One of the most common questions people ask about Bitcoin is: what is to stop someone from creating a better version of it?
If Bitcoin is ultimately valuable because people choose to value and use it, why could someone not take everything we’ve learned from Bitcoin’s history, fix its weaknesses, improve the technology and launch “Bitcoin 2.0”? And if artificial intelligence makes it possible to design much better software, much faster, why could an AI not eventually create something so superior that people simply migrate?
It’s a good question because it forces us to think about what Bitcoin actually is, how money emerges and what role network effects play in all of this. And ultimately, it forces us to distinguish two very different things: creating better software and creating better money.
There is nothing in economics that makes Bitcoin literally impossible to replace. If humanity discovered a genuinely superior monetary good and enough people voluntarily converged on it, Bitcoin could lose its monetary role. But that is very different from saying that someone can write a technically superior cryptocurrency and expect the world to switch.
The mistake most people make when they follow this line of reasoning is treating Bitcoin like an ordinary technology product whose competitiveness can be assessed by comparing features. Bitcoin isn’t just software. It’s an emerging monetary network. To displace it, a challenger would have to do much more than improve the code. It would have to overcome the credibility, security, liquidity, infrastructure and shared expectations that have emerged around Bitcoin and give people a sufficiently strong reason to coordinate around something else.
Let’s unpack this in more detail.
Money Is Not Just a Technology. It’s a Coordination Game.
“Money has not been generated by law. In its origin it is a social, and not a state institution.”
— Carl Menger, On the Origins of Money
When people first encounter Bitcoin, it is natural to think of it primarily as a technology. Technologies compete on features, performance and user experience, and history is full of dominant products being replaced by better ones. Nokia and BlackBerry gave way to the iPhone. Blockbuster gave way to Netflix. If Bitcoin were simply another technology product, the question would largely answer itself: eventually something better should come along and replace it.
But Bitcoin is not just a technology. It is a monetary network, and monetary networks behave differently because part of their usefulness depends on other people choosing the same network.
An informed skeptic might respond: “But MySpace also had network effects, and Facebook still replaced it. Why couldn’t the same thing happen to Bitcoin?”
While this is true, it’s a misleading analogy, because not all network effects work in the same way. A useful distinction is between networks with low switching costs and coordination standards.
Low-switching-cost networks are networks where using multiple alternatives at the same time is cheap and moving between them is largely an individual decision. The network also does not contain significant stored economic value that must leave one system in order to enter another.
MySpace falls into this category. A teenager using MySpace could open Facebook in another browser tab, create a profile there, post on both platforms and gradually spend more time on whichever network their friends were using. They did not need to make an immediate all-or-nothing decision, and they lost relatively little by experimenting with the alternative.
That doesn’t work for storing value. If you have $1 of purchasing power that you want to store, you can’t simultaneously store it in gold and Bitcoin and then wait to see which one everyone else moves to in a few years. You can split your savings between them, but each dollar you put into one is a dollar you haven’t put into the other. You have to make that choice before you know which asset other people will prefer.
That makes expectations about what other people will choose much more important. You don’t just care about which monetary asset you personally prefer, but about which one you expect other people to prefer and, recursively, which one you expect other people to expect everyone else to prefer.
This brings us to coordination standards.
For a coordination standard, much of the usefulness comes from people converging on the same standard. The underlying features still matter — they are often what cause a standard to emerge in the first place — but once coordination has formed around one option, being marginally better is no longer enough. A challenger also has to overcome the value created by everyone already coordinating around the incumbent.
Language is a good example. You could design a language with more consistent grammar, simpler spelling and fewer irregularities than English. Those could be genuine improvements. But someone choosing which language to learn also cares enormously about whom they will be able to speak to, what books and research they can read, what films and music they can understand and what opportunities the language gives them access to.
Esperanto was deliberately engineered as a simpler international language, with highly regular grammar, few exceptions and relative ease of learning, yet almost 140 years later it still hasn’t been widely adopted. Not because its design lacks advantages, but because it couldn’t overcome the coordination standard of established languages such as English.
