Morpho Labs Co-founder and CEO Paul Frambot spoke to Variant Fund Investment Partner Geoff Hamilton at EthCC in Paris about his framework for the two types of DeFi projects, the limitations of existing lending platforms, how to build a truly scalable lending protocol, and the next generation of Morpho's lending primitive.

I'm Jeff, 1 of the investment partners at Variant, and with me is Paul Framboe,

We're going to kick it off with some introductions, and After that, we'll talk about DeFi lending and lending markets and then we'll wrap up with some Q&A So everyone will have a chance to ask questions For those of you who aren't familiar with variant We are a fund purpose-built for early stage investing and we focus on investing in crypto networks that make ownership a keystone of their user experience, which we believe helps them grow faster and scale larger than they otherwise would.

We had the privilege of co-leading Morpho's seed round and It's been amazing to work with Morpho, with Paul, and the rest of the team.

They're some of the most deliberate and thoughtful builders that I've come across.

So I'm very excited to give Paul the opportunity to share his thoughts with you, and I'll hand it over to him for his intro.

Hey, thank you very much, and thanks everyone for coming today.

So some of you know me, I'm Bo Lam's CEO of Morpho Labs, and we're building the Morpho protocol, and today we're gonna expand on what is our vision of DeFi lending and DeFi in general, and how we think the space is going.

And yeah, I'm very excited for this chat.

Please prepare your questions if you have some, and I'd be very happy to chat with you during the Q&A session.

So we have to start it off with something a little bit more exciting.

And so something that I didn't know about Paul until we started preparing this event was that he originally wasn't excited about DeFi and was suspicious of finance in general.

How did you get from that to building Morpho?

So the truth is that when I started building in the blockchain space, I was more focused on consensus and layer 1 stuff.

Like it was too negatively associated in my mind just because it was too related to speculation or whatever and could not see the value for it in the first place.

And what happened is that I was introduced to DeFi, okay, I understood the fact that it was more transparent, you know, more open and more resilient, potentially more efficient, but I could not see myself, you know, spending my whole lifetime building on that space.

And really what really convinced me to actually get into DeFi was mainly 2 things.

The first thing is that DeFi is going to bring, and is already starting to bring, insane cost reduction.

So just like, you think about it, you have more than 50, 000 different banks and institutions in the world.

Like literally, they all have their back-end servers running the same algorithm performing the exact same operations.

The idea that you have 1 open and simple algorithm out there that is performing the operations on behalf of 50, 000 others is kind of clicks in my mind as bringing a really important cost reduction for the users.

And the second point, and probably the most important 1, was about user alignment, and how DeFi was fundamentally more aligned with the end user.

And so you can see that when you look at traditional finance business models, it's all about operating a financial service.

So basically, as a DeFi, as a tradFi business, what you're doing is essentially operating a financial service.

In DeFi, you're paid to enhance a common good.

And this is really what got me into DeFi initially.

So let's start off, if you could just tell people a little bit about what Morpho is before we dive into the rest of the discussion.

So, what Morpho is currently, so it's a landing protocol.

It's currently actually the third largest on Ethereum.

We have something around like $1 billion deposited in the protocol.

And it's different from other landing protocols because it's built on top of existing ones.

So we're basically, you know, optimizing the efficiency of having compound by providing a peer-to-peer matching layer on top of their respective protocol.

And with time, we learned to present ourselves not as optimizers, but more as landing protocol in general, because we provide the exact same service and the exact same user experience as existing netting protocols basically.

the thoughts that you've put out there recently.

For those of you who haven't read it, Paul recently wrote a piece about the 2 paths for DeFi and forms some of the basis of what we'll be talking about today.

1 of the key things that you introduce in that piece is this concept of decentralized brokers and internet protocols.

What's the difference, Paul, and why is it an important framework?

Yeah, I do think this is very important, DeFi in general.

The way we see it is you have 2 rough categories of DeFi platforms.

1 would like to call them as decentralized brokers and the others are protocols.

So decentralized brokers are literally on-chain funds.

Platforms are managing funds for the user.

So you could like, for example, quote Aave, Maker, et cetera, Where the DAO is highly involved in making the experience of holding DAI or the experience of lending to a lending protocol completely passive.

And this approach brings mainly 2 interesting things.

The first thing is you get a passive user experience.

And the third thing is that it's extremely well suited to aggregate liquidity as you're going to fungibilize every single user as 1.

And basically, the DAO will be in charge of managing the risk parameters of this overall platform.

The second approach is the protocol approach, where basically a protocol in that is more of a trustless and unopinionated DeFi protocol, DeFi product, DeFi platform, where you narrow down the usage of your platform to something very specific, and you try to make it completely trustless.

