Panel 4: Searchers and Builders

Panellists
Matt Cutler - Blocknative (𝕏: @mcutler)
Nathan Worsley - Faith Builder / f1b.io (𝕏: @@NathanWorsley_)
Kubi Mensah - Titan Builder (𝕏: @kubimensah)
Tom Zhao - rsync-builder (𝕏: @Tmzhao1)
Uri Klarman - bloXroute Labs (𝕏: @uriklarman)

Moderator
Stephane Gosselin - Frontier Research (𝕏: @thegostep)

I want to introduce our moderator for this panel.

This is Rajiv, he recently joined the Frontier team.

And I mean, this is gonna be a really packed panel.

I think we'll have a lot of spicy debate about the future of block building.

So without further ado, I'll just hand it over to you, Rajiv.

I wanna talk content, so we'll just do really quick intros, starting with Matt.

We're a Ethereum relay block builder, gas estimator, mempool specialist, and hopefully many of you here have used our tools

I'm Nathan bit of a background as a searcher, but I'm currently leading the adventure at Frontier, the builder.

Hi, I'm Kubi, co-founder of Gattaca, the team that operates Titan Block Builder.

We're an infrastructure provider, a block builder, a relay.

If you're doing anything in DeFi and MEV, probably use us or know us.

Also, for everyone, could you each say if you're a searcher builder or not, like for the audience?

So Blocknative operates builders, not searchers?

Yeah, we are not a searcher builder at the moment.

We run a builder, but in general, because we're a service provider, we don't do actual searching.

Maybe we work with other partners who do it, but we ourselves don't do it.

When I think about block building, there's 3 key ways to differentiate to win blocks.

You can have better access to order flow.

I'd love for everyone to stack rank what you think is more important for winning more blocks.

I think order flow is king at the moment.

We would hope that in the future merging would matter more, but just right now It doesn't matter anywhere near as much as AutoFlow does.

I think there's a threshold of certain AutoFlow that is not exclusive today that most builders can have access to.

So in the short term though, technology and being better at merging and infrastructure will give you an advantage.

But without the order flow in general, you don't have the building blocks to build blocks.

I guess if you have a really shit pipeline, then you're not going to win any blocks, but you still need the order flow, right?

OK, spicy take to the different direction.

And I'm not sure I believe everybody sitting in the panel here.

Block building is in the domain of those who can do CIFI, DeFi arbitrage.

of all MEV comes from CIFI, DeFi arbitrage and the amount that's being paid to the validator is not the

but somewhere between 37 and 77 so average 50%.

So maybe you do need order flow because without it then you're at a disadvantage, but it's done by the CFI-DeFi arbitrage.

Maybe they're all very good, so it's not a big differentiator between them, but between them and everybody else, what makes them the kings, not the king among them, but really at the top is CFI, DeFi arbitrage, a lot of capital and the ability to do that, willing to get pushed back against it.

You just said order flow in a different way.

It's just order flow that certain people might be generating themselves, right?

Okay, I took order flow as, okay, the transaction on chain out there, etc.

Well, in reality, Okay, these are counter trades to CFI trades.

I mean, it's just like, it's the same, right?

It's just order flow, it's like bundle flow, whatever.

You just have, if you don't have these transactions to include in your block, like you can't build a block, right?

I think you also have to make a distinction here between the amount of blocks that you have and the value of those blocks.

Yes, if you're talking about blocks-based value consume, then probably CFI, DeFi is kind of king.

But if you're just talking about raw inclusion, then I think that like just general order flow matters a lot more.

The CFI, DeFi trading alone won't get you all of the smaller blocks where there's not the large MEV opportunities.

I feel as if it's the case in which it's more so when you're referring to HFT trading, I feel as if it's you can sort of group it into, let's see, I would call it exclusivity of order flow, so more so both external order flow but also the order flow that you create internally from HFT trading

Yeah, I'm gonna build a spicy take on top of Uri's spicy take, which in the supply chain that we have, you can abstract it down to 3 classes of actors.

There's a trader class actor, there's a validator class actor, and there's a sucker class actor.

So you can go through and you can say user, which 1 of the 3 are they?

What about like, I don't know, relay operators?

So it's like literally just go through and you can say, well, either you're a trader, a different type of trader, there are various types, or you're a validator, or you're a sucker.

And this, I think, is super dangerous for the long-term health of the network.

Do we want to build systems where the customer is a sucker?

So going to the trading side, it seems like the only people really making money on building are the people who are also trading.

So how do people think about just making money in building generally, considering that for any of us to continue at this long term, presumably have to make money at some point?

