Why Privacy Will Be the Biggest Moat in Crypto

Transcript

People claim that users don't ultimately care about privacy. And I think that may be true when it comes to Instagram and Facebook and other social media, but I don't believe it is true when it comes to finance. In order for crypto to actually become mainstream, it's going to have to have privacy. And so then that creates a self-reinforcing feedback loop. It creates a network effect. That means that there will be a handful of winning privacy chains. Everything else will be small relative to those winners. [music] >> Welcome to the A6Z Crypto podcast. I'm Robert Hackett and today we're joined by Ali Yaya, a general partner here at A16Z

Crypto who recently published a thesis about how he thinks privacy will become the most important moat in crypto. We talked to him about his idea, what's given him conviction in it, and its implications. We also talk about the variety of technologies that are available to us to achieve that privacy vision, and whether the outcome of having just a handful of privacy chains that win conflicts with the decentralization ethos of crypto. If you want to read Ali's post, you can check it out on a 16zcrypto.com.

It's part of our big ideas for 2026 package. And here is my conversation with Ali. Ali, you recently put out a thesis. You said that privacy will be the most important moat in crypto. That's a big claim. What gives you such confidence in that thesis? And why are you so sure of it? Now, >> this was inspired by just me spending some time thinking about how block space is becoming functionally the same everywhere. There's this overabundance of uh sort of quote unquote high performance blockchains uh that are coming to market or that maybe in some cases have existed for a long time now.

And also with the fact that we now have very easy bridging solutions that block space that is essentially functionally equivalent is also now accessible from everywhere. So there's this commoditizing force. So I was trying to think what will be the the actual sources of defensibility and value capture and crypto longer term. And it occurred to me that privacy is a unique feature both because it is the one feature that no existing blockchain or very few existing blockchains actually have and also it is a feature that creates a form of lock in that actually strengthens the network effects of a blockchain. And then the reason for that is that in a public blockchain, in the public setting, because of the fact that

it's trivial to move assets from one chain to another, you don't have strong network effects because users no longer care as much what blockchain they're on. And I do believe that longer term, people are not even going to know. They're just going to be operating and using applications that under the hood use whatever blockchain satisfies the needs of the application best. and the blockchains themselves will become fungeible and commoditized with respect to one another. That's the case in the public setting. In the private setting, once you have privacy, that is no longer true as much because moving secrets is much much harder than moving assets. The reason for this is that when you have private states on a blockchain, moving

it to another chain entails quite a bit of risk that some of the private state will be exposed. And the reason for that is that you are different privacy zones have different anonymity sets and moving between one zone and another entails transactions on the memp pool. I want to get into all these different aspects of what you're saying here especially the idea of lock in and this idea of the difficulty the friction involved when it comes to migrating secrets. But before we get into that, I want to interrogate this view that block space is becoming commoditized because I think that there are people working on different teams out there who are different blockchain tribes are working on different

projects. They have different end goals, different visions of what they're trying to build, different product roadmaps, different sort of trade-offs that they're making in their technical approaches. And so I want to interrogate this idea that it's all becoming commoditized because I think some of those teams might push back and say, "Hey, I we're building something that's unique and differentiated and it's, you know, fit for these purposes versus those purposes." Even if we just look at Salana and Ethereum, they've taken different approaches in what they're trying to optimize for. So what would you say to somebody who is, you know, saying, "No, my blockchain is is different." I would argue that at this point performance alone is no longer

enough. So in order to succeed as a general purpose blockchain, you need one of three things. You either need an already thriving ecosystem the way that maybe Solana and Ethereum have. You need some kind of unfair advantage on distribution the way that some of the fintech companies like Robin Hood and others may have. Coinbase with bass or potentially stripe with tempo or you need some kind of killer application that end users actually want to use. Just performance, just infrastructure that is vanilla block space is no longer enough to get people to want to use your chain and to get developers to want to

build on top of it. So that's the kind of concise statement. Privacy is a feature that would allow a chain to differentiate itself, but it also has this added benefit that it creates a strong network effect because now that users have joined a chain that has privacy because they care about their privacy and because moving to another chain entails some risk that their privacy may be lost, now they will be less willing and less likely to want to go elsewhere. >> There's an embedded assumption here that people care about privacy. people have been pretty content to use blockchains without having any of these sort of strong privacy guarantees, confidentiality of their data. What's changing? What is going to cause people

