Ben Chow - Jupiter Exchange Co-Founder, Ep. 2

Ben Chow (Jupiter Exchange Co-Founder) joins The Zeitgeist to discuss Jupiter's role in the Solana ecosystem powering swap infrastructure.

Show Notes

0:44 - Intro / Background
05:04 - How does Jupiter work?
08:44 - Jupiter integration in Phantom
13:26 - Keeping up with the new liquidity innovations
15:47 - What's the process of Jupiter Integration
21:18 - Who is using Jupiter?
24:44 - The Growing Solana Ecosystem
26:04 - What is the vision for Jupiter?
28:44 - Favorite Builders in the space
32:28 - Contact info

References

• Saber
• Orca
• Raydium
• Friktion

Transcript

Brian (00:06):

Hey everyone, and welcome to the Zeitgeist, the show where we highlight founders, developers and designers who are pushing the Web3 space forward. I'm Brian Friel, developer relations at Phantom, and I'm super excited to introduce Ben Chow. Ben is the co-founder of Jupiter, the largest decentralized exchange aggregator on Solana, powering tens of millions of dollars in swaps every day. Ben, welcome to the show.

Ben (00:30):

Thank you for having me. I'm super excited to be here.

Brian (00:33):

Thanks for coming on. And so we have a lot to cover today, including an exciting announcement between Phantom and Jupiter. But before that, could you give us a brief overview of who you are and how you came to be involved with Jupiter?

Ben (00:44):

Sure, yeah. So hi everyone, my name is Ben. I'm one of the co-founders of Jupiter. I became involved in Jupiter actually through a few good friends of mine, Megan and Shawn. Probably unlike many people, I am relatively new to crypto, although at this point in time, I feel like an ancient person. The world moves. I think I've gone through a couple of cycles and a few different crypto projects, so it's been an amazing roller coaster ride of sorts.

Ben (01:13):

But as for my background, I'm a ex-IDEO product designer and longtime serial entrepreneur and startup founder. I've worked in a number of fields to be honest. Probably the most successful ones was that I was actually an early social gaming pioneer. I helped build a social gaming startup that was True Venture backed and later got acquired by Disney played him in a roll up.

Ben (01:37):

More recently, before Jupiter I was the co-founder CEO of a top 100 social app on the iOS store, and I also happen to have several patents and a lot of work in the medical product space. Probably the coolest one was I did an augmented reality surgical navigation system for cataract surgeons that we took to market, which was very cool.

Ben (02:03):

I would say the core team on Jupiter is pretty experienced. Megan and Shawn are probably the OG DeFi folks. They were a founding team or strategic advisor to InstadApp, RAPBTC, Kyber, Blockfolio, and a number of others. And I feel grateful and fortunate that they looped me in when they were thinking about building on Solana. So we're a pretty veteran team. We've all done a number of companies, a number of products. I feel very thankful that Solana is both an incredible ecosystem to build on, and this is an incredible time to be a builder. There's just so much value you can give because we're all super early, but also everyone is really smart, very talented, really great to work with. So it's just been an amazing time.

Brian (02:53):

So you have this wealth of startup experience across a ton of different domains. Was Solana your first blockchain ecosystem that you really sunk your teeth into and got involved with?

Ben (03:03):

Yes. Solana actually is my first. I don't have the stories that people have about Ethereum or Binance, and I don't have as much background in those. I really only know Solana very well. At the same time, I've also had to come up to speed on DeFi and crypto in general. So I have this really interesting depth of knowledge in a certain area, but not as much broad experience. Like when people talk about 2017 and the crypto winter, I didn't have those kind of experiences. I can talk shop now a little bit more around aggregation indexes, concentrated liquidity, derivatives, perpetuals, all sorts of stuff. I think one of the things that attracts me to crypto is that for any of the things I've always done, I just love to deep dive into a problem domain and learn as much as I can. And crypto is like this endless deep well of things you can learn.

Brian (04:02):

Yeah, the rabbit hole.

Ben (04:04):

Yeah. Literally, probably the deepest rabbit hole I've ever been in. And I've been in several. Doing surgical navigation systems are like this combination of understanding the business model, the user experience, the biomedical, the biophysical, every single layer of it is needed to really do something very interesting and cool. And the same can be said even more so with crypto.