So I think in the short term, there is a non-significant amount of revenue that can be made if you are a top builder without trading just by having edge on like the things that we already discussed.

Just So the nature of the auction is such that over time, and it's already happening, the margins on those are going to be compressed.

So 1, are you going to be an integrated searcher builder, so trade, make more revenue that way?

Or do you create additional value and generate revenue that way?

And so I Think and that's also part of our thesis that is significant amount of revenue that you can unlock as a neutral builder And that's what we're going for.

Do you have some examples of things that might unlock additional value as a neutral builder?

So in the short term, if you think specifically about services that are targeted at the most sophisticated consumers of block space today which are traders and so services that facilitate or almost put traders that are not vertically integrated on par with vertically integrated builders that is value-add And you can charge for that.

Yeah, I think I take the same view that the building in general, you're operating an inclusion service for searches primarily.

So what can you do for them as a customer?

Like, you could either vertically integrate, or you're saying, well, you could offer this as a service.

But there's a lot of trust and nuance there.

Do you consider these, well, builders could offer it to their users in general?

Or is this, OK, I can pick 1 or 2 or 3 customers and give them access to it?

It's very similar to, let's say, BlocksRoot offering mempool streaming services, low latency distribution.

So it's almost about commoditizing the edge that an integrated builder has today, and then giving everyone the same access.

So block building itself is a very hard job.

And so I think in order to be relevant as a builder, you first need to control enough block space.

So you're interesting enough for other players to take you serious, which is a hard problem to solve, so a high barrier to entry, and once you are there, start innovating and using the leverage you have.

So, I disagree but I'm going to push it about in a second.

So it is interesting to contemplate sort of how much of the Ethereum network is on the stage right now.

I mean, just between the relays in Uri and BlockNative, it's like

And then among the builders, I would guess it's about the same.

There's well over 80% in the room, if you want to count.

And I think 1 of the interesting opportunities around this space, while the building may ultimately be some flavor of trading or some latency games, there are a lot of opportunities potentially surrounding it, like around observability.

As we think about the complexity of the Ethereum roadmap, 1 of the general vectors is more mempools.

Mempools have really poor observability properties, and that is the why of why everything happens.

And so if you think about mempools as the root of MEV, then observability becomes really important, and entities like us, and obviously like Blockstrate as well, add a lot of value to the ecosystem.

And so we're looking for opportunities sort of on the edges and surrounding to add economic value and to receive.

The key is adding economic value and receiving economic rewards as a result because today this current structure is that receipt of economic rewards is highly asymmetric

Pushing against what Titan Kubey from Titan said our block builders entrenched because as You all agreed Order flow is king And if somebody is a big block builder, if I'm building 30% of the blocks, then everybody who has a bundle, all the searchers will send me bundles.

They can't decide, well, I don't trust Titan, so I don't, no, you will trust, because otherwise you give up on 30% of your revenue opportunities.

So if you're a big block builder, are you entrenched to remain a big blocker?

Order flow begets order flow to a large degree.

It's like you have a position in the market, you become more trusted.

It's sort of like a self-reinforcing cycle, and there's barriers to entry.

Like, most of the builders have to pay you for a subscription, right?

So, there's these like fixed costs that you have to incur to even be in the game.

It's also like, you know, we've seen this with like new builders in the market.

Some searchers find that their bundle inclusion rate with certain builders just isn't up to par.

And as a result, they either consider building or potentially send their order flow to other places.

Would you be willing to pay for order flow today if you could?

And do you think you will end up paying for order flow in the future?

Like if you set up nodes, you're paying for order flow.

Like No matter how you get it, you're basically paying for order

I guess the question is mechanisms around that.

And I guess, does it always need to be exclusive?

So I think the design space for this is complicated.

But yeah, exclusive order flow is really important.

And it could end up being a kingmaker for builders.

I guess not everybody here was necessarily building sort of right around the merge or just shortly after the merge.

Some people were there much faster than others.

a quick question for QB on essentially having a neutral builder and kind of making that profitable, right?

Like you can effectively create an intent layer and allow people to communicate their intent and optimally execute on that.

But how would you then incentivize and monetize that by like taking a portion of the profit if our sink or I mean more concerningly beaver builder could do the exact same thing not monetize it because they have the advantage of getting that additional order flow but on top of that they can then pick off the 20 uni v3 L peoples and profit off that so they don't really have to monetize that side, right?

How do you then compete in that world because they don't depend on the suboptimal?

Execution of transactions like sandwich searches do they don't really care like how how do you solve that?

I think in general, they don't control 100% of the block space, even though there are a few players that dominate certain short-tail strategies, like Beaver and others.

