In a few months since launch, Fluid has completely taken Ethereum by storm to become one of the largest dexes and lending protocols and leveraging venues at the same time. In the process, it created the idea of integrated DeFi, making us rethink the entire premise of pools and composability.
But hereâs the thing that most people still have no idea about - the team, in particularly Samyak, Sowmay, and Thrilok have been pioneering now-common concepts like smart accounts, flash loans, leveraged yield for years, and have now built several billion dollar TVL protocols along the way.
Iâve been working with this team closely for years now, and this is what I find utterly amazing about them â they have a rare ability to consistently turn wild, seemingly crazy ideas into insanely strong adoption and acceptance. And theyâve done it repeatedly, with a spotless track record: zero scandals, exploits, or issues.
With the platform finally getting the recognition they deserve, itâs now the right moment to highlight one of the most underrated, yet undeniably talented teams in all of crypto. Their story not only showcases how talent can emerge from anywhere in the world but also, perhaps, offers a glimpse into the future of Ethereum L1 itself.
I first met Sowmay and Samyak in 2019 while advising Kyber. At the time, they were literally two college dropouts from India with no connectionsânone to the Ethereum Foundation, VCs, or even prominent startups.
Yet, they were already building incredibly innovative interfaces. For example, they simplified Makerâs complex UX from 12 transactions to just 1, developed Makerscan (an explorer for MakerDAO), and created Easwap, an alternative UX for Kyber Swap.
Over the following months, Loi and I met them in Singapore, gave them a grant to pursue whatever they wanted to do, and soon after, Thrilok joined as a key member to help them build out everything from there on. That marked the beginning of one of the most intense periods of DeFi experimentation I have ever seen.
Before diving into all the things theyâre about to do next, itâs worth reflecting on just how intensely crypto-native they wereâeven five years ago. The very first thing they did with their USD grant was convert it all into ETH, deposit that ETH into Maker, and manage their expenses using borrowed DAI.
This stands in stark contrast to the VC-funded projects of that eraâprojects building DeFi while still mostly using traditional banking. True DeFi nativeness is a critical edge - allowing builders to fully align and create truly solutions â not inherited from traditional finance.
As mentioned, itâs mind-blowing to realize that before Fluid, the team had already been pioneering key DeFi concepts for years. With Instadapp, they have been working on smart contract accounts (aka account abstraction) from the very early days. But unlike ivory tower types who created more and more standards and terminologies, they got right down to work to create really cool stuff.
For instance, they were the first team to create the idea of flashloans to allow users to seamlessly refinance positions across various lending protocols, back then maker and compound. Following which, they created the idea of DSA (DeFi smart accounts), which allowed for extremely complicated leveraging and refinancing strategies that once required 10 to 20 transactions were bundled into a single, seamless transaction with great usability.
DSAs also had a feature that is pretty fucking awesome - by writing simple connectors at the contract level, you could compose whatever features you wanted to using very simple javascript - ranging from simple arbitrage using 1inch to complex leveraging mechanisms. Because each contract function was very simple, the usual security complications abstracted away nicely at the start and the end of the composability, it really did take composability and account abstractions to the next level.
IMO, working with them to launch DSA and seeing them tweet extremely cool strategies randomly was probably one of the most inspiring moments of my career and till today, I donât think people quite appreciate how they created a setup where capital could move freely and intelligently between protocols, and complexity could be made invisible to everyday users and devs.
Take Instadapp, for example. It introduced the very first smart accounts (we came out with the name together), setting a new standard for composability in DeFi. The team unlocked flashloans, allowing users to leverage and deleverage assets, and execute a variety of complex money-lego strategies. Of course, fast-forward to today, smart accounts and flashloan strategies have become the blueprint for nearly all DeFi protocols.
Then came Instadapp Lite, which delivered the first-ever leveraged yield strategy for staked assets. At the time, looping in and out of leveraged strategies was overly complex, gas-intensive, and manageable only for whales. Instadapp Lite changed this entirely, allowing users to earn 10-20% APR on staked ETH with a single click in and out, removing the need for manual execution or active management.
The awesome thing is that Instadapp Lite is actually built on top of DSA. However, there was a subtle but critical shift in how the team approached user adoption: instead of requiring users to manage their own smart contract accounts, they introduced a single smart contract account with built-in functions. Now, users only need to deposit into it, greatly simplifying the user mental model. It was a change that would later become the foundation for Fluid.
With Avocado, the team introduced chain abstraction for the very first time. It combined a single USDC gas tank with a unified multisig address across all supported EVM chains. This eliminated the friction of switching between networks and enabled use cases that were previously impossible, like bridging assets seamlessly between chains via a multisig. For teams managing complex operations, Avocado consolidated the chaos of juggling multiple multisigs into one streamlined address that works across all EVM chains.
And, of course, all of these critical learnings and innovations have led to what the team is building todayâFluid.
