This post is also on our new dedicated blog, check it out:

We first started thinking about Caldera (dubbed Constellation at the time) when we saw that the web3 world was in a catch-22.

Blockchain developers — branching across various verticals like NFTs, DeFi, gaming — want to build complex applications. Yet, they’re plagued by the limitation of everyone building on a handful of overcrowded blockchains.

Starting out, we wanted to create rollups to enable expansive, truly scalable web3 games. But we quickly realized that the problem wasn’t isolated to just gaming — most dApps are built on L1’s.

Nowhere is this more apparent than with Ethereum, boasting a whopping 0.19 million instructions per second.

To put this in perspective, that’s roughly equivalent to the Altair 8800, a computer from 1974. Half a century later, compute speed has exploded exponentially. The blockchain is the world’s slowest cloud computing platform.

The lack of processing speed combined with high developer interest has led to extreme competition for computation and storage. For the end user, that means high gas fees and slower transactions.

Both issues are cramping the quality and user adoption of on-chain applications.

The Building Blocks of A Useful Blockchain

The modern computer was built atop thousands of smaller innovations. From the transistor to the silicon chip, to the disk operating system, each level of abstraction added new functionality and brought computing to new verticals.

And similar to the early days of the computer and video game programming, we’re still in the early days of the blockchain. Developers are designing applications around the blockchain’s limitations, much in the same way that we used to when we were bottle-necked by speed and memory.

For example, we’re seeing developers “gas golf” — spending weeks optimizing their smart contract code to minimize number of computations and gas fees. In the process, not only do the developers waste engineering time, gas golfing is often accomplished by cutting corners and unknowingly sacrificing security. These manual optimizations of code are painful and ultimately don’t improve the user experience.

There’s also a tradeoff between finality period and security. A shorter finality period means quicker transaction execution and more complete information for all participants. On the other hand, when we consider disputes or security, a longer finality period allows for more flexible operations. It doesn’t make sense to apply a one-size-fits-all finality period for applications with different purposes and operations.

These limitations haven’t completely stopped the development of innovative applications. But these applications have been hamstrung because they needed to work on an L1 such as Ethereum or Gnosis. Blockchains have already come a long way, but more infrastructure is needed to truly reach web-scale.

Introducing Caldera

Caldera’s mission is to empower developers to create new and exciting dApps and help web3 hit mass adoption.

Caldera is a rollup-as-a-service, making it incredibly easy to launch performant blockchain rollups. Developers are already using Caldera to launch lightweight, highly-customizable blockchains that inherit the security we know and love from other blockchains like Ethereum or Polygon.

We see parallelizing transactions as the most pragmatic way forward out of our current predicament. Just as it makes sense for web applications to run on dedicated servers, web3 applications need their own chain to secure the necessary throughput and features.

We’ve recently partnered with Curio, an on-chain gaming studio pioneering innovative web3 game mechanics like treaties - smart contracts that create smart diplomacy rules within the game. For example, players can create a non-aggression treaty that prevents attacks upon payment of ERC20 tokens to a shared guild smart contract.

Yes - much of this was impossible before rollups. But, sticking to our previous example, Curio consumes around 300 million gas per second. It would take maxed-out Ethereum blocks 120 seconds to process an equivalent amount of gas — so it’s safe to say that building Curio on an general-purpose L1 would have been impossible.

With Caldera, Curio has built a dedicated layer-2 transaction layer and brought latency to less than one second. There’s ample room for scaling, and just as importantly, protection against gas price spikes.

We’re already enabling blockchain developers build the apps that they want, without compromising on performance or features.

Beyond

We’re making it possible to create the same functionality and scale we use on the web today — on the blockchain. We’re excited to have you join us as we develop the next level of infra to take web3 to the scale, usability, and distribution of web2 — and eventually, hopefully, surpass it entirely.