Uma Roy - Truth in a Post Truth World - TOKEN2049 Singapore 2025

Hello, good morning. I'm Uma, co-founder and CEO of Succinct. And today I'm going to be talking about truth in a post-truth world. So, first let's start with the problem statement of this entire talk. What is a post-truth world?

As I'm sure all of you guys have seen recently, the rise of AI models like ChachiPT, but even image models like Midjourney and Sora 2 that came out recently have resulted in a flood of content that makes it impossible to distinguish what is real and what is fake. And the cost of this content is trending down exponentially. And we're already in a world where generating a real image or a real video or something that looks indistinguishable from a real image or a real video is very very cheap, fast and left only to your imagination. And this is a huge problem because when you're browsing social media like Twitter or Instagram or even entries in traditional media like journalism, it's impossible to know what's actually happening uh versus what someone might have just made up. And this is what we call kind of a crisis of trust.

So in this world where it's impossible to know what's real versus fake, there's kind of two ways out. One is the traditional path of centralization where we trust governments and big tech companies to tell us this image is real. This fact is real. This video was taken at this location at this time. But the other one and the reason why we're here is a more decentralized way where we use distributed systems, blockchains and the power of cryptography to kind of prove to us what is truth versus not.

So it's succinct. We're working on zero knowledge proof technology and zero knowledge proofs have are something that have been around for a very long time actually starting in the 1980s by two MIT professors that ended up winning the Turing Award uh one of the highest awards in all of computer science for this work. It's a very complicated mathematical subject but at its core it's very simple. The key idea is that you can prove something happened without revealing all the details. And while zero knowledge has been in the lab for a very long time, uh, for the past 30 years, recently it's kind of broken out of the lab.

And with engineering advances and research advances, it's been able to be something that can go mainstream and help us solve a lot of these problems and prove that things are real, that transactions actually happened, that images actually real and come from a real camera source. And that's what I'm going to be talking about next. So just to give an intuitive concept of ZK that maybe isn't so stuck in the math and details but gives you an idea of how something like that might be possible. I wanted to walk you through this talk a thought experiment. So let's pretend we're all playing a game of where is Waldo.

This is a kind of a classical game where you have someone uh you want to identify Waldo amongst a sea of people. So, if you have some image, uh, in this image, it's probably pretty hard to find Waldo, where Waldo is right. I'll give everyone a moment to try to find him now. First, in without ZK or, you know, any sort of technology, the naive way is that you scan through everyone in the painting and then you have to find Waldo and it reveals his precise position. And then if I do it once, uh, you also have to do it and find Waldo yourself.

So it leads to a lot of repetitive work. So what if there was another way in ZK? What happens is you have this, you can imagine you have this big sheet of paper and you have a small cutout and then what you do is you guide the small cutout over your painting and then you can center it on Waldo. And what's really nice about this approach is that once I found Waldo, I can basically prove to anyone else that there indeed is a Waldo in this painting, but I don't have to reveal his precise location because it's under this piece of paper that covers the entire map. And you don't have to redo all the work because I've given you proof that he is there.

So obviously this is kind of a toy example but it kind of shows this powerful concept that I can do this all this work scan the painting and find Waldo and then I can prove to you and you don't have to redo the work but you also don't have to trust me you know that Waldo is there. So this is a extremely simplified concept but hopefully gives you kind of this illustrative example of how ZK works. So in general with ZK we can not only prove statements about Waldo and paintings but we actually have this new paradigm where you don't have to trust but you can verify that stuff is happening. So now I want to talk about how ZK will actually be useful. So I'm sure as a lot of you have heard in blockchains ZK has already been explored as a solution to one of the biggest problems which is scaling.

So today if you look at Bitcoin or Ethereum mainet transactions are very very expensive. Uh you can't have too many of them and this is because there's a bunch of nodes all across the world that have to redo the work and re-execute all the transactions to come to consensus on the chain. Now in Ethereum world there recently has been a huge push to scale the L1 itself and get to 10KTPS on Ethereum L1. However, Ethereum wants to keep its core chain verifiable, which means they want this property that anyone in the world who's not running some big box in a data center can still keep up and verify the truth of what's happening with those transactions. And ZK is critical to this.

So the concept of what the EF, the Ethereum Foundation is working on is that they want to generate ZK proofs for every single block and then everyone in the world even on a mobile phone can download these proofs and instead of re-executing transactions, they can just verify these proofs and make sure that Ethereum the chain is actually progressing as kind of promised uh while staying decentralized. So Ethereum mainet recently has gone allin ZK to scale the L1. I'm sure you guys have also heard of the concept of ZK roll-ups or L2s. Uh L2s are kind of these chains that settle on Ethereum and leverage Ethereum security. And the way they do that is by using ZK tech to accelerate their finality.

You put money on Ethereum into a rollup. You can do a bunch of computations on the rollup and then you verify ZK proofs to bridge money in or out. And then finally, there's been a very very exciting category recently of perexes. So you have these verifiable per dexes that are decentralized exchanges and many of them are starting to prove that the trades settle in the correct way using ZK technology and that's a third category in blockchains that our team is very excited about. However, I think there's also a lot of stuff in the real world that ZK can help with that I'm also excited to touch on as well.

