Perpetual inflation won’t save your blockchain

Key takeaways

• A lot of blockchains are using perpetual inflation, often called tail emission, to attempt to guarantee miner/validator income in the absence of a robust fee market

• The problem is that all such inflation is denominated in the token native to the blockchain in question, and its future purchasing power cannot be known let alone guaranteed   

• As a result, protocol security in censorship resistant blockchains is a market outcome not an engineering problem and therefore has no guaranteed solution—either there is future demand for security or there is not, and this cannot be known

• Since the future token purchasing power cannot be known, no amount of tinkering with the monetary policy is ever guaranteed to sufficiently secure a blockchain—instead, tinkering erodes the monetary properties of the token, risking a downward security death spiral as tokens are unable to compete for monetary demand and the mining/validating rewards lack sufficient value to secure the chain 

• In short, tail emission is not, and cannot be, a guarantee of sustainable settlement assurance

• On the contrary, I argue that constantly revising protocol monetary policy presents a bigger risk to long term security than simply letting the market tell us whether it wants something or not

A couple weeks back on 12 August 2025, Monero, PoW blockchain underwent a 6-block re-org. Over several days after, it then suffered several 9-block re-orgs.

While this only represents a 12-18 minute history re-write given Monero’s 2 minute block time target (the time-equivalent of approximately 2 Bitcoin blocks), it is still way outside the range of standard non-malicious re-orgs that happen from time to time on PoW blockchains. Statistically, natural re-orgs past two are highly unlikely, and the probability of occurrence decays exponentially with each additional block. 

It is therefore quite likely that the 6-block re-org was an intentional act on the part of Qubic—a strange blockchain acting as a de facto Monero mining pool. And they have publicly claimed that it was. The attack has been widely represented in the media as a 51% attack, however, more sophisticated analyses have not found any convincing proof of this assertion. 

In all likelihood, the re-orgs were instances of a “selfish mining” strategy—a long-known method for maximising mining revenue by miners who control around 33% of hashrate within PoW blockchains. PoS blockchains are also vulnerable to such attacks, often at even lower levels of total stake, although PoS complexity often makes it a less straightforward matter to analyse.

This is not, however, a post about selfish mining. Nor is it a post about Qubic’s peculiar incentive structure which used token markets to, at least temporarily, improve profitability for Monero miners pointing hashrate towards them, rapidly growing their hashrate. Instead I’d like to discuss a less obvious, but in my opinion, much more important takeaway from the debacle: Namely, that tail inflation is not a viable solution to long term protocol security in blockchains.

Let’s first agree on some definitions

One of the reasons discussions on this topic (and most other debates too if we’re honest) are so tedious is that people don’t agree on the meaning of words and terms and therefore end up talking right past each other. In order to avoid that I want to be ultra clear about what I mean when I refer to certain words and terms.

1. Protocol Security — Few terms get thrown around as loosely as this one. It is in turn often reduced to “cost to attack”, which is a loose unspecified reference to re-org cost. When we use the term, we will use it under the strict meaning of settlement assurance as laid out in detail by Nic Carter in 2019

2. Censorship Resistance — The ability for a blockchain to unseat a malicious miner who has achieved 51% or more hashrate contribution and is using it to censor blocks

3. Tail Emission — The use of infinite inflation as an incentive for block creation

4. Security Budget — This is a term I prefer to mostly avoid because it cannot be quantified in a meaningful way within the confines of a blockchain protocol. Therefore it is purely conceptually illustrative, carrying no concrete meaning

There seems to be a problem with Bitcoin’s long term security

Let me try to briefly summarise the apparent problem underlying all of this, which all stems from Bitcoin. One of the major assumptions underlying Bitcoin is that once the supply of new coins runs out, transaction fees will take over as the main source of payment for miners.

Technically speaking, this is already the case. Over the last year on average, if you removed the current block emission of 3.125 btc, approximately 0.06 btc of fees would remain (less than half of that if the average was 30 days). The problem looks obvious: Current fees are less than 1/100th the size of the block subsidy, that’s not very much.

In fact, if fees never improve from here in btc terms, they won’t match the size of the block reward until 6 halvings from now. That’s about 23 years from now. This means that for the mining reward to remain at least at its current level, the product of the bitcoin purchasing power and current total fees per block (btc_purchasing_power * total_btc_fees) needs to grow by a factor of 100 in 23 years.

A lot of people think this is unlikely, so there have been a lot of suggestions of how this can be “fixed”. There’s just one problem with that mindset.

Protocol security in censorship resistant blockchains is a market outcome not an engineering problem

I need to preface this segment with the clarification that when I talk about Censorship Resistant blockchains I am talking exclusively about PoW blockchains. PoS blockchains can per definition not resist censorship, a topic we have covered in depth in previous work. 

