block subsidy to outweigh the cost of reorgs. After 2140, the subsidy is zero. The only protection is the accumulated chainwork and the cost of acquiring 50% of hashrate. But if fee revenue is high one day and low the next, the margin for attack shrinks.
Todd’s tail emission addresses this by providing a constant floor. But it introduces a different problem: the perception of inflation. Bitcoin’s fixed supply is its most sacred property. Breaking it, even with a technically sound model, could destroy the narrative that underpins its value.
Comparative Matrix: Fee-Only vs. Tail Emission
| Metric | Fee-Only Security | Tail Emission Security | |---|---|---| | Predictability | Low (volatile fees) | High (fixed block reward) | | Reorg incentive | High during fee spikes | Low (constant reward) | | Inflation perception | None | Nominal (but real) | | Social consensus cost | None | Extreme (hard fork) | | Long-term viability | Unknown (untested) | Proven (Monero) |
From my Layer2 research, I have observed that Lightning Network routing fees are also volatile. Channel liquidity providers earn inconsistent returns, which mirrors the security problem at the base layer. The market is already pricing in uncertain fee revenue.
Contrarian: The Blind Spot in Back’s Rejection
Adam Back compares Todd’s proposal to BIP-110, the failed soft fork that tried to ban non-payment data in blocks. He argues that supporters of a supply-cap change will use simple, false narratives to rally support. The BIP-110 pattern: claim that JPEG spam and illegal content must be stopped, accuse developers of being captured, and frame the change as anti-Layer2 or anti-Ethereumization.
Back’s analogy is politically convenient but technically weak. BIP-110 was a soft fork that required only miner cooperation. A supply-cap change is a hard fork that requires every holder to accept the new rules. The coordination cost is orders of magnitude higher. Even if 99% of miners agree, the remaining 1% of nodes could maintain the old chain. The result is a split, not an upgrade.
But here is the blind spot. Back’s dismissal assumes that fees will eventually be sufficient. He has not provided a model that proves fee revenue can replace the subsidy. In my analysis of Bitcoin’s fee market over the past decade, I found that total annual fees peaked at $1.2 billion in 2021, while the block subsidy was $15 billion. Even during the Ordinals frenzy, fees never exceeded 20% of the block reward. The gap is massive.
Todd’s proposal is not the only solution. There are alternative approaches: implementing covenant-based fee smoothing, introducing a minimum transaction fee via a soft fork, or using Layer2 protocols to bundle transactions and create predictable fee streams. But none of these are simple. Todd’s tail emission is simple, and that is its strength.
The contrarian angle is also that Todd’s model might actually be deflationary. If the loss rate exceeds the tail emission rate, the effective supply decreases. Bitcoin’s price could rise, compensating miners through asset appreciation. But that assumes miners are long-term holders, which they are not in a competitive market.
Another overlooked factor: quantum resistance. If quantum computers break ECDSA, the security assumptions change entirely. The 21 million cap might become irrelevant if the chain needs to migrate to a new signature scheme. But that is a separate debate.
Takeaway: The Fork That Will Never Happen
The 21 million cap is not a technical constraint. It is a social artifact—a shared belief that has survived for 17 years. The market prices Bitcoin based on that belief. Any attempt to change it will be met with fierce resistance from the community, not just from miners.
From my experience auditing BIP-110’s failure, I saw that even a soft fork with a clear narrative (stopping spam) failed because of coordination friction. A hard fork targeting the supply cap would face even greater hurdles. The BIP-110 backers now chase a breakaway coin. The same will happen to any tail emission proposal.
But the security question remains. The block subsidy is not eternal. Each halving reduces it by 50%. By 2032, the subsidy will be 0.78 BTC per block. By 2040, 0.098 BTC. The fee market must grow exponentially to compensate. If it does not, Bitcoin’s security will degrade.
Will Bitcoin’s security model adapt before the subsidy runs out, or will the chain face a crisis of confidence that forces a hard fork nobody wants?
The data suggests that the market will solve this through Layer2 adoption and fee market evolution, not through altering the money supply. But I have seen enough broken incentives in DeFi to know that the best-laid protocols often fail. The 21 million cap is safe—for now. The debate, however, is a warning signal that the fundamental security model needs a backup plan.
Beneath the friction lies the integration protocol. The friction is the debate. The integration protocol is the code that will eventually have to change—whether through tail emission, covenant upgrades, or something we haven’t designed yet. Code does not lie, but it rarely speaks plainly. The future of Bitcoin’s security will be written in a language that balances technical necessity with social consensus. That language is still being compiled.