There is still an interesting enough amount of block space that is open for neutral builders and even And also the reason why we are we are building out these services for other trading firms to actually compete with these larger trading firms, but are not vertically integrated, they still generate valuable enough flow.

And so that is the position or the value add of a new firm.

I think you also said something really interesting earlier about identifying what are the benefits that other builders are getting for this vertical integration?

How can we turn that into a product and how can we offer it to searchers?

So Max has some great research that says the sex-tax arbitrageurs win blocks when there's high degrees of volatility on Binance, right?

They just need to win the blocks that they need to win, right?

And again, this is this sort of these negative cycles that, you know, you can have these external forces that exert pressure on the network because of the monolithic block auction.

We need to move to partial block building so that the guys at Beaver who are doing sex-dex arbitrage, hypothetically, can win that piece of the block and then others who have other applications.

build an Ethereum for everyone and everything, but it's increasingly an Ethereum for a very elite group of traders, and that's not great.

So I know those sex-dex arbitrageurs are looking for low latency with the node operators to get it in last possible second before the block is proposed.

I'm wondering if relays pay for co-location with node operators to get lower latency.

I don't think anybody pays for co-location with nodes.

We don't, and I don't know of others that do.

We invest heavily in propagating information at the speed of light around the world.

So it's kind of like it's because nodes could be anywhere It's not like well you have 8 centralized exchanges around the world right which next to which you can low co-locate if you're in trad fi But here It's much bigger.

It could be in all sorts of data centers around the world.

So for us, it's a big investment in infrastructure and speed of light propagation.

And Justin came up with the idea of optimistic relaying, reducing, shaving off milliseconds and whatnot, et cetera, in order to be ready at the last possible moment.

But using that as an opportunity to say, latency games still very much exist, whether you like it or not.

So you could imagine that CFI, DeFi arbitragers, who are the strongest actors out there, will have the lowest risk, the closest they are to the execution.

And really wait till the last possible moment.

If you can shave a millisecond, then it's slightly less risky, and you make slightly more money.

So these games still very much exist at the last moment of building blocks and trying to improve them.

Don't know if any of the builders here have something to add.

I don't think latency games are necessarily a bad thing, because they are primary value drivers to end users.

It's just a question of, can you encourage the benefits of latency with appropriate guardrails for users?

Can you give them guarantees around decentralization, censorship, or whatever?

I think that's a better approach than saying, decentralization at all costs, and we'll solve customer problems later.

I would also say we actually like arbitrageurs.

Like if you're here, you definitely have to be like yourself.

And we want arbitrageurs to correct the price at the fastest like possibility.

A lot of people say, well, latency games, that's tried-fi, that's bad.

There's a lot of good things about latency.

There are people getting sandwiched, which is a bad thing.

On the subject of co-location, to Uri's point, the validator-proposer can be anywhere in the world, so it's hard to co-locate.

But there's a pretty small number of relays, like 7 that matter.

And co-location from builders to relays is a big deal.

But everybody asks where you are, because they want to have lowest latency.

And 1 of the areas that we at Blocknanover are working very hard on is basically distributed relaying.

So we cannot have 1 relay in 1 geography, but many relays all over the world.

The great thing about this is it's really fucking expensive.

And the other great thing about this is we don't get paid anything for it.

So, you know, it's awesome that we get to finance the operations of Lido and other folks, because that's super sustainable, and we're excited about it.

Well, I use Lido as an example, and we love Lido.

So if you look at rate, there's many dashboards out here on what we do.

And if you look at rated network, it says quite clearly that the value of an MEV boost block on average is between 300 and 400% of vanilla block.

So the primary beneficiaries of the supply chain are the validators who get a higher return for block validated and therefore offer a better APY, which is great, we all know that, right?

The challenge of course is that there's a whole bunch of layers in the stack that have nice names but ultimately are suckers because the economics for the value flows downhill in the supply chain and that therefore you have many actors who don't have great economic incentives.

And therefore you don't have high degrees of participation and therefore you have counter incentives and therefore you have centralization.

So anyone who's in the relay world knows is tired of me hearing about this, but today we have 7 relays that matter.

And so tomorrow we're either gonna have 2 relays that matter, or we're gonna have 25 relays that matter, right?

And which version of that future do we think is better for the health of the network and what incentives are we putting in place to encourage that version?

And the current incentives today are no incentives and therefore why would anyone stand up a new relay?

Because they're just going to lose money.

I guess there's some benefits of synergy with vertical integration, right?

You'd expect there to be lower latency between builders and relays, and that would benefit the relay.

But that doesn't seem like a good business model to encourage.

See all Frontier Research transcripts on Youtube

MEVday Paris: Searchers and Builders