← head back


Crypto's anti-network effect


Network effects

Much of this millenia's history has been marked by a single observation. Networks grow exponentially ie. Metcalfe's law. Not every industry benefits from network effects, but those who do enjoy unique economics and an extreme upside potential.

Crypto networks

Every crypto project is becoming a network in multiple ways. A blockchain natively needs an asset exchange, a naming system for human-friendly addresses and at least a bunch of protocols to support the ecosystem like liquid staking, lending platforms, NFT marketplaces, stablecoins and aggregators.

There's also a secondary effect where people joining an ecosystem early are disproportionally interested in its success. Yet, in general the bigger the ecosystem the more tokens are listed in exchanges, the more NFTs are listed in marketplaces and the more the people (in terms of addresses) you can interact with. That effect is arguably dominant in comparison to the early adopter effect mentioned before.

In short, one could expect "the bigger a network, the bigger it will become", even in crypto. But that's not the case.

The more popular a blockchain the higher the demand for blockspace. In other terms, the more popular the more expensive they get.

Dreams & pipe-dreams

There's tons of resources about blockchain scaling. Some is technical and inaccessible but from a 30,000 feet view they lie in the following 3 categories - sharding, layer 2s and optimizations (for either consensus or execution). Optimizations for different consensus algorithms are for example the Avalanche's Snowman consensus or Algorand's Pure PoS consensus. Optimizations for transactions execution are for example using Aptos's Narwhal and Tusk mempool or Fuel's UTXO-based language.

Generalized sharding is very network heavy at the moment and research teams are still sculpting out the details. Different optimizations only improve TPS by a small margin in practice which seems not enough for real network effect to take place. On the other hand, layer 2s seems to be the way forward, at least for now. There are already L2s in the wild that have million of TVL and thousands of users.

These layer 2 systems work but each works its own little world at the moment. Dapps and assets in one layer 2 are not available to the other layer 2s. There's a fragmentation problem. We can't have network effects simply by using layer 2. Layer 2 scale layer 1 TPS but not its network effects.

That's really bad right?

Yes and no. The bad news is that there's no silver bullet. There isn't a direct way of scaling layer 2 beyond the limits of a single blockchain. Besides, a layer 2 blockchain is a blockchain after all. The good news is that there are still ways to overcome sharding and layer 2 limitations and build better systems.

As sharding is far from ready and Ethereum is adopting a rollup centric future, I'll focus on layer 2 in this article and the fragmentation problem.

Off-chain computations

A layer 2 is another blockchain with an escape hatch to the layer 1. If something goes wrong on the layer 2 users can still use the (more expensive) layer 1. PS: L2s also have different systems so things don't go wrong often.

This new blockchain has the same limits as the layer 1. For example, Ethereum processes ~1m tx/day and its at max capacity. Layer 2 built for the same validator specs as Ethereum will process around the same number of transactions.

These systems are not scaling the blockchain network. Instead, they scale security in separate networks.

It's simple to suggest to just use another rollup. But that's almost like saying to use another blockchain (besides so more secure bridging options). Using another rollup means researching which one suits you best, it means setting it up in your wallet, it means transferring assets over, it means exploring different dapps and learn how to use them correctly.

Crypto has an anti-network effect.


A network's value grows exponentially to its size. But as blockchains grow so does their costs, even for layer 2. You can't just use another layer 2 because it's complicated and time-consuming.

The only way a multi rollup system could work is by mitigating the fragmentation hassle with cross rollup's systems. There are designs for cross rollup AMMs and Vitalik made a detailed post in ethresear.ch about cross rollup NFTs but there's not much about end to end cross rollup systems that will work seamlessly within a user's wallet, yet.

That's the topic I'll dig into in the second part of this article.

Related Posts

Looking to grow your new product?

👉 Check out my free Funnel Visualization tool

Andreas Tzionis @