Decentralising Finance With Hardware
9 min read

Decentralising Finance With Hardware

Decentralising Finance With Hardware
Photo by Alexandre Debiève / Unsplash

Hey there,

The world of DeFi seems to be rattled by the latest Sifu incident. One of the anonymous leads of a DeFi project was linked with a history of fraud. We will not be digging into that in today's piece. For the curious, I highly recommend reading this RektHQ article on the matter. Instead what I wanted to do today was surface insights I had from a recent conversation with Aditya Palepu - founder and CEO of Dex Labs.

I am not sure if you noticed, but Paradigm and Sequoia invested $1.15 billion into Citadel Securities recently at a valuation of $22 billion. For a sense of scale, Citadel is a ~20-year-old organisation that handles 30% of the trading volume in the US. They are part of the 'secret sauce' behind Robinhood's ability to function as a free application. You see, when somebody sees a random influencer shilling the next Gamestop on Twitter and decides to slam the shiny green buy button on the app, that order flow information is packaged and sold to players like Citadel Securities. Having an edge of nano-seconds on inbound orders can allow market-makers to make small sums in profit, millions of times. Wall Street would literally dig up a thousand-mile line to lay fibre-optic cables to cut 1/1000th of a second. I am not making this up (consider reading Flashboys for added context on this). This is partly what intrigued me about DerivaDAO when I first spoke to its founder Aditya Palepu. Whereas most exchanges focus on a variety of instruments used and chains supported, DerivaDAO's key focus has been to extend the hardware layer. So I sat down with him to discuss what he's building and why it matters.

Aditya is the CEO of DEX Labs – the firm building DerivaDEX. Their stated intent is to build 'capital-efficient, user-centric, decentralised technologies powering web3 economies'. He graduated from Duke University in 2013 and spent the majority of his professional career at DRW as an institutional algorithmic trader trading everything under the sun – from treasury options to commodities and equities. The crypto-bug finally got to him in 2017, so he decided to trade on his own and then began reskilling himself towards becoming a blockchain engineer. That early experience in trading and an interest in how these things work intrigued Aditya about decentralised exchanges. Compared to the speed at which an exchange could process orders in the traditional world, crypto seemed to be running on an archaic infrastructure. For context, until quite recently, most large traders in the industry had two options when trading exotic instruments: One was to go with a centralised exchange which would offer efficient markets, strong liquidity and thick order books but work in a custodial way. The other was to work with decentralised avenues of high security that often had no liquidity. This is the gap Aditya, and his team were looking to address when they set out. DeFi has come a long way since 2018 and compared to the gas wars and multi-minute-long confirmation times needed, we have a whole suite of alternatives and layer-one solutions today. Exchanges like dYdX use Starkware to enable faster confirmation times, but we still have a long way to go.

To summarise, today's exchanges have three key issues:

  • The possibility of front-running bots shaving a few percentage points on each order
  • The order flows being captured at the exchange level to be used by market-makers (like Citadel does with Robinhood)
  • Miners deciding to prioritise certain orders to move markets in ways that can be advantageous
Image Source: Cowswap's breakdown of how MEV works

MEV (Miner Extractable Value) protection is  becoming increasingly accessible through the use of 0x Matcha and Cowswap. The general approach in the industry involves order bundling, in which assets are auctioned off at the end of each block, leaving very small gaps for front running. The image shown above explains how Cowswap typically solves for the issue. Instead of passing on each trade for transaction finality, Cowswap matches orders internally first then bundles similar order types and settles them via AMMs on-chain. What this basically means is smaller orders are settled without necessarily going to a third party exchange and the larger ones go together to save gas fees. But to truly decentralise markets, we need to go to the hardware layer, which is DerivaDEX's current focus.

Enter Trusted Execution Environments

Note : I am no expert in TEE. I have referred to as many resources I could to verify claims here but if you think something's off - hit the reply button. Thank you!

DerivaDEX is a decentralised derivatives trading platform. The platform itself will have an open order book similar to BitMEX, but the infrastructure and cryptoeconomic incentive design will be drastically different. Part of what sparked my interest in DerivaDEX is its focus on the hardware layer. But before we dive into that, I'll need to explain what a trusted execution environment (TEE) is. When you run conventional apps like Google Maps or Instagram, the app has access to data related to your geographic location, usage behavior, and in some instances, even voice recording. Given the right permissions, these snippets of data about the user are uploaded to servers owned by the app, often forever. This is how Facebook likely still has access to embarrassing images of you from 2010. This data could be traded with other parties or used to better understand the user.

Image Source : Link

A trusted execution environment is a component of the chipset where the software is typically pre-installed, and the data that goes through is tamper-proof. That means, if the owner of a device wants to tinker with the code that verifies the validity of inbound data, it may not be possible. Similarly, the owner of the hardware device itself cannot tinker with the kind of data that goes into it.  They are primarily used for using tamper proof computation and giving out an outcome to the non-tamper proof part of a device. TEEs are currently used in day-to-day applications such as fingerprint verification on mobile devices, facial unlocks and payment applications like Samsung Pay. Long story short, it is a way for
(i) data to be collected and parsed at the hardware level and
(ii) using software that cannot be edited by device owners.
I recommend watching this video if you'd like to learn more about how TEE works. It is beyond fascinating.