So that comes with immutable code, for example, and as well as being as much an opinion in general as possible.

So the great example of this approach is obviously Uniswap.

You have similar attempts like liquidity in the field of stable coins.

In both approach, protocols and decentralized brokers have pros and cons, and we can discuss them probably in the case of lending.

And depending on the use case, 1 tend to be more interesting than the other.

And maybe you could tie that to what are the opportunities that building in the protocol mode unlock in lending specifically?

Yeah, so It's been 2 years that we've been building on top of decentralized workers.

So we've been building on top of AVI for 2 years.

We currently are the biggest user of AVI in compounds, and we have more than $1 billion sitting on top of it.

And really, the 3 pain points that we've realized are the fact that it's not trustless at all for many different reasons.

It's not scalable, and it's not efficient.

And so decentralized brokers apply to lending.

In the case of the trust assumptions, basically when you're lending to Aave or Compound, you rely on the fact that the DAO is going to basically set the correct risk parameters on your behalf.

So what that means, basically, is that on Aave, you have more than 500 different risk parameters.

So basically the DAO every day is going to push proposals to update the state of the smart contract and make sure your funds are correctly managed.

So that's the matter of the risk side of things, but there is also the smart contract side of things where the code base is extremely large and There are often you know bugs and fixes that needs to be done with like as Morpho Labs basically reported some of those bugs and basically the DAO is going to upgrade the protocol.

So that's the trust assumption that you make using a decentralized protocol.

Now, this comes with scalability issues because obviously, The more features you want to add, the more collateral you want to list, the more parameters you're going to need to monitor.

The DAO, well, as a decentralized entity, is not necessarily the best suited for that purpose because it's going to be much lower And token holders are not all experts in risk or in lending.

And so it makes the overall thing much, much harder to manage and to scale to new use cases, in some sense.

Which leads us to the last point, which is efficiency.

And this is more of an intuitive point, but when you think of AVI and Compound, they're doing a lot of things at the same time.

So there's a bunch of different assets in it.

And you cannot get the intuition that in order to maintain the risk of the protocol or the platform quite low, you have to consider the worst assets of the basket.

And I know some of you are really familiar with this, that the idea that if I have USDC and CRV in the same base pool, well, I have to consider CRV when setting the risk parameter of USDC.

What that creates overall is a reduction in the efficiency of the overall system where the rates that we get, the capital efficiency that we get is nowhere close to what we have in traditional finance basically.

So building protocols as opposed to decentralized brokers has the potential to unlock greater trustlessness, scalability, and efficiency.

How can protocols actually realize those opportunities?

So, I mean, decentralized brokers have been amazing at providing a flexible and product-oriented financial product, like literally, for lenders.

And so, again, when you lend capital to Aave, you don't get to do anything.

When you lend capital to the current version of Morpho, you don't get to do anything.

What we believe is the path for less trust assumption, more efficiency, and more scalability is the protocol path.

So, reducing all the scope of the protocol to 1 simple thing in order to do it extremely efficiently and completely in the code.

So, in practice what that means is that all, that means that most of the complexities that are inside of the protocol might be pushed back to the user.

So, for example, risk management is like 1 of the big components of a learning protocol.

You have many of them, you have rates, risk management, etc.

Risk management is 1 thing that could be pushed back to the user.

And with this, because the DAO does not have to do risk management anymore, you can come up with a completely trustless primitive, And then you can require other actors that are the risk experts to sort of recreate the passive experience on top.

So basically in that framework, you can literally recreate the decentralized broker experience on top of the protocol base layer basically.

And to us at Morpho Labs this felt like the obvious way to go after like 1 year building on top of a decentralized broker.

Like we see you know Morpho has like the biggest DeFi growth in the last year like relatively And we just don't see it working much further than that to the $5 billion that Aave has reached to us.

It's like literally the cap of the decentralized brokers path for lending.

And this is why we've been working on protocols for a while now.

And we do believe that's the only way to scale to the trillions of dollars that traditional finance is able to get to daily in the over collateralized lending markets.

So 1 of the things that you mentioned was that building protocols in the way that you've described pushes some complexity to the users.

I'm guessing that is the intuition that many of us had in hearing your comments as well.

How can protocol builders address some of that complexity?

What's the approach that makes the most sense?

So I guess the way we see the thing is that DeFi needs a very efficient and trustless core that is completely unopinionated, such that more complex builders, users, basically like product-oriented people actually rebuild the abstractions that are needed to basically reach for the masses.

So, what that means in practice is that we're going to create a base protocol that is ultra-efficient, maybe harder to use, But because we put everything in the trust, the scalability, and the efficiency of the core primitive.