Brian (04:33):

No, I agree a hundred percent. It's an endless learning experience. And so you mentioned your background. Back in the day you were at IDEO as a product designer. Most people, I would say, know Jupiter today because of the website you guys host at jup.ag. You guys have a really slick few clicks and you're swapping the best rate between different tokens on Solana. But you guys also offer a couple other ways to interface with Jupiter. Can you speak a little bit to how Jupiter works under the hood and how both users and developers can interact with what you guys have built?

Ben (05:04):

So we have an SDK and a API, and our vision for how we see Jupiter and how it relates to the rest of the ecosystem is that we want to be the swap infrastructure for all of Solana. And the single place that any project developer or user can easily access all of the liquidity and discover the best price trades for any token that they want to trade on in Solana. What that means is that if you're a developer... And I would say one reason why we've gotten a huge traction on developers building on top of Jupiter is that we make it really easy to do not only trades, but arbitrage trading. So if you're using our API, it's actually one of the easiest things you can do because it's language platform agnostic. It's literally one REST API call to get the best price routes, and then another API call to also get the transaction objects you need to send to actually do the swap.

Ben (06:12):

On the SDK side, it's a little bit more involved, but it's about three or four steps. You just have to load up the library and then put in your trade that you want and compute the routes. Right now, computing all the routes is done off chain and offline, and the reason for that is that we integrate currently 15 DEXs on Solana. And what that means is, basically we connect and route every single market there is on Solana. So you can think of all of Solana's liquidity as one big virtual DEX, and we have to create the routes for all that. That means that there's tons of routes, thousands and thousands of routes. And calculating the best price between all of them is actually computationally expensive and something that is very very hard to do on-chain. And so that is done in our SDK off-chain, and our API also leverages our SDK. So whenever we make improvements to our SDK, our API automatically gets them.

Ben (07:13):

There's a lot of work that happens on a daily basis. One, to optimize the routing algorithm, to make it more performant. Because as you can imagine, anytime you increase the number of tokens, and there are thousands of tokens on Solana, anytime there's new liquidity sources for those tokens, the complexity grows exponentially. So the number of permutations and combinations of routes grows ever greater, and calculating them in a way that's very fast for users. In some ways, a lot of market opportunities and best price routes don't live very long, right? So the route itself could be best price for only a block or a couple blocks before someone else snatches it. You want to be able to calculate these things very, very quickly. And so we do a lot of work optimizing the computation for it, and then also there's a lot of infrastructure work for anyone using the API to be able to make sure that it scales and is performant for all the people who are using the API.

Brian (08:21):

So I'd love to dig in more of the details, but before we do, we do have an exciting announcement to bring up today. And that is at the time that this show is being recorded, Phantom is set to release its brand new in-app swapper that's powered by Jupiter. We're super excited about this because it opens up a huge wide range of new markets and gives our users the best possible price across all pairs they're interested in trading. Can you guys speak a little bit to this integration and what this means for DeFi on Solana?

Ben (08:48):

So basically I think along with our mission, what it means to be a swap infrastructure, is that there's actually just a lot of work to make all this liquidity and any token available for anyone. What we want to do is make it easy for any project or any wallet to access this liquidity as simply and as seamlessly and as easy as possible. And that is so you can support the rich number of use cases there are for swaps, which could be streaming payments, being able to, if you're streaming a token that the recipient can receive the token in the token of their choice at the best rate possible. This is like purchasing NFTs, being able to purchase NFTs with any token, not just SOL or whatever the collections token is.

Ben (09:35):

If you don't have those things to be able to do that easily, simply in one click to power payments... So Solana payments, we work with MtnPay to help facilitate that. So you can purchase anything with any token that you have, not just USDC, but the person the recipient can receive in the token that they want, whether it's USDC or SOL or whatever, but someone can pay with the token that they want. So facilitating these interactions, that means that the simple thing that most developers do is they connect with one DEX, right? But that DEX may not have the best rates. That DEX may not have all the tokens.

