Business Privacy, Harder & More Valuable Than You Think | Paul Brody at Pragma Prague 2025

[Music] I want to bring on Paul who is from EY to tell you about why we should care about privacy if you want businesses to adopt more crypto. So, please give Paul a massive round of applause. Thank you. Listen, I I know you're all here for me, but just as a courtesy, the next speaker deserves your attention, too. Like, when I'm done, don't just get up and leave.

This is a little self-deprecating humor. So, business privacy. So, I want to talk about privacy. Privacy is hot. It's on the road map.

It's something I care about a lot. But in particular, if you think if you spend time thinking about business privacy, it turns out to be way more difficult than I would call like traditional consumer privacy that we spend our time thinking about. So I want to talk about like why it's valuable, why it's important, why why we should pay attention to it and and what we can learn and is difficult about it. So um coordination I'm going to spend just a couple minutes talking about and I'm going to go kind of fast because I have a lot of slides and a little amount of time. I want to talk about briefly like why why should we we care about privacy?

What is the value proposition if you are selling into the enterprise and the answer is kind of the same as what we have in the rest of the ecosystem which is coordination is the problem but the nature of the problem has changed enormously. So if you were to go back to the 1950s and fortunately we don't have to although some people are trying but if you were to go back to the 1950s manufacturing was vertically integrated like the the in in 1950 you ask somebody well like what is the greatest factory on earth they would have said oh Ford's factory at River Rouge like basically highly vertically integrated dirt went in on one side and cars came out on the other there was a steel mill there was like a tire factory all these things in there and the job inside of that organization was to coordinate internally. Coordination internally turns out to be a relatively easy process. If you're the boss, you say, "Listen, everybody, you're going to all do exactly it this way. You're all going to use uh this software."

First we had MRP and then we had ERP, right? And and so those things internally get done relatively well today. But if you ask somebody today, what is the preeminent factory in the world? I would argue that like the most typical factory of the 21st century is the Shenzen Foxcon assembly facility where they really do nothing except for screw together iPhones. And all the stuff that is really really valuable about the iPhone is actually done in other locations and thousands of companies in like 30 different countries work together.

And so coordination is the problem because uh when you cross enterprise boundaries, all the really cool things that we have spent the last 30 years building inside the enterprise stop working, right? You spend millions of dollars building sophisticated ERP systems with analytics and your suppliers would like you to fax them their orders or better yet send them an Excel spreadsheet or let's have a Google spreadsheet. Whatever it is, it doesn't work. The whole thing falls apart. And the value proposition of being able to have a shared space with tokenization and smart contracts is unbelievably compelling.

Even though Patricio just yawned, um, I'm going to choose to believe that that is from jet lag and not because this topic is so exciting. Yeah, I know. I know. I'm like a broken record. He's heard me before.

Um, I want you to trust me when I tell you that this is worth a crazy amount of money. And I'm not going to spend a lot of time on it today, but basically if you can do this this picture of tokens moving from left to right and money moving from right to left and all of it being governed by smart contracts, you can run business processes in 99% less cycle time with 20 to 30% less inventory and about half the administrative cost. Trust me, and this is crypto, so it's a very trustworthy industry um when I say that this is true. So if it's true, how do we do this? And the answer is that none of it can be done without an awful lot of privacy.

And specifically, there's two kinds of privacy. Because every business agreement on earth boils down to something very simple. I have money, you have stuff, real, virtual, imagine, whatever it is, and we are going to exchange my money for your stuff under the terms of some kind of an agreement. And we need privacy about the money and the stuff because no company wants to tell their competition what they're buying and where it's going. And then secondly, we need privacy around the terms of the agreement because if you can read the agreement that says I agree to buy a thousand widgets from you at this price, you've kind of like you've spoiled the whole secret already anyway, right?

So you need both kinds of privacy and both of them are really really difficult. Now there are uh a lot of companies in this space now that are doing really cool stuff. Uh I will explain to you how we are approaching it at EY, not because we're the only player in here, but because I at least can tell you with confidence what I think we're doing now. I do not have a PhD in math. So, the truth is I don't really understand what we're doing, but I'm going to pretend that I do.

Basically, we do two things. We have a system called Nightfall, which is a ZKZK rollup for payments and transfers under privacy. And then we have a system called Starite, which allows you to write Solidity smart contracts, mark them up with the parts that you want made private, and it turns them into zero knowledge circuits that can run on the Ethereum mainet. So, money for stuff and the agreement working together. Uh, and the QR code here is a link to our GitHub.

Both of these are public domain open-source applications. We have spent, I think, $30 million over the last eight years building them, and they are a full donation into the Ethereum ecosystem. There is no secret, special terms and conditions, anything like that. If you want to use it, you can. You can copy it, you can modify it, you can fork it, you can contribute to it.

So for enterprises though, privacy is hard and there's five key things that you should think about. Number one, regulatory compliance. I'm going to talk about that a bit more. Number two, complex business logic, right? It's not most personal payments are just like I have money.