And then on top, you have potentially risk experts rebuilding abstraction for passive lenders.

You'll have more basically risk management across 1 pool, but also across different assets.

You could have like real world assets management.

You could have like liquidation productions.

All those kind of things can be rebuilt as layers on top of the efficient core basically.

To us, this vision is just like, offers much more flexibility for the space to grow, both vertically, but also horizontally.

And we do believe that This is so much of a stronger model than what we are currently having in the DeFi landing space.

And again, this is what we've going that path for the last 6 months.

So it's clear that by bringing in some of these actors, you unlock greater scalability.

And again, you take some of the burden off of the protocol and governance.

Any thoughts on how do you get attract the right actors to the ecosystem to help perform these functions for on behalf of the protocol and users?

So, open source is extremely powerful for that purpose.

So, in the vision of a stack, and we see a lot of more and more web-free businesses adopting the stack model, we're open sourcing all the tooling that's necessary and getting the tools out there for any builder to actually, you know, build what they're good at doing.

And you know, they don't have to care about anything else.

And you know, the tool is out there for them to basically have their product up and running on top of like the base core protocol is obviously very important.

And really the way we thought of it is like it has to be products built on top of protocols has to be venture backable.

Like I really think of it this way, is like, you need to build a base protocol on top of which you can be able to, build a billion valuation startup, literally.

And, When you think of the IP protocol, that's literally what that is.

It's just like a very inopinionated layer that's connecting, you know, servers and applications.

And it's like some sort of narrow ways through which all the traffic goes.

But on top of it, you have so many different applications.

And obviously, IP protocol is like, there's so much value contained in it, like so many multi-trillion dollars like companies working and building on top of it.

So, yeah, I think those 2 like open source and making it such that it's venture bankable is truly important and a key component for the success of like building a stack on top of a web-free protocol.

So the economics of building on top have to sort of stand on their own.

has to be highly profitable for builders to build on top.

So, on 1 side you have the gain, which is make it highly profitable, and the other hand you have cost reduction, which is open sourcing and bringing as much tooling as possible.

The power here is that the more tool you put out there, the easier it's becoming to basically build on top.

Like, you know, there's countless libraries, you know, abstracting the IP protocol, obviously, like all the library interacting with the web are like, you know, open source.

And this is what makes it so much easy to interact with the internet that we don't think about it nowadays.

And I think this is really a key component to the success of the stack model.

And you see that a lot with optimism going the LP stack model, Z-Casing also going the ZK stack model as well.

And yeah, obviously more for also going that path in the coming months.

So you made a really strong case for protocols as having a lot of potential in defi lending and in general.

Did you see the protocol approach taking share from the incumbents who are mostly built along the decentralized broker model?

We have to give the audience something spicy to react to.

It's clear to me that protocols are the right way to go in terms of specifically for the lending use case.

I think depending, you have use case like stablecoins, I think decentralized brokers are a very, very good approach.

I think Maker does an awesome job at doing decentralized risk management.

In the case of lending, there are so many different verticals, so many different collaterals, et cetera, that we have to narrow down to protocols and then re-aggregate and re-abstract everything with brokers, basically, centralized or decentralized.

And to me, it's very clear that in the long run, this is a completely dominant model.

I could not see a decentralized broker governed by token holders handling trillions of dollars.

Just think of how do you, where would you put all the world's financial value?

Would you put it into a decentralized lending broker?

Do you think that the incumbents, decentralized brokers, can evolve towards the protocol model successfully, or are there structural impediments that prevent them from doing that at this point.

If they were here, maybe some of them are hearing what you're saying and agree because you've made a pretty strong case, what can they do to incorporate this?

Or is it just, again, not something that they're able to do?

It's like we're moving away from this decentralized broker approach for a protocol approach.

It's easier for us because most of the complexities around risk management were taken from the Average Protocol directly.

And we did not have to really do risk management just because Morpho protocol was taking that from the other protocol directly.

What I think is very hard with protocols that have already decentralized is that the purpose of the token holders themselves relies in the fact that there are decentralized brokers, that there are risk managers.

So if you want to pivot from this, it's basically changing the purpose of the token itself, which obviously, the stake is distributed in such a way in existing decentralized brokers that they don't really have the financial incentive to move into another direction.

And now on the operational side of things, DAOs are not at all a Silicon Valley startup that are able to pivot very easily.

We've seen very few cases where DAOs were able to actually pivot.

And I think, you know, you can make the case that DAOs are relevant for very scalable protocols that have rich product market fit.

This is where you decentralize the most and you get, you know, you move very slow because every change involves like trillions of dollars.

And this is where DAOs are, you know, structurally much better than Silicon Valley startups.