Ben (10:15):

So a lot of tokens don't launch on every DEX. They launch on, it could be a Serum market or, or maybe a Raydium permissionless pool. And so being able to look at where the liquidity is, where the rates are, and make sure that people can easily access that with one API call or using our SDK is really important. And again, as Solana grows, as the number of tokens grow, as the number of projects come along, as a new markets come online, this stuff gets harder and harder, and more so because the latest DEXs that are coming online are actually much more complicated than the constant product AMMs that Solana first launched with.

Ben (11:00):

So we're seeing concentrated liquidity market makers, we're seeing proactive market makers, which are using Pyth or other price oracles to enable their market making. These things are very complicated to deal with and to integrate on top of them, but the reason to use them is that they offer amazing prices. So LIFINITY is one of the DEXs that we integrate with, along with Crema and Orca whirlpools, and their ability to give you competitive great prices for some of the popular token swaps is amazing.

Ben (11:38):

Because they have a lot more dependencies and they're a little bit more complicated, there are issues and there are corner cases to deal with. For example, because LIFINITY relies on a price oracle, in times of congestion that oracle could get behind and not be updated, and thus their ability to market make is hampered. Along with that, since everything is really early, we're usually the first people to work with anyone's SDK. It's usually in the early stages, and so doing things like having a really nice way to handle these errors and surface them to users where it makes sense to them, like, "Why is this something happening?" is an important job, and we're generally becoming sort of the interface between integrators and DEXs in terms of helping people like troubleshoot where bugs are in order to provide the best user experience for the end user.

Brian (12:38):

No, I think that user experience is everything at the end of the day, and that's why we're super excited about this integration. I remember a time on Solana when it was so simple, I could just open up maybe Saber, Orca, Raydium, a few other DEXs and have them all on one screen, and I could just compare different prices. But as you said, there's been a complete Cambrian explosion in different markets and different types of markets. And so you guys are doing great work to normalize that and make that a great user experience for everybody.

Brian (13:05):

Just real quick. You hit on a couple different types of these types of markets. AMMs, the classic example, and then there's concentrated liquidity, Orca's whirlpools have been a huge success recently. Solana's also unique in that we have central limit order books on chain, Serum. And then there's this concept also of professional market makers giving liquidity. Is there any other types of liquidity that you guys are aggregating, and how are you keeping up with these new innovations that are happening?

Ben (13:31):

There are a lot of interesting things going on. Probably the next set of integrations that will be more interesting for folks is that we are starting to integrate directly with minting and staking programs and not just liquidity pools. And the benefit there for folks is that normally these routes will have either no to very low slippage. So if you wanted to, for example, go from SOL to mSOL and you wanted to convert a million dollars worth of it, usually if it's in a pool, there's a great deal of slippage. But with these direct minting routes, there's almost none. Also, a lot of these minting, staking routes will have zero to low fees as well, because you're working directly with the protocol itself and there's no need to... A lot of times the fees are an incentive for LPs to deposit their liquidity and give them some incentive. If you're talking directly to the protocol itself, there's no need to take fees per se. So they could be zero to low things. So there's some really cool advantages.

Brian (14:47):

And you guys don't add a fee also, correct? You guys don't add a fee any of your trades.

Ben (14:52):

We also take zero fees, yes. We take a very long term view with Solana. We think it's super early. We want to grow the volumes in Solana. We want to grow Solana itself as well. And also we're very focused on making sure we deliver the absolute best price swaps to you, and charging a fee on top of DEX fees is not in our interest. We're not really here to extract value, more to give value to end users, developers and anyone that we work with.

Brian (15:25):

I think we're very aligned on the long term vision and the potential of Solana there. So what is the ultimate process like for integrating these new markets? I imagine, given the scale of your guys' infrastructure, it must be a fairly manual process to review these and normalize different markets. Is there a plan at all to make this in a way in the future where this could be a semi-permissionless thing, where people can add their own markets to Jupiter, or how complex is that integration today?

Ben (15:52):

For the most part, we've automated everything. So it's really not possible to manually check everything. If there's liquidity on Solana and a token in the token registry, we have it listed on Jupiter within minutes. And that's because we built out the infrastructure to automate everything as much as possible. So for instance, what that means is that for any DEX we aggregate with, we ask if they can provide an API on their markets that they are adding or removing. But because it's early days, not everyone has an API, so we have to do very interesting things, like if it's an open source SDK that the DEX has, we'll crawl the GitHub repo to see what markets they've added to the SDK. Whatever we can do to automate, we have done.