to actually begin to care about this stuff? People claim that users don't ultimately care about privacy. And I think that that may be true when it comes to some of the more incidental data about their lives and has been true when it comes to to their usage of things like Instagram and Facebook and other social media apps of the sort. But I don't believe it is true when it comes to finance. I think people very much care about the privacy of their financial activity. And in order for crypto to actually become mainstream, it's going to have to have privacy. And that's not only because users will care about it. People will care that their salary is not visible to everyone. They will care that the transactions they

make online are not visible to everyone. It will also be true because enterprises and companies that use crypto for their own financial reasons will absolutely require privacy for their operations. So is it the case then that what's different now is all of this new institutional adoption and interest is what's going to be driving this privacy trend because for them it's non-negotiable. That's part of it. I think the well the two things are privacy matters much much more when we're talking about financial use cases and also it is inconceivable that the enterprise use cases that we would like to see on chain can happen without

privacy precisely for that same reason and in the case of finance it is even more important for the enterprise for fintech and for financial institutions to have privacy than it is for consumers but even for consumers given given that it's financial financial data that we're talking about, privacy is essential. >> You mentioned one instance of something people might not want to reveal their salary information. But um I'd love to just make this very concrete for people when we talk about secrets and the kinds of secrets you might want to keep on chain. What are we talking about? What are some like examples of things that should be kept private?

>> What are the kinds of things that you buy on Amazon? Is that something that should be visible to everyone in the world? What kinds of websites do you have subscriptions to? >> What friends are you paying and how much are they pay are you paying them? How much money do you have? >> What is your salary? How much do you spend on rent? How much do you spend on all sorts of other services? What are your preferences? What are what are the kinds of things that you're into? All of these things are very legible from your financial activity. And therefore, I would assume most people would not want those things to be public. There are all these sort of debates raging about what crypto is for and whether it extends beyond the financial or whether it really is just a financial movement. A

lot of the things you cited are financially oriented. But is this is there also a bigger sort of world here of other data that could be encompassed that should be kept private? Yeah, it's becoming very clear that the initial use cases for crypto will will primarily be financial down the line because again our thesis has always been that blockchains are general purpose computers that have this special property which is that they can make commitments about their behavior long into the future. We will see blockchains be used for other things as well. There will be applications that are maybe not purely financial in nature that still leverage the fact that a blockchain allows you to write programs that are free of interference from anyone

including the people who originally wrote them and including the people who run the physical machines the computers that that execute those programs. I think one of the good examples that we haven't seen work yet but I still believe will will work eventually is a kind of social network that lives on chain. The social graph is on chain. the interactions between users can be onchain and it is not controlled by a centralized monolithic tech giant the way that today social networks are and where users they can control their own say recommendation algorithm that shows them a feed of the activity in the network and they can change it if they'd like and it can create some kind of competitive ecosystem of people who provide the various different components

that make this network a great experience for the user in much the same way that the internet works. The internet is not really controlled by a single entity. It's it's this ecosystem of participants that have come together to create the experience of the internet. In that world, privacy unlocks all sorts of applications that are simply not possible on top of a blockchain that's purely public. It could also be that in the gaming world, you don't want the full state of the game to be visible to all players at all times. Most interesting games have some aspect of the game state that isn't visible, that is private. There's some uncertainty in the game. If you and I

are playing, say, poker or if we're playing a board game that is interesting or a video game that is interesting, usually there will be some uncertainty as to what's going on in the game. That helps make the game fun. I feel like maybe we need a whole podcast to talk about applications in crypto that are not just financial. Yeah. because there's a lot of nuance and a lot of rationale for why we believe that those applications still make sense. But I think at least in the next several years, the most interesting applications will likely be in the more kind of the financial realm.