Money has this property even more strongly because people are coordinating not merely over communication today, but over purchasing power across time.
Now someone might reasonably say: “I can understand why an established language like English is difficult to replace. But Bitcoin is nowhere near as established today as English was when Esperanto appeared. How do we know that Bitcoin has already become enough of a coordination point that it will not simply be replaced by Bitcoin 2.0?”
The answer is that Bitcoin does not have to be as established as English for the analogy to hold. Coordination standards do not suddenly become difficult to replace only once everyone uses them. Their network effects strengthen gradually as more people, capital, infrastructure, security and expectations converge around them.
Bitcoin is clearly still in the process of monetizing, but it is also no longer a blank-slate experiment competing against a field of equally unproven alternatives. Over time, substantial liquidity, holders, miners, nodes, developers, exchanges, custody infrastructure and financial markets have accumulated around the same network. More importantly, Bitcoin has become the obvious reference point for the idea of scarce, non-sovereign digital money.
That creates path dependence, meaning that someone evaluating a new cryptocurrency today is not making the same choice someone made when Bitcoin first appeared. Instead, they are choosing between an existing monetary network with years of accumulated history, capital and credibility, and a new network that has to bootstrap those things from scratch.
As a result, if I am considering moving my savings from Bitcoin into a new asset, it is not enough for me to believe that the new asset is technically better. I also need to believe that enough other people will reach the same conclusion, that markets and infrastructure will follow them, and that the new network will remain the preferred choice rather than being replaced again by the next improvement.
A new cryptocurrency therefore cannot beat Bitcoin simply by being marginally better. Its advantage has to be large enough to overcome the network that has already formed around Bitcoin.
But this raises an even deeper question: which of Bitcoin’s important properties can actually be copied, and which of them have to emerge over time?
Credible Scarcity Has to Be Earned
One of the easiest things for a Bitcoin competitor to copy is its supply schedule. Anyone can launch a cryptocurrency tomorrow with a maximum supply of 21 million coins.
But writing a supply cap into code is not the same thing as creating credible scarcity. The economically important question is: why should anyone believe that the cap will still hold decades from now, particularly if changing it ever becomes tempting?
Merely writing the supply cap in the source code isn’t what makes Bitcoin’s 21 million limit credible. What makes it credible is that there are tens of thousands of nodes all over the world independently enforcing the supply cap. The incentive structure of the network itself also creates credibility. Increasing Bitcoin’s supply cap would require people and institutions who hold Bitcoin precisely because of the 21 million cap to devalue their own holdings voluntarily. In other words, in Bitcoin, changing the rules requires convincing the very people it would hurt most to hurt themselves.
Most importantly, this resistance to change isn’t purely theoretical. Bitcoin has already gone through major conflicts over its rules. During the 2017 Blocksize War, a powerful coalition of miners (representing more than 80% of Bitcoin’s hashing power) and prominent companies supported a change to Bitcoin’s consensus rules, but was ultimately unable to impose that version of Bitcoin on the wider network. What counted as Bitcoin still depended on what users, exchanges, businesses, miners and node operators were actually willing to recognize and enforce.
That is the difference between declared scarcity and credible scarcity. Anyone can put “21 million” into a new protocol. What cannot be copied on day one is the accumulated evidence that the network will keep defending that rule even when powerful participants want it changed.
Bitcoin Could Only Be Launched Once
There is another part of Bitcoin that a competitor cannot simply recreate, which is the conditions under which it came into existence.
When Bitcoin was launched in 2009, it was experimental open-source software, there was no established market price, and early participants acquired bitcoin through mining. At the time, people did not yet know what they were competing to own.
Bitcoin then accumulated users, miners, developers and value gradually, before governments, regulators or major financial institutions understood its significance or took it seriously.
Importantly, in its early years Bitcoin was far more vulnerable than it is today because it had not yet accumulated the decentralization, security, liquidity and geographic dispersion it has now. But by the time it became valuable enough to attract serious attempts at control, suppression or capture, the network had already become unstoppable.