When the mad scientist Samyak first laid out Fluid to me nearly two years ago, it felt like the polar opposite of how someone usually thinks of DeFi - it actually felt like culmination of years of raw, unfiltered experimentation with composability, leverage, liquidity, and abstraction, serving real users and an unwillingness to be constrained by what DeFi is supposed to be.
Fluid, at its core, answers this question: âWhat if we integrated lending, leveraging and dexes into a singular layer, abstracted away all the ux and security issues for users for complex operations, and max-optimized gas, capital efficiency and risk for users and ecosystem?â
It is a mouthful for sure, so lets break down each of these 3 core principles:
Traditional DeFi separates lending, trading, perp pools, etc. Fluid integrates all of the same assets from different protocols into a unified pool, eliminates the fragmentation seen in protocols like Aave, Compound, or Uniswap - allowing the creation of a fluid system where both collateral and debt act as productive assets.
Note the similarity here to how Instadapp created this within the smart account system, but now it is within a liquidity layer instead.
In traditional lending and trading, managing liquidity or leveraging assets requires manual execution, constant monitoring, and high transaction overhead. In Fluid, borrow positions dynamically adjust based on trading activity. For example, if a trader swaps 500 USDC to USDT, the system reduces USDC debt and increases USDT debt for borrowers, maintaining constant total debt while generating trading fees.
This operational abstraction not only simplifies the user experience but also enables features that were previously impossible, such as seamless swaps between collateralized positions, and passive generation of fees on debt positions.
Crucially, much like with Instadapp, having everything integrated into one layer is the crucial thing here to doing complex operations with max security and ux.
Within Fluid itself, there are key algorithms that push capital efficiency, liquidation mechanics, and gas optimization to their limits, achieving financial parameters that are unthinkable in other protocols.
These algorithms include integrated DEX positions within the money market and the most advanced liquidation mechanisms. Collectively, these optimizations create insane DeFi parameters like:
All this sounds like magic, so letâs look at each of the following using some common examples:
Letâs start with the most basic example - the USDC/ETH borrowing vault, which has by far the industryâs most advanced liquidation mechanisms. Right from the offset, what stands out is the 87% LTV ratio, and incredibly low liquidation penalty of 1%, and liquidation only of what is necessary,
This means that in the event of liquidation, only the minimum amount of collateral necessary is sold to restore the position below the LT. For example, if a borrowerâs debt-to-collateral ratio rises to 96%, the system will liquidate only enough to bring it back to 95%. In practice, this could result in a penalty as low as $38 for a $3,800 debt, compared to hundreds of dollars on other platforms.
Fluidâs liquidation process is also gas-efficient, enabling multiple liquidations to occur in a single transaction with minimal overhead. For example, if 13 users with varying debt sizes need liquidation simultaneously, Fluid aggregates these into one transaction, keeping gas costs comparable to processing a single position. On Fluid, gas usage for liquidation typically ranges between 120k and 150k for liquidating all positions at once, a fraction of the 300k to 1M costs seen on other platforms for liquidating only 1 position
(Consider explaining how much it means in USD terms at average ~gwei gas fee as non-devs have no idea what is âgas usageâ.)
Importantly, unlike other platforms, liquidation is not a standalone process but occurs seamlessly as part of regular trades. This means that for instance, when a trader swaps USDC for ETH, their trade can indirectly liquidate a flagged position. This approach minimizes the reliance on dedicated liquidation bots while providing discounted collateral to traders and maintaining the health of Fluidâs ecosystem
If the advanced liquidation framework does not sound cool enough, letâs make things even more interesting with the idea of smart debt, which uses not a single asset as debt, but an inverse LP position.
Letâs use the familiar eth - stable example, but this time using a vault with a smart debt: the eth - usdc/usdt vault. This means your debt is not in usdc or usdt, but rather held in a virtual USDT-USDC LP position. So when a trade is routed through the smart debt, the trader is repaying one part of your debt and borrowing another part of your debt on your behalf.
Imagine the smart debt pool is currently (virtually) 500 USDT and 500 USDC. When a trader swaps 100 USDT for 99 USDC, that incoming 100 USDT repays part of the USDT debt, while simultaneously âborrowingâ 99 USDC. As a result, the debt composition shifts to 400 USDT and 599 USDC. In the process, the pool earns trading fees from the swap.
This is kinda mindblowing â how can a seemingly negative liability be used for productive activity? The answer, of course, lies in the fact that all the protocols in fluid share common underlying liquidity pools - so while the debt position is virtually represented, it can still be used, thanks to the existence of the integrated liquidity!
Now let us cover the opposite case - where users are able to use an LP position as collateral, for example in this vault, where the collateral is a WBTC-CBBTC LP. In lending protocols, your collateral will be a static asset, while on dexes, you will not be able to do anything with the LP token.