>> Okay. So in the short term, we're thinking mostly financial data and things of that nature. Let's talk about this idea you have about how secrets are hard to migrate and the moat that that creates then for private chains, privacy chains. What makes secrets hard to migrate? Is it a technical issue or is it something more social about it? Like what what is the cause of that friction? Yeah, it is it is a fundamental technical issue that stems from the fact that any privacy system has to operate within what's known as an anonymity set and that your privacy is guaranteed by virtue of the fact that your activity is sort of mixed with the activity of many

other users such that it is not possible to attribute any one action to you or to anyone else. The larger the anonymity set, the better, the harder it is to track someone down. >> Strengthen numbers. >> The smaller the anonymity set, the more risk there is that you might be tracked down. Whenever you have different anonymity zones, different chains are different zones in that they have their own anonymity set. You have fragmentation of the anonymity sets that you that you're inhabiting. And crossing from one anonymity set to another entails some risk because there's a lot of metadata that gets leaked in that transition including timing of transactions, the

correlation of transaction sizes on one anonymity set versus another one, network activity that happens at a far lower level, even at the sort of the TCP IP level. There are all these technical reasons for why there's just some risk when you move from one from one chain that's private to another chain that's private or when you move from a chain that's private to one that's public especially in that case that your privacy will be lost. I'm not saying that it's impossible. I think that we likely will spill systems that preserve users privacy even when they're moving between chains. But it is much harder and there's some unquantifiable risk that whenever you do that, whether because there's some very clever observer that can track all of the

metadata or because there's a bug in the way that you're actually making the transition, there's a risk that you lose your privacy. And so then as a user, you might think twice. You might want to pick the chain that has the most users, that has all of the functionality, and then just stay there because moving might mean that you you get exposed. And that's a strong network effect. It means that it's no longer as easy to move from one chain to another as it is in the public chain world where really it is just trivial with with bridging protocols and messaging protocols like layer zero. In the privacy world, that's no longer the case. There's a an additional hurdle that users will have to get comfortable with in order to want to move. And so then that creates a self-reinforcing feedback loop. It

creates a network effect that that means that we'll likely have more of a winner take all or winner take most type of dynamic. there will be a handful of winning privacy chains. And then if we if we extend that to the fact that we believe that most interesting applications in crypto will require privacy, that might suggest that there will be a handful of chains that have privacy as a feature that own all of crypto and most most everything else will be small relative to to those winners. >> What you're describing is basically there's these these two forces. There's a push and a pull. The pull is for a privacy chain. The fact that you have more users on the network provides better privacy and security to the other

participants. So that is like a fundamental part of the network effect there. You want to gravitate toward the biggest kahuna out there. Then on alternatively or on the other side of the equation, you've got this push where you don't want to leave that network because every time you might leave that pool of users, you are risking unshielding or or revealing some aspect of your metadata that could out you. And so you have this attractive force and also this deterrent for wanting to leave a given network. And that's that seems to be the basis of this network effect that you're describing.

>> Yeah, that's right. It kind of reminds me a little bit of, you know, if you just look at the financial sector and as we've seen the biggest banks getting bigger over time and community and smaller banks, regional banks struggling, you know, because people want to put their money somewhere that they have confidence will be safe and stable and that the government will not let fail. And so this sort of naturally causes the bigger banks to get bigger. >> And so it's interesting, yeah, to see that play out. Yeah.

>> In traditional finance, but also in this new area. It's a different kind of network effect. But >> yeah, and I would say that in this case, it's even stronger because in the traditional banking system, even if you bank at a small regional bank, you can still send a wire transfer to anyone anywhere, regardless of who they are and who they bank with. In this case, if you're on a privacy chain, again, sending money out of the anonymity zone into another zone risks that transaction getting exposed. So, even though it's possible, you will probably still be able to send money to everyone to anyone anywhere. It's a little bit harder than to just interact with people who are already in your privacy zone, which would be inherently safer. And so when deciding which chain to join, you would

probably want to join the one that will enable you to interact with the most users. And this is the classical network effect. If you're joining a social network, you probably want to join the one that has the most users because there's more value in being able to interact with more people, >> more people to talk to. And once that flywheel gets started, it is this self self-reinforcing feedback loop because the value of that network scales quadratically with the number of users because now you have like this pair-wise number of possible interactions that can happen in that network and the more users you have the more pair wise connections are possible. >> You've used this phrase privacy zone.