I'm giving you some money. It's a remittance. It's it's a payment, whatever it is. Um, companies have logic, a lot of logic embedded like if I buy so much, then you're going to give me a discount or a rebate. Thirdly, almost everything that companies make is basically nonfgeible, right?

Apple makes a couple hundred million iPhones a year, they're all nonfgeible. The minute you slap a serial number on that thing, it's a nonfgeable token, right? And by the way, non-f fungeible tokens very easy to keep track of without privacy. Enterprises do lots and lots of transactions with each other. Samsung and Apple do business probably to the tune of 40 or 50 billion dollars a year.

That's a lot of payments. If I'm just paying you once, that's easy to disappear into like the flood of stuff, right? But if I'm paying you 40 or 50 billion dollars a year for hundreds of millions of DRAMs and and and uh displays, if you have good analytics, there's no obfiscation system other than proper zero knowledge privacy that's really going to hide that. And then lastly, people have multi-step business networks, right? It's not just me and you.

It's actually me and you and like 500 other companies that are in this ecosystem together. So, uh, getting that all done is quite a bit different than a lot of the traditional things like privacy pools or mixers that work because you put money in and then you take it out and it's a one-time payment or a person-toperson payment, right? Um, let me take a moment in particular to talk about regulatory compliance. This is a big deal and it is going to become a bigger and bigger deal because regulators care a lot about where money goes and at least at EY we have decided my personal objective I'm a big believer in privacy. I'm a big believer in decentralization.

My goal is the maximum level of privacy I think we can legally get away with. To do that we are really focused on what something I would call confidentiality. So, Nightfall, which is our zero knowledge payment layer, you can use you cannot use anonymously. You must have a user identification. In our case, we chose something called an X.

509 certificate. And we picked that specifically because those are issued by the internet certificate authorities. The process for issuing them is an open standard and the tools themselves are all open standard which means that there can be no gatekeeper that stands in your way of saying hey I am the sole issuer of this type of credential. You can get them from lots of different parties. As long as you have one of these identification certificates you can use it.

Now that compares to like private and anonymous systems. These are going to get a lot of regulator attention. And I would encourage anybody who is building for privacy to think really carefully about regulatory compliance, you don't want to be in the crosshairs of regulators. Um, as I like to say, you know, to our clients, part of the value proposition of working with EY is that you are less likely to go to jail. It's a joke.

At this time, none of my staff have gone to jail and nobody's under indictment. So, you know, so far so good. Um, beyond regulatory compliance, what are the things that we want out of privacy for enterprises? Number one, scalability, right? I've got clients who are literally want to move millions of NFTts per day, right?

Gas costs need to be really low for that, right? Secondly, programmability, right? I need to be able to put in embedded business logic. Thirdly, composability. I need to be able to connect all these pieces together.

And then lastly, permanence. This is something that's especially different from traditional privacy systems, right? Companies want to run their entire uh privacy infrastructure supply chain permanently under privacy. It's a big dark pool, not something that we go into and then out of, right? And that that's particularly important, right?

This is a permanent layer that the entire process and ecosystem is going to run inside of. Now, everything by the way that you want to do without privacy is much harder with privacy. Supply chain traceability, where did that token come from? Who had it before? Where has it been?

All of that stuff is much harder. It's solvable, but these challenges are quite a bit more complex when you try to layer in privacy. When it comes to privacy technology, there are, we think, four major technology choices. The first are zero knowledge proofs, snarks and starks. Secondly, multi-party computation.

Thirdly, homamorphic encryption. And then fourthly, trusted execution environments. And one of the things that's really cool right now is we are living in I want we're getting, I think, quite close to what I would call the golden age of privacy technology experimentation. Right? I I could not have made a slide like this even three years ago with this number of different companies in this race.

Right? We've been at EY, we've been working on this for I want to say uh eight years now, right? So, my joke at at EY meetings, which none of my managers find funny, is that we are currently eight years into what I thought was going to be a two-year project. Um, but uh we have chosen zero knowledge proofs as our approach. Um, I think the other ones are pretty interesting and cool.

The only one that I think I would that I am specifically passionately against are trusted execution environments because I do not trust them and anybody who says you should has some hardware to sell you, right? Hardware that probably has some significant security flaws in it or possibly back doors. So, uh we have also looked at what could we do to combine these? Could we make a better solution? All of these have pluses and minuses and I'm happy to share them with you.

um you get a lot of silly silly cat pictures. We had fun with them. Our conclusion at the end of looking at all of these different sort of reasonable possible combinations is that none of them are particularly attractive. They solve some problems, but they create way more problems. And frankly, I I don't think I could hire enough PhD cryptographers to tackle all of them.

So, in the end, we stuck to the basics, which is uh zero knowledge proofs. And we also look at things like uh snarks and snarks and which ones are um for example quantum resistant. The truth is that snarks are more or fully quantum resistant whereas snarks are not. Um and you have to one of the key decisions we'll have to make at some point is can we adjust or adapt if we think that sort of quantum uh supremacy is really here. So far I feel like it's it's a lot of demos and not a lot of reality but um it depends on honestly on on who you talk to.

