Fireside Chat with Henri Stern and Kartik Talwar

[Music] All right, we're going to jump right into our second session for today. And for this one, I'd like to bring on Henry from Privy on stage. So, please give Henry a big round of applause. [Applause] Hey, Andre. Hi, So uh anything uh big happened this year for you guys?

Um a number of things I I think it's been a very exciting year. Uh last few months uh a lot of my attention was on uh the work that we've undertaken with Stripe. So we we got acquired recently uh and we are now uh you know crypto team working at Stripe and doing the things that we do but with just a lot more muscle uh behind us. I mean, awesome and congrats. Um, you obviously kind of just said that nothing really changes for your day-to-day.

Uh, but what are kind of the things that you now get to add in from the Stripe Network? Yeah, I mean I guess jumping into the things that don't change first, one of the the the things that got us excited in the beginning when when um uh Stripe reached out was was the question of what would it take for us to think this is a really good place for us to do work and and the the sort of core three tenants were one uh we want to remain uh fairly autonomous. We want to be uh moving at the pace that we're moving at. we will stay in our office like very uh lowbrow things like that but that end up making a huge difference. Um, two, we have to be able to work with everyone across the space as we have.

I think obviously everything that's happening around stable coins is extraordinarily exciting. There's on the back tail of that like a lot around tokenization of other assets, but I think beyond that uh you know I come from uh protocol labs. I think a lot of web 3 is very important to me. I think there are a lot of other aspects to the space that matter deeply and it would be such a miss to lose that focus that we've had. And so I think we've we work with a lot of fintexs, we work with a lot of uh trading platforms, we work with a lot of consumer products and that doesn't change.

And then I think the third was um we need to be able to work with you know stripe competitors, bridge competitors like ultimately the space is so nent there is so much exploration and experimentation to be done across the board and we want to be able to work with any party that's doing something interesting and frankly that our customers want to work with which is the main part of how we drive you know any partnerships that we do. Um the things that do change I think really are maybe twofold. First from a you know purely uh logistical you know muscle standpoint we just have a lot more resources to play with and that comes in sort of two flavors. One um Stripe is a lot of how we've modeled building our own APIs and our own developer tooling. They know a lot.

it is a gold standard still

and so yeah and so we want uh you know a lot of resourcing and they have offered to to help us on a number of ways in terms of you know everything from um API semantics to uh scaling infrastructure to uh you know continuous hardening of security surfaces which is vast majority of our work. Uh and so across the board that just means we uh have our team but we have uh a really leaned in basically massive engineering and product team at our heels sort of pushing us forward which is really great. The second is you know I think uh wallets a lot of our work is in sort of you know uh so so privy for those who don't know builds embedded wallets. We build basically digital asset accounts that you can embed into your product directly. So if your user already knows about crypto, has an existing wallet, they can bring it with them.

And we help support basically connectors into those thirdparty wallets. But if the user has never heard of crypto, doesn't care about crypto, and simply will get access to assets as part of the product experience, we help set them up basically or set the developer up with key management that enables them to spin up self-custodial wallets in their product. And um a big part of that is basically then connections with the rest of the money movement space. And so that's the second big thing is ultimately um how do these digital accounts work as part of bundles specifically with things like uh money movement, payment methods, stable coin orchestration, KYC, a lot of these areas that we've just seen are part of what our customers need to get off the ground really quickly and build global businesses on crypto rails. So that's the second really interesting

unlock you get you get direct access to that rail and u enable that on on day one. That's amazing. Um, has that changed how you think about product now or was that already part of the plan and you just now basically got there faster? I

I think it has. I think in in in two ways. First, you know, we're a startup and so in a lot of ways a lot of our thinking on product was always based on focus. We don't get to do multiple things. We get to do one thing, maybe two if we're lucky and we have to do them really well.

And I think most of our work is driven by customers and most of our work happens you know we have yearly goals and that's typically a sentence I would like this to be true at the end of the year which translates into quarterly goals to actually get there which translates into sprints and then a lot of it is basically based on on our partnerships our customers and what they need us to do. Um I think the first difference is we do now have some call it multi-year plans where uh beyond privy you know stripe is an extraordinarily uh driven hungry but also humble home that is very excited about this future that hopefully is being built um and that wants to be a big part of it and I think um understanding where we can kind of put spin on the ball on things happening at Stripe to help influence product decisions there in ways that we think will benefit privy customers down the line is the first piece. So actually interacting with basically teams that are working on longer tail projects outside of area but where we think there will be like really incredible synergies down the line of this is something that we can actually really use for privy customers and this is the shape that you know a given neo bank would want for this product so that we can work with it really well when uh integration time comes or something like that. So that's the first one and then I think the second one is uh really much more looking at the sort of you know entire arsenal of tooling at our disposal. We have a much deeper bench of product capabilities that we can call upon to build the right solutions for our customers.