What do you mean by zone? By zone, I just mean a single unified system that has one anonymity set. And usually a say a chain that has privacy as a feature will have a single anonymity set. And the bigger the number of users that join that chain is, the bigger that an anonymity sets gets and the stronger the privacy guarantees become. >> So you're basically saying like these are just it's just a generic word to describe different areas where you get different systems, but if you move between them, maybe you could You're moving between anonymity sets. Yes.

>> Okay. >> Yeah. Exactly. >> Could bridges get good enough where moving between these different privacy chains becomes totally secure and you don't have to worry that you're leaking any metadata at all because the thesis seems to rest on the assumption that these movements are going to cause you to expose yourself. What could be done to prevent that? >> Yes. Yeah. I think in the long term you will have bridging solutions that can bridge different privacy zones fairly securely. Uh but there is this fundamental problem that will continue to exist which is that privacy depends on the size of the anonymity set that you that you inhabit and whenever you cross the boundaries between different

anonymity sets there's always there's a greater risk you may be exposed in so far as the anonymity set of a new blockchain that you're moving to is still small. There's still risk there. uh in so far as the number of transactions that you make across anonymity sets is large, maybe there's also some risk there that even if the bridging solution is perfect and manages to kind of re-encrypt your state from one chain to another, there will still be some information that is leaked every time you move between chains just by virtue of the fact that you're joining a different anonymity set than the one you were in before. And that's a fundamental problem that no amount of technology can solve. And I think that the risk of

being exposed is unquantifiable. And so then as a user, you probably will think twice about moving or sending money between zones more than you would if if you lived in in the fully public domain. I think longer term this may not matter as much. If you do have again, if you have a few chains that win that have massive anonymity sets, then moving between them will never present any risk. But then that is the concluding point of this whole thesis that you end up with a small number of winners all of which have privacy. But it's not the case that you have say hundreds or thousands of chains that have the same feature. So then the way that you get that guarantee is is by only moving to chains that are massive. So this idea

that you could have these winner take all or winner takemost dynamics in crypto that might be unsettling to people especially when they hear the word lock in that might trip some people up. Crypto has historically been all about having an open-source sort of ethos. People are very interested in composability interoperability permissionless exit, these sorts of things. How do you reconcile this apparent trend toward consolidation with those core fundamental values of crypto? The key thing about this is that every privacy chain that's built in the right way will still be decentralized. It will still be permissionless in that anyone

can participate. Anyone can either use the chain, build on the chain, or become a validator for the chain. and open source. So the code itself is also visible to everyone even though the state of the chain at any given time is encrypted. And so if those things are true then the ethos of the space continues to be preserved in that you have a chain that has certain rules that are enshrined as part of the chain and changing those rules can only happen through some governance process that involves all of the participants in the network. And that may depend between chains, but a truly decentralized network would still be representative of

the entire community that uses that network and not under the control of a single entity like a Google or a Facebook or any other kind of centralized actor that could change the the rules of the game on a whim. So you get the best of both worlds and that you get privacy of all of your personal information and your transactions and your activity in the network, but you still get the guarantee that the underlying network is credibly neutral and won't all of the sudden change its rules in a way that might screw you. And the way that that's verified is by virtue of the fact that the the code itself is open source and control of the network is distributed across a large set of independent actors that all come

together to make the network do its thing. >> So what you're saying is that this value of decentralization and this possible trend toward consolidation are maybe not even in that much tension. Maybe you actually get both the best of both worlds as you said where you can have this consolidation but things are still decentralized. >> People think decentralized they think sort of you know many >> many nodes on a network. >> Yeah. Well decentralization refers to control. Yeah. So when we say that a system or a network is decentralized, we mean that it's not controlled by a single monolithic entity, whether it be an individual or a company that could make unilateral decisions about the way

the network functions. And so the internet is decentralized. Solana, Ethereum, networks like that are decentralized. And we have that the famous line, the line that we like to use, which is that can't be evil is better than don't be evil. Google had that famous cultural value which was don't be evil. But it is much better to be able to enshrine that into the network in such a way that the network cannot be evil even if the people involved wanted it to be evil. So those principles, those values continue to be upheld even in a world that has privacy because the chain itself is decentralized, incredibly neutral, and still makes a long-standing commitment to the rules of how the network