When I previously said that Bitcoin’s long term security/settlement assurance is based on an assumption I meant just that. It is not possible to prove that it will be “sufficient” over the long term. Either there is market demand for settlement assurance through transaction fees or there is not. It is almost exactly the same question as to whether there will be demand for bitcoin in the future or not. 

If the answer to either of those questions is “no”, Bitcoin will fail. I also can’t imagine it’s possible that the answer is “yes” to one, and “no” to the other, so the failure condition is almost exactly the same. If people don’t need bitcoin (the money) or Bitcoin (the protocol and network), it will fail. The argument is self-evident.

This reality does not tend to sit well with the engineering types of the world (I am one). As a result, many  “solutions” have been proposed. These “solutions” tend to fall into two categories. Either they aim to change the supply or demand side of the fee market, or they aim to use tail emission as a perpetual reward for block creation.

The problem is that neither of those approaches can provably work. No matter how much one tinkers with either the supply or demand side of the fee market, one cannot guarantee that there will be any demand at all, let alone “sufficient” demand, whatever "sufficient" means in this context. This is obvious enough to most people.

Still, just because it cannot be 'guaranteed' or 'provable', people still think there are theoretically better designs than at present, and that is why these things are proposed. That’s a fair enough position to have, I just cannot stress enough the fact that thinking they will work is as much of a leap of faith as thinking the fee model will work. The truth is that nobody knows if either will work.

A tail emission “security budget” can never be proven sufficient 

It seems to be less obvious to a lot of people that tail emission has the same exact problem of not provably being a solution to incentivise perpetual settlement assurance. At what level is the tail inflation “sufficient”? Setting tail emission at 1% targeting some arbitrary “security budget” can never guarantee settlement assurance at any given level—it’s all pure guesswork.

Since you cannot know the future purchasing power of the token, you are always at risk of the emission being “too low”, necessitating endless changes in monetary policy, and in the extreme, runaway inflation. This further erodes the monetary properties of the token risking a security death spiral as plummeting token prices requires ever-higher inflation rates to finance the “security budget”.

Monero’s vulnerability should be a shot across the bow for tail emission blockchains

As mentioned above, the main point I want to drive home here is that tail emission is no guarantor of long-term blockchain security. Monero implemented tail emission in 2022, and among the community it was expected that it would guarantee sustainability of the block reward. While I suppose that is technically true, as we’ve all just seen, it has also been meaningless in terms of providing workable settlement assurances.

While there’s a guaranteed perpetual block reward available for Monero miners, Monero the token cannot compete with bitcoin on monetary properties so nobody stores any value in Monero. The outcome is clear as day, Monero purchasing power has been relatively stagnant over the last decade. Against its strongest competitor, bitcoin, the situation is much worse, with Monero plummeting in value over the years.

In other words, tail emission sounds fine and well, but if your token isn’t worth anything no amount of tail inflation will be enough.

This should be a wake up call for blockchain inflationists. It is widely accepted among bitcoiners at least that inflation is harmful to fiat money and society at large. I therefore find it remarkable that there are any bitcoiners at all that believe it wouldn’t be harmful to bitcoin.

The main driver of long-term value in any type of money comes from its low-velocity users looking for an instrument that holds the value of their savings over the long term. These users have a strong preference for a monetary unit they perceive as having the strongest possible monetary properties. If your blockchain doesn’t provide that, they might use it to transact with high velocity, but not to store their wealth while they’re economically inactive. That’s bad news for your token value.

We must resist the engineering temptation to tinker with Bitcoin’s monetary properties

In an earlier paragraph I mentioned that in order for Bitcoin to retain its current settlement assurance over the next 25 years or so it needs to grow the purchasing power of transaction fee rewards ( btc_purchasing_power * total_btc_fees) by about 100 times. I actually don’t think that’s unlikely at all.

Given that bitcoin price has grown by more than 100x over the last 10 years, and over the same time we’ve had several periods of sustained fee levels more than 20x higher than current ones in BTC terms. In other words, 10x growth in price coupled with a 10x growth in fees over the next two and a half decades doesn’t sound that improbable to me.

That is, I don’t think it’s unlikely so long as bitcoin retains its exceptional monetary properties. If we mess those up by changing the block size, adding infinite inflation, or falling into the general Ethereum mindset of constant and endless revisions to monetary policy, I think that’s a much bigger risk than just letting the market tell us whether there’s demand for bitcoin over the long term.

In fact, the very reason bitcoin is not already at $10m per coin is that we (as one big society and market) fundamentally cannot know whether the assumption of a sufficient future demand for bitcoin turns out to be a good one, or not. Figuring out the answer to that question and pricing the probability accurately as time goes on is exactly what markets are for. Let it do its thing.

 ¹This is a theoretical level that does not account for a whole range of confounding factors that create a more blurry range around the actually necessary real-world conditions.