And so that's had us kind of broaden our thinking in terms of you know how can we help these folks? How uh much deeper do we want to get into uh fiat to crypto conversions? What can we do to help with you know uh licensing or KYC? these questions that again were before a little bit out of reach for us.

That's amazing. Um I mean I've kind of seen Privy from kind of the beginning to kind of where everybody is is now. Um and the interesting part has been the users that have been the customers uh for different websites that are using privy for for the login and now you kind of have a whole variety of uh categories and sectors of people using it. And the question I have for you is I guess in 2025 are there any kind of obvious patterns that you notice on like people who are building kind of cryptonnative products and how they think about this versus traditional non-crypto companies who are also integrating privy. How do they kind of treat the product?

How do they treat the the solutions in general or what's kind of uh obvious when you kind of talk to both of these segments?

Yeah, I I think broadly there you know there's a number more but there are two core archetypes that stand out and the first is the cryptonative team. These are typically smaller teams. They have uh slightly more like homogeneous stacks. So, you know, we started for example with uh React library that um had a number of presets that worked for crypto teams but um that uh for example gave them a lot of control over some of the guts of like blockchain interactions being able to configure slippage, gas prices, etc., etc.

Um today we work with call it you know fortune 500 like fintex these are thousand person engineering teams uh usually not react based uh they want direct access to our API they want a lot more a lot more flexibility across a number of fronts related to authentication UX and other pieces conversely they want a much deeper abstraction when it comes to the cryptogs they want a lot more of the decisions having been made for them with regards to uh what are the right ways of uh bridging that gap app between off-chain and onchain in a way that enables their users to use chains directly and they're very sensitive to different things. So I think a lot of our work as we've seen this develop over the last two years has been taking the sort of crypto system that we've built uh and uh unbundling it so that we could offer sort of deeper hooks into the guts of the package to uh enable different customer types. But ultimately we've found that serving both sets of customers is extraordinarily self-reinforcing in a sense because the the cryptonatives so to speak kind of show us how to use the tech whereas uh uh fintex today at least help distribute it on a much larger scale. So

so are you saying that effectively you're still building the same product and you're just exposing different stacks of it or are you actually now have to kind of cater for two different audiences and two different sort of subproduct decisions for everything that you ship? Yeah, I mean this is an interesting we we we I think we this up and and corrected it, but basically we started assuming the latter, which is we're going to have to build basically two different products. It'll be the same underlying tech platform. And we launched this in December of last year, which was like a different wallet type that we thought would be more uh appropriate for

Is that the embedded wallet?

This was like server wallets. We made this and and we basically realized that the distinction like our assumption was like the ven diagram overlap of the customer segments is like basically two mostly separate circles and it's actually just not true across so many different like facets of the product. It is a complete overlap of the circles and um what we've moved towards it is it is actually the same product but you can pick where in the abstraction ladder you want to dive into and I think one of our core strengths you know I I uh I I see it when we talk to customers who have been sort of fudded about privy in the past and part of the FUD is well privy serves a consumer product why could it be right for you as a you know multi-billion dollar asset holder and I think the reality is that actually uh sort you kind of temper infrastructure by working with solutions across the board. So we work with you know consumer products who have insanely spiky traffic coming from uh hotspots across the world and so that's allowed us for example to build server collocation in Singapore in SF in Frankfurt and that means that um we have really good latency across these places because a game had to be able to be responsive for you know a spike because some you know I don't know Korean uh blog wrote about it once um and then conversely we work with neo bankanks that are holding these assets which means that we have to have modes of operations for a product where uh the product runs policy engines, the product has MFA on any transaction, you have a key quorum of you know MFN users needed to sign to actually do a transaction. And so all of these pieces sort of strengthen each other, but the key and hard thing is to make sure that uh a developer can quickly get to the product configuration that they need for their users.

I guess one thing I'm kind of um and and this is actually something um as a full-on separate followup we we never discussed before. Um you effectively are in a position where you can given the distribution and just the defaults that you can present to different customers who have massive distributions, you kind of get to influence the future of wallets altogether. Um, do you actively kind of think about that or I guess how do you kind of foresee the the wallet ecosystem? Do do you feel like you're a participant or do you feel like you get to actually dictate some of the decisions that somebody should or should not make? We're seeing embedded wallets.