operates. Just to tie it all together for people who might be concerned about lock in, they are familiar with the lockin that happened in web two with big tech, why would lock in or the sort of lock in that we would see among privacy chains. Why would that be superior or preferable to what we've seen in past eras? >> Well, okay, there are two kinds of people who may be thinking about this. One of them is the end user who they may they may be thinking well if I join this network I want to be able to later change my mind and join a different network and there's also developers who may be deciding if I build on this network maybe later I want to change my mind and move over move my application

to a different network and for each of those users there may be different answers for the end user even though uh there's some friction in moving between chains that support privacy. It is not infinite friction. If the chain does ultimately succumb to some kind of governance attack or all of a sudden isn't as functional or as credibly neutral as you wanted it to be or isn't doing it for you for whatever reason, it is still possible to move and there may be some risk that you lose some privacy in the process, but it is still an open and permissionless system. And so there will be bridging technology that will allow you to migrate elsewhere. So you're not locked in permanently in the

way that maybe you would be in the way that you you kind of are in existing social networks today like like if you're a user of X and you have a large following, it's pretty hard for you to leave because you can't take that following with you. You'd have to start from scratch if you go elsewhere to some other network. For developers, it's a bit of a different story because the thing that they care about the most is platform risk. Can the underlying platform screw them over? Can they all of the sudden change the rules and decide that maybe they want to charge more or extract more value or censor some of their transactions or some of their activity? That's the thing that as a developer you care about the most. And that's guaranteed by the fact that again the rules of the game are enshrined and

encoded in the network and it's a decentralized network. And the only way those rules can change is through whatever governance process that network follows to get those rules to change. And if the network's properly decentralized, that should give you some trust and some some ability to believe that you're at least somewhat safe from platform risk. Much more so than when building on top of a fully centralized API like the web 2 APIs, which were the cause of of the death of so many startups, especially in the late 2000s and early 2010s. Mhm.

>> In the case of crypto, you're much more protected against that risk because of the fact that the network itself guarantees that the rules will remain incredibly neutral. So, if you're right about all this and if privacy is this moat that will lead to just a few big winners, why does it seem like not much attention has been paid to privacy, at least compared to performance? Everybody has been obsessed with things like higher throughput, lowering fees.

Privacy has historically sort of been lower on the agenda than those things. There's always been a loud and vocal minority that has been small throughout the ages. We've actually invested in a lot of them. But I think that the reason that it hasn't been the top priority from the start is that it is fairly difficult from a technical standpoint. And there were so many other technical problems that needed to be solved just to get started, just to build a blockchain that can scale, just to get to the performance level that we now have such that we can say that it's now table stakes that just cause people to focus on those problems first because that is it is table stakes. If you don't have a blockchain that functions and it

and actually scales and can support meaningful activity, then privacy isn't even a consideration. But now I think that now that the infrastructure has clicked into place to the point at which we can have say the arbitrary movement of any amount of money from anywhere in the world to anywhere else in under a second for under a penny as you can on say like Solana. Now that that's the case then now the the bigger lift will be to do that but also do it while preserving users privacy. >> What technologies are available to us to achieve the kind of privacy that you're talking about? So [snorts] there's four major technologies. One of them is zero knowledge proofs which is a

cryptographic approach that many many projects are using which allows you to prove that a transaction happened and that it was valid without anyone being able to see exactly what that transaction was. There's also FH which stands for fully homorphic encryption which is a more powerful kind of cryptography and has additional capabilities but that is much farther out and today I think most FHE solutions don't have the level of performance that you would need to be able to practically deploy them into today's blockchains. It's extremely compute intensive.