Those conditions can never exist again. Simply having knowledge of Bitcoin makes recreating Bitcoin’s initial conditions impossible.
This is the deeper idea behind what Bitcoiners sometimes call Bitcoin’s “immaculate conception.” Bitcoin had to become decentralized before the world understood the value of controlling it. It had to distribute ownership, mining and infrastructure while the potential prize was still largely invisible. By the time that prize became obvious, ownership, mining, development and infrastructure had already become widely dispersed, making the network increasingly difficult for any single actor or coalition to capture or control.
Satoshi’s disappearance completed that process. Bitcoin had a creator, but its creator ultimately stepped away, leaving no founder who could claim permanent authority over the network or speak on its behalf.
That sequence is what makes Bitcoin’s emergence so difficult to reproduce. A new network can copy the launch mechanics — proof of work, no premine, open-source code, even an anonymous founder — but it cannot recreate a world that does not yet know what Bitcoin taught us.
Bitcoin was able to become resilient before the world fully understood what it might become. Its success permanently closed the historical window that made its own emergence possible.
Security Has to Emerge
Bitcoin’s network effects extend beyond holders, liquidity and infrastructure. The economic value stored in Bitcoin also helps secure the network itself, and that security has its own network effect, creating a reinforcing loop.
As more economic value accumulates in Bitcoin, the rewards to miners for securing the blockchain become more valuable, which supports greater investment in mining and increases the cost of attacking the network. Greater security, in turn, makes Bitcoin more credible as an asset in which to store value, which further establishes it as the coordination standard.
A competitor can copy proof of work, but it cannot copy this economic security. It first has to make its own token valuable enough for miners to commit comparable resources to securing it. Miners follow economic incentives, and those incentives are strongest where the block rewards and fees they earn are most valuable.
This creates another powerful feedback loop: monetary demand finances security, security strengthens monetary credibility, and that credibility supports further monetary demand. The stronger this loop becomes around one network, the harder it becomes for a challenger to bootstrap an equivalent one from scratch.
That creates a difficult problem for any “Bitcoin 2.0” because it needs substantial monetary value to finance strong security, while strong security is itself one of the things it needs in order to attract monetary value.
Bitcoin has already crossed that threshold.
Better Technology Does Not Mean Better Money
So far, we have treated the word “better” as if its meaning were obvious. But what does “better” actually mean in this context?
Does it mean faster transactions? More privacy? Or maybe more programmability? What about lower energy use? While all of these may be desirable properties, improving one dimension doesn’t necessarily improve the monetary system as a whole.
Every change involves a trade-off, which makes engineering distributed systems so complex. Increasing throughput can raise the cost of independently verifying the network and hurt decentralization. Adding functionality can increase complexity and attack surface. Making governance more flexible can make a system easier to upgrade while compromising the immutability of the key rules.
That last point is immensely important for money. Someone storing wealth for decades will likely care more about knowing that the rules remain difficult to change than about having the most feature-rich protocol.
The relevant question, then, isn’t whether a competitor can beat Bitcoin on any particular benchmark, but whether it can materially improve the properties people actually care about in money without weakening others that matter just as much.
Additionally, a feature that genuinely improves a monetary property doesn’t necessarily require a new monetary asset. Bitcoin can evolve, meaning that improvements can be incorporated into the protocol itself, built on higher layers or implemented through better infrastructure around the base layer.
This creates an important asymmetry in favor of Bitcoin. Bitcoin may be able to reproduce and incorporate a competitor’s technological innovations, while Bitcoin’s accumulated credibility, history and network can’t simply be reproduced by the competitor.
A challenger therefore needs much more than just a clever feature. It needs an improvement that matters enough, can’t be adequately absorbed by Bitcoin, and actually makes the alternative a better money overall.
But suppose a challenger clears even that hurdle. Suppose it really is a better money overall, in ways Bitcoin cannot adequately absorb.
Would that be enough to replace Bitcoin?
Better Isn’t Enough
Imagine, purely for illustration, that Bitcoin is 97 out of 100 as a monetary system and a new cryptocurrency is 98. That one-point advantage does not mean people should immediately move their savings.