Ben (16:39):

Part of the promise that we have for Jupiter is the best token selection. So that means if there's some hot token that launches anywhere on Solana, that we've a DEX that we've integrated, we want you to be able to trade it instantly. And that's reason why any project can integrate with us. So if you integrate with us and you're a wallet and all of a sudden, some token blows up, you know that if you're using Jupiter, we've got it, and we've got the best price trades for it. There's a lot of infra work that happens on a daily basis. And again, I don't think anyone should have to replicate this work. So the goal for us to be infrastructure is that, look, you shouldn't have to worry about new tokens. You shouldn't have to worry about performance on routing or where the best price is, or what these dependencies and errors are from these DEXs.

Ben (17:34):

If you use Jupiter, our goal is to abstract away a lot of these things. So for you, it's just, "Okay, I want to trade this token. Give me the best route. Let me trade it." Now, we're still early. So I would say there's a lot of work to be done, again, not just on our part in our SDK, but for us as an interface to other DEXs and helping them discover what their bugs are, helping them have the right way of surfacing those errors to users in a way that makes sense to them. And then maybe giving users and integrators a choice of what kind of things that they prefer.

Ben (18:15):

For example, we've been learning that not everyone cares about best price. There are some people that are more like, "Look, I don't need best price. What I would like is like a route that would fail less often." So for example, Serum's liquidity being an essential limit order book is not very predictable compared to an AMM, right? So if you have tight slippage settings, as a trader, your trades may fail more often, whereas you'd really just want to make the trade. And so some folks may be like, "Well, if the best price is on Serum, I'd be okay with maybe not the best price if it's on a constant product AMM or something that I know if I do the trade, it's more likely to succeed." If you're one of those traders, we want to give you that option.

Brian (19:07):

Yeah, that's great. You know, you've hit on a lot here. One, just being the sheer complexity of what you guys do, given that just the exponential number of new routes that are coming online. You guys also, I believe you allow for multi hops between different routes, right? So it's not just, if I want to go SOL to USDC, I have to go find that pair. You can find an intermediary, same mSOL and then go to USDC. Is that right?

Ben (19:30):

Yes. So we support multi hop routes. Right now, it is up to two hops. And the benefit there is that one, if there's no direct market, you will still be able to swap between the tokens that you want in one transaction. So you don't have to do it manually in two or more. The other benefit is that it just so happens that you get the best prices by hopping through an intermediary. Because in Solana, the gas fees are so low, hopping through markets is very cheap. And a lot of markets that are either highly volatile or have some price inefficiencies allow you to route through them in order to get more token, really.

Ben (20:18):

Part of the complexity is actually just finding all these routes. So if you move to two hops, now the number of permutations is even more. And I would say, actually, one of the things that we're working on as a future roadmap item is actually to go beyond two hops. So more than two, say three hops. That's going to allow anyone to access even better price savings and, of course, increase the computation load. But what we're finding ways to optimize on that.

Brian (20:48):

So you guys have a wide range of users who are using this product. You hit a lot on the users who want the best price. This is probably your everyday retail trader like you or I who just wants to swap between different tokens. You also hit on the trader that just wants to be able to fail less. I imagine you have professional traders looking potentially to arbitrage via Jupiter. And then you also have this incredible SDK. I saw you guys recently announced you have over 42,000 downloads on this thing. Can you speak a little bit to who has been integrating these SDKs? Are these power users? Are these other dApps on Solana? And who is your kind of ideal consumer for these developer tools that you're building?

Ben (21:26):

So actually since that announcement, we're now at over 52,000 downloads.

Brian (21:33):

Two announcements on this podcast. I love it, breaking 50k.

Ben (21:36):

Yeah. So, you know the surprising thing, I mean, we have a lot of different folks who use us. I know of VCs who are using Jupiter to use them for arbitrage trading or playing with arbitrage trading. But the most surprising one is actually new developers, literally developers who are new to developing, and they are new to developing on Solana. And I know this because we spend a good deal of time supporting them and trying to help them out. We've lowered the bar for people who are interested in arbitrage trading. And so now it's just easier for, for folks to get more involved and actually learn to code. People are learning to code, so that's been a surprising thing to see. And there's nothing like free money really to draw people to learn to code.