>> Extremely compute intensive. And so maybe down the line, but I think for now it's more of a theoretical technology. Then there's MPC which stands for multi-party computation where multiple people can collaborate to perform a computation without any one of them knowing what that computation is. And often times MPC is used in combination with some of the other technologies. And then the final one which is >> people are probably most familiar with that when it comes to like splitting up keys, key management. >> Yeah. Exactly. And yeah, multiple signatures. >> A multi-IG is a very basic example of a multi-party computation, but you can do many more sophisticated things using MPC that are more complex than just signing a transaction. And then the final one is

uh which is which may be the most pragmatic one is TEES which stands for trusted execution environment which usually relies on trusted hardware that's manufactured by a hardware vendor like Intel or Nvidia or any hardware manufacturer that you can use to perform a computation inside of a special region in the in the processor that is not accessible by any other part of the chip or the software and that is fairly hard to access even if you have physical control of the machine.

>> Of the menu of options that are available, what do you see as the most credible path forward both in the short term and the long term to achieving a privacy chain? >> Well, the two most promising ones would be trusted and execution environments, TEES, and we have a number of portfolio companies exploring that direction simply because they are the most pragmatic. like the performance level of using a TE is fairly high and comparable to what it is even if you had zero privacy on a public blockchain. And then the other one is zero knowledge proofs which have improved dramatically and our own team have made great strides in making zero knowledge proofs more efficient with projects like Jolt. So those two are by far the most promising.

oftentimes you might use NPC for some aspect like you might combine TEES with NPC for example to give yet another level of security with sort of defense and depth even if the TE itself the hardware were to be broken if say the person who holds the physical machine is able to break into the enclave in which the trusted transactions are you could still have a layer of multi-party computation that gives you additional protection and so even if that were to happen privacy isn't lost So you can stack or layer these technologies to get additional asurances and levels of security.

>> Of the projects that are working on this right now, are you expecting the handful of winners to be from the existing set of chains that are out there or is this going to require completely new formulations of technology? That's an open question, but I I am very excited by a number of the existing projects. >> How does this thesis that you have affect how you're going to be making investment decisions in the coming year? Like given the size of the opportunity, is it the thing that you are most concentrated on now or is it one of many other things that you're looking at?

Whilst we certainly have invested and maybe will continue to invest into privacy solutions, that is certainly not the whole universe of possibilities in the space and and of interesting applications that will be built. Many interesting opportunities might actually just be building on existing solutions or building on top of the privacy networks that get built and those opportunities will themselves be interesting venture bets. I think the main thing will be an awareness just like our own awareness, the privacy will be an important piece of the puzzle and wanting to see that same level of awareness in the founders that we invest in just having making sure that the the founders that they have some kind of

story for how in the very long term privacy makes its way into whatever they're bullying. So as we look further out, how concerned should people be about the risk of a, you know, powerful quantum computer decryting things or making networks insecure? What level of concern should we have about this? And is enough being done to prevent or mitigate those threats? >> Yeah, this is a very important question, very topical these days. Everybody seems to be seems to have a strong opinion. I I always turn to our research team who I think is the closest. Both Justin Taylor and Dan Bonet are probably the best

people. And our our latest sense is that realistically we will not have a quantum attacker that is capable of breaking modern cryptography for another 15 years. That being said, it is important that we start preparing as soon as we can because the transition to a quantum secure alternative to our current cryptography will take a very long time. So I think it is absolutely correct that people have this topic as top of mind and are working towards better approaches. But it's alarmist to say that anytime in the near future our modern blockchains uh cryptographic

systems are going to be broken by some kind of quantum attacker. There's always some uncertainty as to what the timelines will be but uh even with very big error bars is not going to be for a while. Okay. So one last question for you. People are expecting the internet to become filled with sort of agentic AI that's going to be transacting on people's behalf. One thing that agents need is data. They need context and they need lots of it. If we start seeing a world where a lot of data that is important to us, you know, me personally as a consumer, if that is owned by me and concealed in a way in some sort of privacy chain, very broadly, where do you see the collision between your

thesis of the rise of privacy chains and the AI powered internet that we are quickly approaching? I'm sure there will be many others, but one obvious one is that in a world of AI, of prevalent AI, we will all be living in a kind of ponopticon where every single thing we do is training data for the next generation of AI model. And so then the desire for privacy may become that much more salient and important for people who are using the internet. >> I think I'd have to agree with you.

>> Cool. Ali, thanks so much for your time. >> Thank you. It's a pleasure. Heat. Heat.