The reason is that changing money is itself a coordination problem. Moving early only makes sense if you believe enough other people will eventually move with you. Otherwise, you may have exchanged an established monetary asset for a technically superior one that never develops sufficient liquidity, security or recognition to become money at scale.
There is likely a replacement threshold that must be overcome. A challenger has to offer enough additional value not only to outweigh Bitcoin’s existing liquidity, infrastructure, security, historical credibility and network effects, but also to compensate people for the uncertainty of leaving an established coordination point for one that does not yet exist.
That burden is another asymmetry in favor of Bitcoin. Bitcoin does not need to remain the best imaginable form of money forever. It only needs to remain good enough that no alternative becomes sufficiently better to justify the risk and coordination cost of moving away from it.
What Is “Better by Enough”?
“You never change things by fighting the existing reality. To change something, build a new model that makes the existing model obsolete.”
— R. Buckminster Fuller
We now know that being marginally better money may not be enough for a new monetary asset to replace an established standard. The new money has to be better by enough to overcome the network effects, credibility and switching costs of the existing network.
But monetary standards have changed before, so the more interesting question becomes: what does a sufficiently large monetary improvement actually look like? And, closely related, what could make something better by enough relative to Bitcoin?
The best way to understand and answer both questions is to look at gold.
Over centuries, humans converged on gold as the monetary standard because it had the best combination of monetary properties of all physical goods. But gold has a lot of shortcomings. The most important limitations of gold are that it can’t be moved across the world at the speed of information, it’s cumbersome to divide for small transactions, expensive to verify at scale and difficult to store securely without eventually relying on vaults and intermediaries.
The key point here is that all of these shortcomings are tied to gold’s physical nature, which is what made gold so difficult to improve upon with another physical good.
Even if we found a metal that was rarer than gold, easier to divide or superior along some other dimension, it would still be operating inside the same technological constraint tied to physical existence. This metal would therefore only be marginally better than gold, without being better by enough to overcome the network effects and thousands of years of accumulated monetary history around gold.
The only thing that could overcome gold’s shortcomings therefore had to be something of a nonphysical nature — something immaterial.
This is where Bitcoin comes in. Bitcoin competes with gold not by being a better physical good, but by being something fundamentally different: digital rather than physical.
Bitcoin introduced scarcity that could, for the first time, exist in a digitally native form. Its nonphysical nature is what allows Bitcoin to overcome many of gold’s shortcomings: it can be sent across the world without physically moving it, divided into tiny units without melting anything down, independently verified without an assay office and held without depending on a bank or vault. At the same time, its supply can be governed by rules that anyone can verify for themselves.
This is a fundamentally different kind of improvement, and I think it gives us the right way to think about what it would take to replace Bitcoin.
If a new cryptocurrency improves on Bitcoin in some way, whether through faster transactions, better privacy or something else, that improvement still happens within the same paradigm of digitally scarce, decentralized money. That is like trying to improve on gold with another metal — you stay within the paradigm of physically scarce, durable money.
To overcome Bitcoin’s accumulated network effects, credibility, security and path dependence, I would expect a successor to require something much more profound than another cryptocurrency with better specifications.
It would likely have to escape some fundamental constraint of Bitcoin in the same way Bitcoin escaped the physical constraints of gold.
The problem is that we can’t know what that would look like, just as someone 500 years ago couldn’t have imagined a digitally scarce, decentralized money that could be sent over the internet to anyone in the world. That person wouldn’t even have a concept of what “digital” means.
That does not mean we should assume our own imagination marks the boundary of what is technologically possible. Something may eventually emerge that changes our understanding of money as radically as Bitcoin changed the constraints of physical monetary goods.
But if that happens, its successor will probably look less like Bitcoin 2.0 and more like Bitcoin looked from the perspective of gold: something that changes our understanding of what money can be.
How Much Better Is Left?
There is another reason the jump from Bitcoin to a hypothetical successor may be fundamentally different from the jump from gold to Bitcoin, which is that on several important monetary properties, Bitcoin may already be close to the theoretical limit.