We're seeing smart wallets being kind of reintroduced all the integrations like how does the world change um according to you over the next kind of couple years? It's a very big loaded question but curious to get any thoughts. Yeah, I I mean I think we are a participant and as a participant we are a driver uh you know modulo the success of the customers that we get to serve and so I think a lot of our focus is how do we make our customers as successful as possible and I think a lot of our sort of product demo has been uh we exist to enable developers to build better products secure products self-custodial products for their users uh but beyond there our agenda is not to push a given technology or a given way of doing things. Our agenda is to help them build, you know, experiences that can delight their users and enable them to do new things with value and ownership on the web. Um, so so we've kind of taken a sense of we have things that we'd like to see and this is where the defaults we introduce hopefully again put a little bit of spin on how you should use these things by default.

For example, uh this is a spicy take, so I'll give a spicy take, but uh we we thought a lot about how should we present wallet connect to our customers and I think wallet connect is an extraordinarily important part of the stack. They are the lynch pin to so much of how uh connectors work today. And they've done an incredible job setting out standards that basically enable a very fragmented wallet space to be usable for developers. And yet our thinking was, you know, if somebody's never heard of crypto and they're using a wallet that they, you know, don't know about, and the experience that we're telling them to go through is click connect, which is already a little weird, then click wallet connect, which is a protocol you've never heard of, and then go down and find the name of that wallet that you maybe were looking for. There is no world in which they're going to get there.

And so the decision was

switch devices and part of that process.

Yeah. And so the decision we made, it's it's it's stupid but but we kind of sweat it is um let's have it be more wallets like privy itself by the way but this is not us like uh so so we we we kind of hid the underlying protocol in the same way by the way that privy as a brand we don't have a consumer wallet we simply serve our customers so that they can build onto cryptostack and so these are I think kind of uh first principles that we have about the way crypto usage will evolve in the world that I guess we are pushing through the ways in which we built our abstractions Um, but you know, ultimately if a customer says, "I really want to show this," then we're like, "Okay, it's your product and we're here to to let you do it." So, to actually answer your question, I see embedded wallets as a growing part of the ecosystem. I think, you know, that the analogy I like to use is I I I think of wallets as as digital identity passports. One of the things they allow you to do is hold assets.

I'm a huge wallet user across, you know, a number of of of wallets traditionally. Um, but also sometimes it's just too heavy to bring your own wallet in experience. If I want to trade, having to like confirm a transaction every time. Um, if I'm on mobile, it's a pain in the ass. Like there's a number of places in which it doesn't work.

So I I guess the the two core takes I have is the idea that a user would have a single wallet seems to me very misguided if only for privacy purposes. That seems to not be the way anyone operates in the world. Um,

the multi-chain world you got to factor in.

Yeah. Um and then the idea that um every user will want their own call it personal wallet which is to say these are the traditional you metamask phantom rainbow that you think of also seems to me like not the way the world is headed. I think basically people will get into crypto through experiences through the apps that serve them or give them access to these rails and then at some point they may become mature enough that they decide to say hey I actually want to hold this in a uh container that is fully built around the crypto experience rather than as part of another experience that I'm going through and so we see ourselves as very complimentary to traditional wallets. I think both will continue to grow massively. But the analogy I use is I, you know, people don't get a passport to feel American, to feel French, to feel whatever it is they feel, they get a passport to take a plane.

And I think our goal is what is the fastest safest path to getting you on the on the flight. And then I think from there, if you're traveling a lot, maybe it's you should get a passport and maybe you should do these things. But that's kind of how we're seeing things evolve. and then a lot more internet connectivity between the wallet and the rest of the infrastructure being built across DeFi across you know various fiat money movement rails uh and so on and so forth.

So you got the global entry game here.

Exactly.

Um so two two questions and uh we obviously I don't have a lot of time so I'll I'll see if um I can make these quick. It feels like the world is headed towards manage wallets for the mass audience. Um, how do you think about that premise and also just a security of this where there may be more than one person who effectively has the ability to uh to uh trigger events and uh and kind of manage your funds instead of you owning the the keys?

