Scalability problem: is the end of Bitcoin near?

Question

I was reading in Wikipedia that the blockchain can only handle around 7 transactions per second:

https://en.wikipedia.org/wiki/Bitcoin_scalability_problem

With the soft fork for SegWi, I can imagine it could handle a little bit more, maybe 10 transactions per second.

That would mean that in one day we could have:

60 second * 60 minutes * 24 hours * 10 transactions = 864000 transactions/day

Here the official chart showing that we are often getting close to 400000 transactions/day

https://www.blockchain.com/charts/n-transactions

Does that mean that when we double the number of transactions, Bitcoin cannot scale anymore? Am I missing something?

If that's the case, considered the recent bull crypto market, we can expect there is a high chance Bitcoin reaches the maximum number of transactions per day, which will cause the crash of the Bitcoin price and a big shift in the crypto world.

Answer 1

TL;DR:
No, the end is not near—we're just getting started. 20× is boring, let's figure out how to do 100,000×.

---

Let's first talk about what we are trying to achieve. We are not actually trying to increase just the number of transactions. What we're actually trying to do is to grow the network's utility: allow more people to use Bitcoin with better privacy, UX, features, and security.

While increasing the count of transactions would increase the utility, it would do so only linearly. Ten times the cost for ten times the utility—it's a bit trite. It also leads down a path that endangers other properties that we relish about Bitcoin: e.g. financial privacy, rules without rulers, censorship resistance, and the ability to validate the blockchain individually. So, instead, we aim to get leverage on the cost.

For example, stop thinking in terms of transactions, but rather think in terms of payments. By batching multiple payments into one transaction, the transaction count doesn't increase, but for a slight increase of the transaction weight, we can increase the payment count. You mentioned segwit which introduced new output formats that have less transaction weight for equivalent effect. The upcoming taproot soft fork will add another weight efficiency improvement in the same vein.

Networks scale in layers. We can use the base layer of Bitcoin as a foundation to build other services on top. These other layers can operate with different trade-offs and make use of the base layer as a "court" to resolve contracts and settle disputes. An example is the Lightning Network which facilitates instant payments and enables a user to perform manifold payments via a single base layer transaction to open their channel. Timestamping of events and documents has moved from one nulldata output per timestamp to off-chain data structures that anchor to a single output per batch. Sidechains like Liquid and Rootstock provide new scripting and privacy capabilities, but limit the on-chain cost by performing most transactions in their own blockchains.

By being restrictive in the use of the base layer, the network forces us to do the hard work to make efficient use of a scarce resource. It's also honest in setting expectations: we would never be able to put every imaginable payment on the base layer, so kicking the can down the road by doubling the payments a few times only delays the necessary work. However, after exploring other avenues, and squeezing the most out of a small base layer, we can always come back to talk about increasing the base layer capacity, eventually.

Answer 2

The answer to this question is not entirely clear, as nobody can predict the future to see how successful the bitcoin network will ultimately be, and scaling to meet the financial demands of the world is quite the task.

That said, there is an important point to understand before exploring this question further: the number of transactions that happen on-chain do not map 1:1 with users making payments. The word 'transaction' is maybe a bit of a misnomer here unfortunately, it might be more appropriate to call it a 'settlement', as a single on-chain transaction can account for and settle a potentially unlimited number of payments. I think this article does a good job of describing this in detail.

What this means is that there are a lot of different ways that a user can store and transact value that is backed by the bitcoin network, without actually sending their own on-chain transaction! In some cases, the user may not even need to send an on-chain transaction at all.

---

Scaling via 'off-chain' solutions

There are a number of known methods to scale the number of transactions that users can engage in (see Murch's answer here):

• The lightning network (and other payment channel techs)
• Lightning channel factories
• Federated sidechains
• Custodial off-chain services

The general idea of these solutions is that a single bitcoin transaction can be crafted in a way that allows users to exchange value without sending further bitcoin transactions. Each of these methods requires additional trust/security considerations, but many users are willing to make these trade-offs for increased utility/functionality.

There are some other ways to transfer bitcoins without sending an on-chain transaction, a few of which are detailed in this answer. But some of those solutions (eg the OpenDime) are interesting novelties, more than widely employed solutions.

---

Scaling via 'on-chain' solutions

There are also ways to scale the number of bitcoin transactions that the network can handle, to copy paste from Murch's answer that is linked above:

Examples:

• Blocksize/blockweight increase
• Faster blocks
• The witness discount of segregated witness
• Smaller size of Schnorr signatures
• Bellare-Neven signature aggregation
• Key aggregation

Additionally, batching payments can provide savings in both blockspace, and transaction fees.

Note that there are very important engineering considerations when attempting to scale the network by altering these on-chain parameters. As example, increasing the blocksize drastically can provide a linear increase in transaction throughput, but it also greatly increases the resources required to run a node on the network, which is damaging to the network's health. Without a large number of independent node operators, some of the network's most important properties (eg. censorship resistance) will be substantially degraded.

---

If that's the case, considered the recent bull crypto market, we can expect there is a high chance Bitcoin reaches the maximum number of transactions per day, which will cause the crash of the Bitcoin price and a big shift in the crypto world.

I don't think so, it seems that the current rhetoric of users (investors) is not "this is valuable because we can send a huge number of transactions!", but rather, "this is valuable because it is a scarce digital object with strong guarantees of censorship resistance and an incorruptible economic policy".

Of course, serving a larger audience of people is desirable, but just because the system hits its max throughput does not mean it will suddenly fail. In fact, such situations serve as great motivation for users to find more efficient ways to transact. As example, a recent period of network congestion lead to a great number of users turning their attention to the issue of poor fee estimation present in some wallets.

If nothing else, consider that even if bitcoin cannot scale to meet daily global transaction demands, it may still be an incredibly useful and valuable technology nonetheless! As analogy, consider that gold is extremely valuable, yet it is very rarely used in day-to-day transactions. (I don't mean to imply that this is the ideal future for bitcoin, I am hopeful it will continue to scale, but it is something to be considered in the extreme of OP's question "will it fail if it hits max transaction throughput?").