Scarcity is the clearest example. Gold is scarce, but its above-ground supply still grows by roughly 1.5–2% per year and can respond to higher prices. Bitcoin’s terminal supply is fixed at 21 million and can’t respond to higher demand at all. Once scarcity is absolute, there is no meaningful category of “more scarce” beyond it. A cryptocurrency with ten million units instead of twenty-one million would not be twice as scarce because, as long as the units are divisible, you would simply be changing the denomination, not the scarcity.
Divisibility is also close to a ceiling. At the base layer, one bitcoin can already be divided into 100 million satoshis, which is more than sufficient for today’s commerce. And if Bitcoin’s relative value ever reaches a point where divisibility into satoshis is insufficient, finer denominations can be implemented on higher layers.
Portability is another one. Using the Lightning Network, Bitcoin can already be sent around the world in seconds or less. So, at the level of information, Bitcoin can already move globally at essentially the physical speed limit of communication. Maybe user experience could get faster, or settlement architecture could be improved, but there is no future monetary technology that can send information faster than the laws of physics allow. Any remaining improvements in portability are therefore incremental rather than categorical and could potentially be implemented in Bitcoin.
Verifiability is similar. The monetary ideal is not having to trust someone else to tell you whether the money is valid or whether its supply rules are being followed. Bitcoin already allows anyone running a full node to independently verify those rules. Verifiability could become cheaper, easier or faster, but that would mostly reduce the cost of obtaining the property rather than create a categorical improvement in how independently the rules can be verified.
That doesn’t mean Bitcoin is perfect or that every dimension of money has been exhausted. Privacy, robustness, security and other properties may still have meaningful room for improvement.
But it does mean that a successor faces a problem that Bitcoin didn’t face: some of Bitcoin’s most important monetary breakthroughs are not merely good implementations of a property but may represent the endpoint of what that property can logically become.
AI Doesn’t Solve the Hard Part
This brings us back to the question we started with: what if artificial intelligence eventually becomes capable of designing cryptocurrencies far more sophisticated than anything humans could build today?
AI may discover better cryptography, write better software, identify weaknesses humans missed and make it almost trivial to create technically impressive monetary networks. But that does not mean it can automatically create substantially better money.
First of all, as we just saw, some of Bitcoin’s most important monetary properties may already be close to their theoretical limits. AI can’t invent something more scarce than absolute scarcity, and there is only so much room to improve an asset that can already be divided almost infinitely, transferred globally at the speed of digital communication and independently verified by anyone running a full node.
And where meaningful technological improvements do remain, they do not necessarily require a new monetary asset. If AI discovers a better cryptographic technique, scaling method, privacy technology or networking improvement, Bitcoin can potentially incorporate that innovation itself or build it on higher layers.
This creates the same asymmetry we saw earlier. A competitor’s technological innovations may be transferable to Bitcoin, while Bitcoin’s accumulated history, credibility and monetary network are not transferable to the competitor, even if the competitor is engineered by AI.
And even if AI does design a genuinely superior monetary protocol that can’t be absorbed by Bitcoin, it still hasn’t solved the hardest part. AI can write a new supply schedule, but it cannot instantly create decades of credibility behind it. It can design a secure consensus mechanism, but it cannot manufacture the economic value that pays for that security. It can create an excellent protocol, but it cannot instantly create its liquidity, infrastructure, ownership distribution or the shared expectation that millions of other people will continue to recognize the same monetary asset.
In fact, if AI makes technically sophisticated cryptocurrencies abundant, that may make Bitcoin’s distinction stronger rather than weaker. The easier it becomes to produce clever code and attractive features, the more obvious it becomes that those were never the hardest parts to reproduce in the first place.
This doesn’t mean that AI could never contribute to something that eventually replaces Bitcoin. If it discovers a genuinely new monetary paradigm that is to Bitcoin what Bitcoin was to gold, then the argument changes.
But simply designing a technically superior cryptocurrency would not solve the hard part.
AI may make creating better software dramatically easier. It does not make creating better money dramatically easier.