The Rise of DEXs Begins Move to Layer-2s Like Starkware

The decentralized spot and futures trading app dYdX once required prospective traders to deposit $10,000 to get started. This was just to cover the gas fees involved in trading on the decentralized exchange. Unlike a centralized exchange (CEX), DEXs don’t live on servers. Instead, they exist as a series of smart contracts on a blockchain. 

While this means that there’s not the same overhead as a centralized exchange (consider that Coinbase employes close to 1000 people) for most of the past year each transaction comes at a higher cost given the complexity of the contract execution on Ethereum and the demand for blockspace. 

DEXs have been around for about three and a half years since the launch of the Bancor Network in April 2018, followed by the original version of Uniswap on November 2, 2018. In crypto-land that’s comparable to a decade. The launch of dYdX — which is typically written in like this, inspired by the derivative function in calculus — changed the adoption equation by providing trading infrastructure that’s quite usable and similar to a CEX, explains Dragonfly Capital’s Haseeb Qureshi. 

“We always believed it was a matter of time until decentralized perpetuals took off, and dYdX has really shown that the big thing standing in the way was an incentive for traders to onboard (to solve the cold-start liquidity problem) and just having a good enough fully-fledged trading experience that mimics CeFi,” Qureshi told Blockworks by email, explaining his firm’s investment thesis. 

But this is also a double-edged sword. The market enthusiasm for dYdX made it prohibitively expensive, with Ethereum gas fees shooting up into the stratosphere as the market crowded in to trade on its crowded pipes. 

Part of the problem comes from how congested the Ethereum network has become. The future of finance may well be on Ethereum, but as a result, its rails are currently a congested hot mess. 

Layering and Scaling

The market has been evolving to fix this problem with scaling solutions that move transactions and computation off of the mainnet. Alternative layer-1 protocols tend to take a more centralized approach to boost transaction speed or accommodate more users and data thus bringing down costs. Examples of this are protocols like Polygon or Solana.

Another layer-1 method is sharding where the blockchain database is broken into smaller pieces, and thus makes it easier to process — the approach that Ethereum will take with its forthcoming 2.0 version next year. 

In contrast, layer-2 solutions are apps or technology that operate on-top, or separately, from the blockchain allowing for transactions to occur in a sandbox before being placed back on-chain. One example of a layer-2 solution are sidechains, like Liquid, which allow peer-to-peer transactions on a separate chain before synchronizing them with the main blockchain. Optimistic Rollups debuted in recent months with the launch of Uniswap V3 on both Arbitrum and Optimism. The latest method is Zero-Knowledge Rollups (shortened to ZK Rollups), which batch together transactions and then transforms them into a single settlement transaction on the Ethereum network.   

Starkware, how dYdX rolls 

If DEXs are going to “rule crypto” as dYdX founder Antonio Juliano proclaimed earlier this year, keeping the status quo isn’t going to work — they need an alternative to compress thousands of trades into a small size to run on Ethereum more efficiently.

And with the adoption of ZK-rollups, thanks to Starkware’s layer-2 solution, the exchange is now open to the little fish and not just the big whales who can invest a minimum of $10,000.

“We like to think of Ethereum as this prime real estate you only use for displaying really expensive stuff. It’s like a display window on Fifth Avenue or something,” Starkware’s co-founder Eli Ben-Sasson told Blockworks in an interview. 

But layer-2s currently have a centralization problem. After all, they rely on an outside operator to manage the transactions and send them back on-chain. 

Starkware is well aware of this and is actively moving to develop Starknet — a decentralized version of its ZK-Rollup mechanism, due out at the end of the year. 

“Things like Starknet are going to be the manufacturing plant, producing a lot of the stuff that then gets displayed and curated on a layer-1,” Ben-Sasson said. 

But does all of this take the “magic” out of Ethereum, as research house Santiment argued in a recent note? Not the case, believes Ben-Sasson.  

“I think that overall this would not decrease the value and usage of Ethereum, it’s the other way around,” he said. “Just like how you could move manufacturing facilities out of Manhattan or the center of London because you’re utilizing them for more expensive and important stuff.”

Ben-Sasson doesn’t think concerns over the centralization of layer-2s are as pressing of an issue as some might think, pointing out that if Starknet vanishes tomorrow, all the funds can be retrieved via a layer-1.

“Even though there was a central party running part of the system, the funds are always held in a non-custodial fashion, and you can revert to a layer-1 to get your funds back. But we’re not content with that. The path forward is decentralization,” Ben-Sasson concluded.

But decentralized or not, considering the uptick in volume over the last week — Dragonfly Capital’s Qureshi puts that on attractive token rewards and not a China exodus —  the market doesn’t seem to mind. After all, it is pushing through more volume than Coinbase. Maybe the future isn’t that decentralized?