So one very real example it uh similar to the AI example or problem I was highlighting earlier is that today it's hard to distinguish real images from fake images. However, imagine a world where every single camera ends up signing on the image that's taken at the point of capture. So you'll have a signature from a real physical device that an image was taken at a certain location at a certain time with certain raw pixels. Then you generate a ZK proof on top of that signature uh to prove that an image was compressed or edited or cropped in a certain way and that when you verify that proof, you know that that image was derived from a real camera. And you can imagine a future in which all social media timelines have a nice verifiable check mark that links to a proof uh to know that those images are actually real.

I'm also excited about role of ZK in things like KYC and compliance where today there's a very very uh ownorous KYC regime where you have to submit all your information whether it be your passport or driver's license to exchanges and often these things get hacked uh and reveal very sensitive personal information. Imagine instead that you submitted a ZK proof of your identity where you can prove that you're over 18, that you live in a certain country and then you submit only the proof to the exchange and they don't have any of your other personally identifying information. Recently, the Coinbase chief legal officer has been writing about how ZK can help improve these concepts of the Bank Secrecy Act and compliance. And that's another really big use case of ZK in the real world. And finally, today the internet broadly is gated by gated APIs that only certain websites let you access and it's up to the website owner whether or not you can access and take your information from one account to another account.

ZK can also help take that information or any from any API or any website and then you can prove that there's certain aspects of your account whether it be you have a certain balance on an exchange whether it be your trading history on other on another platform or even other stuff like you spent this much money on Amazon or you know you're loyal to this brand and you can make data much more portable using this technology. So the main point of this slide that I want to highlight is that not only is ZK useful for blockchains, which I think is a theme that's been explored, you know, very widely over the past five years, but that ZK in the real world is another really exciting category that's starting to emerge and especially relevant in this age of AI. So it used to be the case that you know building with ZK is too hard and building with ZK is so expensive but actually that's the problem that we're solving at Succinct. So ZK as I mentioned earlier in this talk has been around for around 30 years but it was always you know a research concept in the lab with what we built at Succinct a general purpose zero knowledge virtual machine or ZKVM for short. If you're a developer, you no longer need a PhD in cryptography or need to understand some specialized niche languages.

You can just write normal code and normal software and then prove it in ZK. And actually SP1 or ZKVM is one of the fastest growing ZK libraries on GitHub, which is really exciting. And we have people proving all sorts of things in ZK using SP1. Now using ZK is still a bit expensive and challenging even with SP1. Even though it makes the developer experience easy to generate proofs really fast and cheap, you still need specialized hardware GPUs.

And so our protocol, the succinct prover network is our protocol where we have GPUs all across the world, similar to data centers who have been mining Ethereum or Bitcoin and they are able to connect over our protocol to people who want proofs and provers who have that sophisticated infrastructure. And it makes using ZK as easy as requesting proofs on our network. And I think that ZKVMs are going to be the endgame of ZK. And it kind of follows a lot of trends in other complicated technical fields where you have this initial layer that's you know very application specific and hard to use. And then you have something like ZKVMs that make it very easy and accessible to all participants uh through normal software and normal code.

Similar to how LLM's made AI accessible to normal people by making it so that you could just type English and get AI, our ZKVM makes ZK accessible to any piece of software and any developer who knows how to code by writing normal RS code and getting ZK. And then we all connect this together by basically having our network and our token. And our token is kind of this utility token of the network where basically provers get paid in it. Users request proofs in proof. And then as this network spins, you have more and more provers coming online which drives down the costs which makes more use cases accessible.

And then you have the use cases getting generating more and more demand which brings more and more provers to the network. And I like to kind of call this broad concept the proving economy. Now within crypto we already work with a lot of the best brands that you may know. So for example mantle which is a rollup uh associated with bit they're using SP1 in production to secure their layer 2 chain. We also work with polygon katana and the ag layer which is another rollup with around half a billion in TVL all secured by succinct and because zk is such a broad technology it's not just rollups that we work with we also work with bridges other ecosystems like celestia and we even connect ecosystems like cosmos to ethereum mainet and the numbers kind of speak for themselves today we're securing 4 billion in TVL we have millions of dollars in fees and this is just with our network launching a month and a half ago and we have many many customers for all of these rollups.

ZK also plays a very critical part in their product and roadmap which is very awesome and means we provide a lot of value. So Mantle uses ZK to make fast finality and make their product very appealing for their end users. Instead of depositing money in and only getting it seven days later, you can deposit your money in and with ZK get it out in less than 30 minutes. Similarly for Katana, we're kind of providing a similar service and a bridge called Across which connects a lot of the layer 2 is also using us to make their finality for their service very fast. And I think overall over the next 12 months, we're going to see incredible tailwinds for for ZK.

Whether it's the Ethereum Foundation going allin ZK for scaling DL1, whether it's teams like Offchain Labs, the people behind arbitrum who we recently partnered with to bring ZK to all arbitrum chains. I think it's a very exciting time for ZK in general and succinct. And I want to end the talk by really emphasizing that ZK is not just a technology only for blockchains, but something that actually can go truly mainstream. And we can see this in the form of, you know, institutions like the White House writing about ZK being important for the future of KYC and compliance. institutions like Google actually using ZK and embedding it in their identity wallet uh and using it to prove facts about people's driver's license and identity in a privacy preserving but compliant way and even institutions like JP Morgan who are also excited about ZK.

So overall I think this is a really exciting time for ZK. I think finally all the ingredients are there. It's easy, it's fast and cheap and I think it's time for ZK to eat the world. Thank you.