So uh where are we in this kind of overall roadmap? So despite the challenges that we have had right I think we are moving past the early stages of this marketplace and we are moving into a time when we can do real applications. Now, we showed I mean, it's funny actually. It's it's a nice come coming home. In 2018 at DevCon, we debuted the alpha, the prototype version of Nightfall.

And uh in that version, it cost us a $100 to do a single transaction. And this was, by the way, when ETH was $265, right? that is not in any way, shape or form scalable, right? We could prove a proof of concept. That's it.

Where we are now after 8 years uh and with both Nightfall Starite and things like the Aztec test net is we are in what I would call early high-V value use cases. That means things like financial services. So the price of doing a privacy enabled transaction is not low enough for you to process a million NFTts a day and build a whole supply chain. But it is low enough for you to do things like uh build a DEX, right? It's low enough for you to buy securities.

It's low enough for you to make payments um and that kind of thing, right? So, we are in the early stages. One of the cool things that we have built is a customized smart contract tool that lets you implement business rules and guarantee to your customer that if they negotiate at a certain price deal, it's always implemented. and that any efforts by a supplier to submit to you an invoice that doesn't match the terms of your agreement get rejected. Right?

So, you can you can sign an agreement and you can be confident it'll be implemented. Now, some of you are thinking that's that's pretty basic, bro. Um, and the answer it is, but the truth about large companies is that they are great at negotiating deals and they are terrible at implementing them. Five minutes after the deal has been signed, they're like, "Wait, what were the terms? When do we get our discount?"

Right? Somebody is laughing. They know this is true. Right? where we are headed in the longer term is what I would call generalpurpose kind of deployment right so it's easily programmable there are mature tool sets around it and you can just write an application and deploy it and and customize it it will run at scale and perform very very well right we are not there yet but I think we are we are going to get there in a relatively reasonable amount of time right uh what makes me think that so what makes me think that is the history and our own direct experience building Nightfall.

As I said, back in 2017, we got started on this. In 2018, we showed the prototype, right? Fast forward to 2024, we had our fourth version. And our own estimate is that computationally performance has increased by a factor of roughly 10,000 during that time period. So if you think about you know today's version of Nightfall or Aztec's uh uh Aztec network right we're doing we're both able to do about one transaction per second right now imagine in five years time that we can do 10,000 times better performance.

So even though ZKP got frankly speaking a bit of a slow start and it was very complex the performance improvement over the last few years has been absolutely phenomenal like you really could not ask for kind of a better matur maturation of the technology. So we are going to get to a point I think in a very reasonable period of time where the outcome is uh very scalable, very deployable. We're going to have prices and operational costs that are are substantially lower than even typical mainet transactions today. Of course by then mainet transactions will be even cheaper and simpler. So with that I will conclude by a couple of things.

Number one, shameless advertisement for EY. We are an ETH maxi firm. Like check out the t-shirt. We are proudly ETH maxis. I don't think there's any large company on earth that is more Ethereum maxi than we are.

Right. Uh I am the chairman of the enterprise Ethereum alliance. I even got Vitalik to write me a little blurb for my book. Um and please note the book is called Ethereum for business not blockchain for business or how to use blockchains in your business. It's called Ethereum for business because that is what we are going to use.

People often say Paul what is your crosschain integration strategy? My answer is I don't have one because I don't think it will be necessary. And finally, uh, when you spend $100 million investing in blockchain and build your own blockchain infrastructure, you can mint tokens. It's not a PO app and therefore not as good as a PO, but you can mint a token. And um, I do try to make a new one every month or so.

And so if you would like that, uh, please let me do that. And with two minutes and 30 seconds left, I'm going to say thank you. And if if I can I use the two minutes for questions. Are there any questions, complaints? I also accepted gratuitous flattery.

I also have some vacation pictures from my last I actually really don't. I mean, no one has ever asked to see No one has ever asked to see the vacation pictures. So, I'm I'm honestly unprepared for that. I think those should be the NFTts that you meant. Nobody wants to see my kids.

If you have any questions, raise your hand. We'll get you a mic and uh we'll do two. The gentleman with the headgear. Uh hi there. Uh exoid here.

Do you do you see any other uh big four that will be entering the space? So, I think we're quite unique in that we are investing in the core technology. So all the big four firms have consulting practices. None of them are as good as ours. Uh shameless self-promotion, but I think we are quite unique in sort of wanting to actually build on Ethereum and have investment in the core technology.

Um and that's something I'm I'm really proud of. It it takes a lot of persuading. When I early on when I was at EY, I sat down at this this partner dinner and I sat down with a bunch of the audit partners and I explained who I was and they said, "Oh, when are you going to stop wasting all this money on blockchain stuff?" Right? It's our money.

I'm like, I'm a partner to my money I'm wasting. But the truth is, you know, 10 years later and I' I've just celebrated my 10th anniversary at DY. We are a large, substantial, and a very profitable business for the firm. So, I I would like to think that we are we've reached the point where it's too late for them to turn back. All right, I think that's our one question, and I want to I don't want to take any time away from this next uh uh speaker.

Thanks, Paul. Please give him a big round of applause. [Music]