Many believe that general-purpose blockchains like Ethereum, Cardano, Avalanche and Solana will come to power everything on the web, including financial apps, social apps and even Amazon-like marketplaces. But there’s a show-stopping problem that’s being widely overlooked: on-chain storage.

While today’s general-purpose blockchains have worked well for storage-light applications like decentralized finance (DeFi), they cannot scale to handle storage-heavy applications like social apps and marketplaces. Imagine a world in which every “like” or follow on a decentralized app cost $1.00+ in storage fees.

Unfortunately, that reality is now because of the storage limitations of all general-purpose blockchains on the market today. As such, in order for Web 3 to reach its full potential to disrupt Web 2 and the systems of the past, new blockchain architectures will be required.

Nader Al-Naji is head of the DeSo Foundation and founder of social media platform BitClout.

From finite state to infinite state

Today, all general-purpose blockchains on the market were built to power what we call finite state applications. These are applications where the amount of data or state that you have to keep on-hand for each user is, well, finite. For example, in order to build a financial app, all you really need to know in order to validate transactions is each user’s account balance.

Users could transfer funds between each other millions of times, but at the end all you need to store is just a few numbers indicating the final balance of each user. Put another way, the state you have to keep around grows as a function of the number of users rather than as a function of the number of transactions.

Perhaps surprisingly, virtually all DeFi consists of finite state applications.

What if we want to look beyond finance? Infinite state applications are ones where the amount of data you need to store grows indefinitely with the number of actions that each user performs. For example, consider a typical social app: Users can create a profile, make a post, follow people and do other such things, all of which add state.

The difference is that, with social applications, all transactions are state-augmenting rather than state-neutral, as is the case with DeFi. With social, instead of just having to keep a few account balances in your state, you need to be able to store an indefinite amount of data. Even worse, this state needs to be frequently queried by other users on the network, requiring it to be highly available.

In order to handle the storage and indexing requirements inherent to infinite state applications, blockchains will need to be custom-tailored to the application at hand, such as DeSo for decentralized social. This is because, without being able to make assumptions about the type of data that will be stored, the costs of storing, indexing and querying the data will skyrocket, making applications built on the chain uncompetitive.

The cost of storing just 1 gigabyte of on-chain state varies significantly across blockchains. Importantly, these costs are only expected to increase for general-purpose blockchains because they weren’t designed to scale storage.

These high on-chain storage costs prevent the vast majority of Web 2 applications from being implementable on today’s general-purpose blockchains, even when using bridges to storage-focused blockchains like Arweave or Filecoin. At current prices, even storing a simple link to Arweave or Filecoin on a general-purpose chain costs $0.10-$1.00+, which is prohibitively expensive.

Moreover, even though many blockchains claim to be able to handle thousands of transactions per second (TPS), this metric does not take into account the storage properties of the application at hand. There is a big difference between 50,000 DeFi transactions, which may generate zero bytes of new state data, as opposed to 50,000 social transactions, which may generate tens of megabytes that need to be stored, indexed and queried.

Today’s most advanced blockchains fail completely at handling the latter type of transaction. This limitation is blocking the development of some of the most interesting Web 3 applications.

Read more: Web 3 Is Where the Young People Are

We at the DeSo Foundation been researching this challenge and have concluded that all storage-heavy Web 3 applications, such as social apps and marketplaces, will require new types of blockchains to develop because these applications are infinite state applications rather than finite state applications.

The difficulty of storing and indexing data in a scalable way is something that has been underestimated by most of the crypto space. For a long time, the entire space has been limited to finite state applications without much consideration for the wide range of infinite state applications, like social apps and marketplaces, that make up the majority of Web 2 applications.

In order for Web 3 to reach its full potential to disrupt Web 2 and the systems of the past, blockchains that are custom-built to support new use cases will be required because of the storage and indexing limitations inherent to the existing general-purpose chains.