Brian (22:32):

Right, to incentivize a new skill.

Ben (22:34):

Yeah. For us, we want to support anyone who wants to use Jupiter. And probably the more interesting ones are projects that have really interesting use cases. Obviously wallets are an amazing use case for us because you're at a really incredible touchpoint, a very close touchpoint with users. We're really excited about payments. I think while payments is early, it's got a huge potential. And letting anyone buy anything with any token is just super cool. And it doesn't have to be payments. It could be anything. I mentioned NFTs before, but we have a few projects that are doing stuff to make it easier to deposit into vaults. So Friktion uses us because a lot of their vaults take one type of asset, but the user may not have that asset. So why should they come to Jupiter swap, go back to that thing?

Ben (23:31):

By integrating us, their user gets to stay more on that site. It's fewer clicks, it's a quicker transaction. And because they're using Jupiter, and we want to make Jupiter work for our partners, we allow any partner to easily add their own sort of platform fee so that swaps can be a revenue source for them. And so actually, Friktion, I think they've done a good bit of volume just by adding Jupiter and having that as a revenue source. And there's a few others too. Like some DeFi Land was an early launch partner with us. So if you're swapping anything in their game, you're using Jupiter. We're starting to see some more interesting things come out of this.

Brian (24:15):

Yeah, I think this highlights two interesting things about Solana in particular. One, the composable nature, obviously. If you're doing something in Friktion with your portfolio management or you're playing a game with DeFi, it's as easy as just adding another instruction essentially, to swap into Jupiter. But then you also hit on another piece that I think doesn't get highlighted enough. This dovetails nicely with your story, that there's a lot of folks who Solana is their first experience at all in crypto, or as you highlighted, maybe ever even coding, which is, I think a really neat opportunity, bringing in new people with new perspectives on what this technology can do. It's pretty great.

Ben (24:52):

I really see Solana as the practical blockchain, right? It's the one that's going to be accessible for most retail users because it's cheap and fast, but also it's just got a great developer ecosystem of people building games and NFTs that are also just as accessible to the non-crypto enthusiast person. They're like, "Oh, I just want to play a cool game." Or, "I really love this NFT project," or something, "This collection." And to be able to do it with this great, fast, user experience, I feel like it's a hallmark of Solana. It's less about theoretical white papers and more about, "How can I build this great experience for someone to be able to engage with crypto in some way?"

Brian (25:40):

Yeah. Practical and also building something that users actually love and they enjoy using. So we've hit on a lot today. You guys are integrated across tons of different dApps. You have professional traders using you guys, VCs using you guys. Now with Phantom, we're making this super easy for anybody to swap an extension or on mobile on their phone, get the best price. What is the ultimate vision in your mind for Jupiter? Is this something that will always remain an aggregator, and is cross-chain something that would ever be on the horizon for you guys?

Ben (26:13):

Good question. A lot of people ask us this question. Currently, we work with some great partners to support the cross-chain use case. I'm not going to rule out cross-chain. There's definitely ways to do it. We're, I would say, a pretty lean team by choice. We just have really good people and I don't think anyone really wants to scale this into a big company with middle managers or anything. We just really like working with good people who are awesome at a lot of things and just moving fast. And because of that, the things that I think give us pause about cross-chain, at least for me, and this is something that we think about a lot is how to execute well on these things.

Ben (26:56):

And beyond the tech, the bigger thing to really tackle if you're going to do, cross-chain or multi-chain is how do you build a community on all these different chains, and without scaling to huge sizes? And so that's one reason why I think we work with partners, because it allows us to focus on what we do really well, and it allows us to focus and invest in the Solana ecosystem and the users and the community in Solana. We've got an amazing community, and it's because we spent time with them, and we invest in them, and try to give value to them. And I know the other ecosystems and communities on the other chains are very different. And so I don't know if we have yet a model for how to be very successful in those other chains and staying, operating the way we are currently operating.