Why does all of this matter for a derivatives exchange? Hardware providers in today's settlement layers are the often-ignored players in decentralising finance. Using a network of TEE based node providers for liquidation, price feeds and order settlements add a layer of security to DerivaDEX that insulates it from several issues plaguing decentralised exchanges today. The miner-extractable value is almost entirely eradicated since the order flow is encrypted and cannot be accessed by anyone, even by individuals with physical access to the hardware. Unlike Robinhood, like with most present-day AMMs today, front-running could also be eradicated as third parties have no view of these orders until they hit the exchange. Each node operator will be expected to do the necessary computation and send the final state of the transactions to Ethereum's base layer every ten minutes or so. This means the team combines its network of hardware operators with Ethereum's existing L1 security. Transactions on DerivaDEX are bundled and recorded on Ethereum once every ten minutes or so.

This focus on decentralisation extends to the team's token economics design. DeFi-native applications have found users in the past by giving users tokens in exchange for active participation. Compound used to offer tokens to those lending or borrowing on the platform. Centralised exchanges had a variation of this where they gave tokens to users that traded the most. Often referred to as transaction mining, the model incentivises users to wash-trade and thereby create unsustainable behavioural patterns on the network. Thus, if users receiving tokens sell, it only creates a constant sell pressure on the network. DerivaDEX uses elements of this to ensure that traders and stakeholders help build the ecosystem. Close to 50% of all tokens are allocated towards liquidity mining and insurance mining on the platform. However, these tokens are not given in a very short time frame. Instead, the tokens are rewarded over a 10-year horizon, basically eradicating interest from individuals looking to do short-term transaction mining and vanishing. Side note - 10 years seems really long by most benchmarks. Average vesting for most tokens I have seen comes to around four to six years.

What intrigued me about DerivaDAO is that the team has committed to all aspects of the venture being driven by the DAO. Aditya explained that he sees DerivaDEX as a crypto business, not a crypto project. Most DAOs typically run into an issue of being non-profit cooperatives. Part of the reason for this is the team's lack of effective governance control. The open parameters for voting are not defined before a token is released into the wild. In comparison, DerivaDEX has explicitly stated that platform fees, operator incentives and the pace at which they are granted will be determined by the community over time. I think, compared to what we have with the likes of Deribit and, lately, dYdX, DerivaDAO differentiates itself on two fronts. First is its obsessive focus on building a decentralised infrastructure layer for the product. Second is its philosophy around building the product to be DAO first. In many ways, Derivadex looks like an intriguing experiment focused on solving some issues that plague decentralised markets today. By reducing the ability to front-run orders, it is making a relatively more fair and equitable market. It passes ownership to the platform's users over time instead of holding on to it.

Much of the $ captured in Derivadex currently goes towards their insurance fund. 

None of this suggests that the team behind DerivaDEX functions purely on an altruistic basis. Part of the token supply vests to them over time, although the DAO holds the right to vote against it if the community believes it needs to be done. Compared to a conventional exchange where the entity behind the venture derives value from equity appreciation and trading fees, the team makes its money from the rise in the allocated tokens' value. This lets the platform use the fees generated for what it determines as best. Platform fees generated on DerivaDEX go towards the insurance fund – a pool of capital used to make users whole if the platform malfunctions and cannot do so due to faulty trades. This post by BitMEX is a good breakdown of how insurance funds are used in derivative exchanges. The DAO can realistically suggest ways that this fund can be used. It could be used for anything ranging from research to marketing.

There has been little focus on the infrastructure layer or order flow in the DeFi ecosystem, even as the space revolutionised finance in a matter of years. To me, DerivaDEX is a return to the basics – where the stack has been redesigned from the ground up and is slowly approaching a point in time where complete decentralisation is possible. Remember the bit I mentioned early in this article about why Sequoia and Paradigm are acquiring Citadel? DerivaDEX makes that move redundant by offering a far better solution. There are, however, challenges to the model:

  1. It would be interesting to see what kind of volume the DEX can generate as it goes live.
  2. Given how DAO apathy works in general, it may be a while before members proactively begin contributing to the DEX.
  3. The token emission may be so long-term that a very short attention span market considers alternative opportunities (such as NFTs) to spend its time on.

If you are wondering how to get involved with DerivaDEX, their test net is already live. Hop on to their Discord and join in on the conversation around how and when the main net launch will happen. I am most interested in seeing how the team will bootstrap a network of hardware providers worldwide to truly decentralise the network because that is what makes DerivaDEX censorship-resistant. Since the venture itself is structured as a DAO, hopping into the Discord and proactively contributing could be one way to get hired if you want to join a web3 native venture full time. Ultimately, Aditya sees his team's role in the DEX declining over time. They describe themselves as a tech shop currently doing a lot of R&D for what the market may need in the years to come. I believe the world needs more individuals doing just that, given how DeFi native ventures go down from time to time and routinely censor users from different parts of the world.

We will be hosting an AMA with Aditya the coming Tuesday. You can RSVP for it here. See you on Monday. Also make sure to come hang out in the community Telegram to discuss more on the venture.


1. Financial entities I am linked to may have exposure to assets or ventures mentioned in the article.
2. Not financial advice. Digital assets currently have a high amount of volatility. Do not invest what you cannot afford to lose.


Job Board

We launched our Pallet a few weeks back. The plan has been to build a small curated list of jobs, ideally with ESOPS or carry from vetted founders in the community. Make sure to check the list of 25 jobs we have listed if you are looking for something new. Roles cover everything from CTO roles to opportunities in VC and non-tech roles. Get in touch with us if you are looking to list a role there.

Apply here.