In this case though, your collateral is being used productively as both a liquidity provider and collateral. For example, a trader who is swapping 2 WBTC for 1.9 CBBTC (according to the AMM curve and after fees) will in essence be making the asset more productive than the collateral sitting alone!
Finally, letâs put all these together. Just like how you can leverage loop collateral and debt (remember Instadapp Lite, where we leverage looped stETH and ETH), you can also leverage loop smart collateral and smart debt.
Taking this vault as example, where both the collateral AND the debt are BOTH wstETH/ETH positions, both the collateral and debt are smart positions. And remember the very high 95% LTV?
This means:
So in summary, hereâs the 3 integrated layers that make it all work
Now and future, all new pools and protocols built on top of Fluid will be contributing to the same liquidity and utilizing the same liquidity - perps, options - all of these will add depth to the underlying liquidity pools, reducing borrowing cost and increasing flexibility of protocols on top of it.
All that is good and well, but how did it perform during key moments?
First, lets us look at liquidations - during extreme market events, such as the 16% ETH price flash crash, Fluid demonstrated its capability to handle liquidations swiftly and without creating bad debt. Users were liquidated efficiently with minimal penalties, and gas costs remained low, even under high transaction volumes. On my end as a defi builder, it is really quite amazing to see such a crazy ass advanced liquidation system work in real time.
In addition, due to the insane leveraging of asset liquidity -> trading liquidity (converting $1 into $39 worth of trading liquidity), Fluid became 3rd largest DEX on Ethereum within 3 weeks from launch, crossing $1B in volume in 3 weeks of launch, making it the fastest growing DEX ever on Ethereum with only 3 DEX pools live.
As for volatile pairs, Fluidâs ETH-USDC DEX has already handled about 10% of the ETH-USDC volume on Ethereum mainnet with a modest $20M cap in a few weeks, an insanely impressive accomplishment by all measures.
In other words - the core concepts of Fluid works, and of course, it is just the beginning.
Now, the story of the FLUID is a really great story in and of itself. Shortly after instadapp started getting traction, and before instadapp v2, sowmay, samyak and i started talking about the token, eventually naming it INST.
However, back then for some reason, all three of us knew that INST token was going to be a temporary name although we couldnât quite figure out the right name for it yet, so we left it modifiable.
The INST token went off to a great start initially powered by the great hype, but soon tanked off due to a lack of understanding of the protocol and overall market. Most teams would just give up or make empty promises or hype or just simply just stagger along, Sowmay, Samyak and Thrilok did none of these things.
They continued to build great products that people love, build up an amazing reputation in the space that allowed them to rack up TVL and usage for whatever they do, recruited a team of top-tier talent that can execute things and crucially spent 2 years building up an amazing product, extending both Instadapp, the powers of platform, understanding things and leveraging things and finally into what Fluid is today.
Impressively, in massive contrast to the vast majority of teams, they also didnât sell any tokens and instead focused on creating actual revenue-generating products for the team that the community, and of course building insanely hard away at the hardest problems in defi.
Special credit to Igor (DMH), who saw Fluidâs potential when Fluid was just announced over a year ago. Quickly joined the team, took the role of BD and within a few months became the core member and massively contributed to Fluidâs success through education & collaborations.
All that of course built an amazing foundation for the FLUID token, which they executed seamlessly. Itâs a simple rename with no drama and whatever not, and of course with a much better platform, much better community and much more excitement for the future.
Ethereum has led DeFi innovation but has recently lagged behind other chains in terms of actual usage, but probably more importantly, also in terms of optimism about its long term future. Still, the best part of DeFi is that anyone can jump in and redefine whatâs possible and push boundaries, which is exactly what Fluid did.
Built by three kids from India, far from Ethereumâs usual centers of power, Fluid stands out as one of the most innovative protocols around. Beyond being a breakthrough in DeFi, it could set an example for Ethereumâs future by showing what the community should value, appreciate and celebrate.
While many look to famous VCs, fancy super concepts and foundations as the future of Ethereum, Iâve always believed the true future of Ethereum is always where it should be - the users, the builders and the leveraging the core foundations.
There is really no need for 4 year roadmaps, magical thinking on how separate blockchains can operate as one, or pretending that all the factions on Ethereum have the same economic interests or vision as the mother chain.
All it takes is a focus on the users who have real needs to meet (like low liquidation) and desires to fulfil (like APY), and builders who understand the full potential of the network (deep, deep liquidity), how to work around limitations (like gas) and deeply aligned with the ethos.
And with Fluid, you have a team that embodies all of these.
The future of Ethereum is incredibly bright, we just need to look in the right places.
I am a long-term adviser to instadapp and fluid. I have INST tokens since the very beginning, never sold.
Much thanks to DMH for helping with much of the detailed Fluid explanations, Ignas for helping with the formatting and Julian for helping with the proof-reading.