And I think that some of the daz we have in the ecosystem right now are just not structured to pivot from 1 model to the other.

Yeah, that makes sense to me and I completely agree.

So before we move on from this topic, the broad framework of decentralized brokers and protocols and what it enables for DeFi lending.

Is there anything else that you want to share with the people in this room in terms of what this means for them?

Are there things they should be doing, ways they could be helping bring forward this more protocol-centric future?

Yeah, so obviously like Morpho is like building a learning protocol, like a true action.

And so, you know, we've been talking to most of you about this and we are obviously, you know, looking to work with many of the people here to, you know, toward building more, you know, clear and trustless and efficient and scalable financial stack.

I think it's just also a big work of education that we have.

The beginning of DeFi was mainly, DAOs had a big role in DeFi, except for Uniswap.

In many, many DeFi protocols, DAO had a big role.

And we're so used to this that we don't put it so much into question nowadays.

And I think there is a good, you know, a lengthy and hard job of like trying to educate people and like that all the paths are possible if there is like more work and like the stack approach with a core protocol at the base layer is actually possible and we're not stuck in the loop of like, you know, necessarily having a DAO that does the hard work, some of the hard work of the protocol.

So I think realizing that is the first right step and understanding deeply why it hinders trust, scalability and efficiency is the right first step.

And then building with this is like a good second step.

So We're going to shift gears a little bit actually and talk about some of your experience building Morpho and in particular 1 of the things that I think is very differentiated about Morpho is the approach that you take to research.

So first, can you share a little bit about some of the most exciting research or areas of research you're doing?

And then I'd love to talk about your process and how that works.

So we've obviously, so the way Moff Labs is structured is like we have a separate research team.

So we've been obviously looking at any areas of research that you have to get involved to when you're doing lending.

So which involves interest rate models, basically liquidation mechanism, MEV around liquidation is also an interesting areas of research.

Then you have obviously like the base, solidity, data structure and architecture around how do you build like the margin engine that enables someone to take a loan from a collateral.

That's also a very interesting sort of like design space.

And moving to like how should we structure this.

I do think there is a fundamental difference on how do you structure a decentralized broker than you know a company working on a decentralized broker and the 1 building on top of a, for a protocol I'm in.

On the decentralized broker pass, obviously you want many, many different contributors.

You want like 1 base code that is very large and you basically run it the same way you would run it, a Silicon Valley startup.

So it's like you upgrade your protocol many times, you add features at the protocol level, like at the platform level, you add features and you know, there are bugs, but okay, no worries.

We're just going to upgrade and fix the bug.

And those practices are inherited from Web 2, basically, and again, Silicon Valley, et cetera, which is good to make product-oriented

To me, it only works for the upper part of the stack.

For the base layer, you really have to think of DeFi protocols as internet protocols.

And I do believe that This does not come from like, you know, it's rating.

Obviously, IP could not iterate with its user when it was designed.

So it more comes from like, you know, a very thoughtful like process of like thinking about what you want from the service to be executed, narrowed down to something very simple, could be swapping tokens.

Like, you know, the only thing Unisoft has been doing is like focusing, a little bit later, focus on swapping tokens and making it very, you know, as easy as possible and as efficient as possible of swapping tokens.

And then designing the right simple and opinionated protocol to actually make this happen.

And so, yeah, and this requires a different structure to work on those kind of problems where you may want to keep like the protocol team as small as possible to reflect you know high talent like high talent density at the protocol team and you only scale the parts like that are around the protocol itself.

So it's like much more slow pace, much more vision oriented, not so much user oriented in the end compared to like web 2 approach, which I think is, you know, very counterintuitive as entrepreneurs.

Like, but if you look at Like all the successful like DeFi or cross-chain, like, you know, Layer0 is going that path and they've been very successful and they've done the opposite of building the products.

They've been like literally a low-level protocol and they're killing it.

You know, Lido is going with that path with Lido V2.

What they're doing is like a very unopinionated protocol to root basically stakers and people like buying STEs and infrastructure.

And they're not building products, like literally.

Then they rebuild products on top, of course, but it's just like focusing on building the unopinionated protocol is like a very long process that, you know, takes more of a vision than user feedback.

Of course, user feedback is part of it, but it's just like less in the weights than what it is in, you know, Silicon Valley startup, I would say.

So we're about to make the shift over to do audience Q&A, so people think about what you want to ask Paul.

Before we do that, Paul, is there anything else from your experience building Morpho that you think would be interesting for the people in the audience, builders, investors, other things you've learned during the journey?

See all Variant transcripts on Soundcloud

Morpho CEO Paul Frambot at EthCC 2023