Ben (27:49):

It doesn't preclude us from... It could be we figure out something, some way that really works for us and go after it. But for right now, we've got some great partners and we'd rather work with them. And that allows us to focus. I can't say this enough. I think great teams focus, because there's a lot of attractive, shiny new opportunities out there. And what would it be if we weren't relentless in our focus to give best price swaps for people, to make sure that we're always integrating to make sure that we're always improving the performance, to make sure that we're solving these issues. If we went off and chased after some shiny new thing, we really wouldn't be infrastructure. There's a lot of projects and users who depend on us. And so we want to make sure we can continue to deliver on that promise.

Brian (28:38):

Well, I think you guys have done an incredible job so far delivering on that promise. I know you guys are very highly respected in the Solana ecosystem, which brings us here to our final question we ask all our guests. Who is a builder that you admire in the Solana ecosystem?

Ben (28:53):

Besides you guys? Can I plug you guys? I mean, look, I'm going to plug you guys because, look. We're big fans of Phantom. I'll give another one, but I do want to say, I'm really personally very excited about our integration and work together. I think you guys have built by far one of the simplest and best wallet experiences out there. One reason why I'm really excited about our partnership is that I think you guys have taken that same simplicity and applied it to Jupiter and our integration, so not only do your users get the power and benefits of Jupiter's ability to get you the best priced token trades, but in a really simple interface that is very accessible for people who are not crypto natives, who don't know what slippage is, who don't know what price impact is, who, who don't know about the various DEXs or how these things work. And what they really want is just to be able to trade from one token that they have to something they heard is pretty awesome or they're really interested in.

Ben (29:57):

So I'm really excited personally for this integration. That said, separately, I will give a shout out to some other folks we really respect as builders. We're big fans of Friktion. I love their vibe as builders. And the new thing for me being in crypto is how important vibe is. So I don't think I've ever said that in any other industry or any other thing ever, but some of our best partners are just people that we vibe with. We get along. We share values around building. We share values and how we think about things. And it's funny to say these things here, but it makes so much sense now. And I feel one reason why I enjoy working in this space so much is that I never used to think about, do I share values with a partner?

Ben (30:49):

It's just more like, "Hey, is this valuable to you? Am I valuable? Are you valuable to me? Okay, let's do something." But now it's not just about are you doing something where there's mutual benefit, but do we share the same long term value? And a lot of this is mainly around, are you looking to extract value from the ecosystem? Are you looking to build value, give value to others and just build and ship? So I think Friktion is a great one.

Ben (31:18):

I've been also a fan of the Orca team for a long time. Also recently, we had STEPN on as an AMA and got to deep dive with Jan, one of the co-founders and have huge respect for that team and their journey as builders. For those who don't know those guys STEPN came about because they failed to raise funding. It was a bad time in the market for them, and so they had the ship. That was it. That was the answer. We can't raise money, so we got to build. We don't have enough time or resources to build a full game. Let's build this simpler thing.

Brian (31:58):

Yeah, the focus there.

Ben (31:59):

That focus and that urgency of shipping and building is great, is really, really great. And so I have huge respect for that.

Brian (32:09):

Well, you mentioned a bunch of great teams there. I know the feeling is mutual on Phantom's side, both between you guys and also you hit on Friktion, on STEPN, and Orca, many others as well, too many to list in this ecosystem. But yeah, the long term alignment, the focus, I think these are all great things that we believe in a hundred percent. Ben, this was a fantastic discussion. Thanks so much for coming on the show. We'd love to have you on again some time. Maybe we can check in on how the swapper's doing. In the meantime, how can people get in touch with Jupiter and how can they learn more about what you guys are building?

Ben (32:38):

Yeah, you can follow us on Twitter. So we're @JupiterExchange. I highly recommend anyone to join our Discord and our community there. We've got a fantastic community. You can join through discord.jup, jup.ag. Usually our community has the first look on any new announcement or any new cool thing that we have. We also are very, very responsive. The team is listening and responding all the time, both to developer and users. And a lot of the things that you see on Jupiter today came from the community, and a lot of the partnerships also came from the community. So we listen very closely and we also have good discussions with folks. And so if you get involved, you will have a very strong voice in where the future of Jupiter goes. So really encourage people to join the community.

Brian (33:33):

I love that. That's a great way to end it. Thanks so much, Ben.

Ben (33:36):

Sure. Thank you.