Yeah, it's a great question. So, we we use manage wallets in a certain word and I think like you know they're broadly three dirty words in crypto. Uh decentralization, trustlessness and custody. uh custody self custody whatever kind of custody like you you based on the marketing you're trying to posit you'll you'll you'll spin it to your needs which is true of all of these words and all of these are very nuanced and I think words that require a lot of precision um I we define managed wallets as custodial wallets so that's that's our our definition I think the way you use managed wallets

I was referred to as custodial as well

okay I I think um so I I'll give an example something we saw we we built out a policy engine uh because of agentic trading platform forms. I want to give an LLM access to my wallet so they can trade on my behalf. Um, and I'm worried that a hallucination in the LLM will completely wreck my portfolio. And so I want a set of policies that limit what the LLM can do. So we build this for this use case.

And then, you know, it turns out it's extraordinarily useful for NEO banks and global remittance products who want to, for example, limit a wallet to only trading USDC or something like that. Um, and so this is part of the, you know, we see this virtuous cycle of of the customers that we serve. um the LLM needs to be given access to your wallet. So this is you know your self-custodial wallet but you have now created what we call a scoped signer but you can call it whatever you want. It's a permission entry point into the wallet with certain permissions that you've explicitly given the LLM of I will enable this LLM to trade up to X with these contracts.

For example, I only want it to be the in swap V4 contract um and only this much like per day or something like that. um ultimately you remain the sole party in control of the wallet in that you can always revoke that permission and you can always basically you know export your keys and leave the system. So I think the way I think about it uh which is maddeningly nuanced and I it's a very complicated topic but the way I think about it is in terms of like intermediation ultimately I remember the old like you know bitcoin paper wallets

you're still delegating in this flow

and and with meta mask this is not to make it like uh nihilistic argument that like everything is the same. I think these are extraordinarily morally veilent questions. But with MetaMask, um, you are being intermediated by the client and the software that is shown. There is a swap button on there and MetaMask has chosen which swap rails you're going to be using. In fact, monetizing these swap rails.

Um, there is a send button and they've decided to, you know, make it ETH first rather than ETH and Salana even though they're Salana now. like we are all like the victims or beneficiaries of a hundred product decisions that have been made by the infrastructure providers that are there and the way I see our work is to basically give a little bit more control to the app developer or the product developer who wants to utilize these rails so that they can make decisions because for example maybe a pick Robin Hood a Robin Hood user doesn't actually want to like you know see the gas that they're going to be spending on like each and every little thing we

or Robin may want to sponsor it for the users. Yeah. And so we're kind of removing agency from the user in enabling that sponsorship because maybe the user wanted to pick a different like setting for gas or wanted to like use a different like node provider, but a lot of users don't. And so, uh, I see the wallet I I think there are going to be really hard lines that have to be drawn around this sort of spectrum where there are places and this is a lot of what we work on is making sure that we don't introduce food guns into the experience where the developer is given so much power that they can put the user at risk and where the user uh has to give sort of proactive consent to the things that they're enabling the wallet to do. And so I think this is where we need to sort of continue to draw bright lines so that we have definitions of this is a self-custodial experience, this is a custodial experience and so on and so forth.

And I think a lot of the next few years are going to be about that. But I think beyond this um what we're doing is we're kind of opening up the design space so that you can really tailor the experience to the audience that you're serving. I mean, the one thing I I'm hearing which I'm really happy to hear is that you are fundamentally saying that you believe in modularity and effectively you get to kind of toggle these settings based off of a users and a developer's preference and ultimately they get to combine the stack and do the combinations and that's a much better route than enforcing defaults or simply just limiting a few customizations.

Yeah. And I guess this is my my my my beef with the crypto industry by by and large is is like you get the two sides of the coin which is extraordinary tribalism and uh to some extent a lot of paternalistic instinct of this is how you use this tech and no other way.

That's a really good way to phrase it actually.

Uh and and so I think we are trying to be uh uh responsible on how to set these things. We set a floor with security. We use you know a mix of tees and tremir secret sharing to ensure that we have the ability to uh uh give the developer a lot of flexibility based on what they need and give the user a lot of guarantees with regards to how keys are managed and the fact that they were the only ones in control of set keys. Uh but ultimately how you should use this technology should really be up to the people who benefit from it not to you know uh the folks who read a paper ones that really caught their fancy. No.

And um I think that's exactly how it should be too because that's how you actually make good products, not not just going to showcase tech that is meant to do something that may not actually come to uh come to fruition. Well, this has been a wonderful chat, Henry. I want to thank you for taking the time and uh please everybody give him a big round of applause. Thank you. [Music]