Connectivity Credits: Tokenizing Telecoms for A Better Huma/I/nternet

Transcript

Good afternoon. It's, you know, hearing all of the discussions this morning, um, and being really immersed in this AI moment, we do hear a lot about energy as the driver for, uh, for kind of compute and infrastructure. Our team has been working for the past six years at another part of the defining railroads of the future. And we consider that part uh just as important as electricity, although maybe a little less known. We work on tokenizing infrastructure assets, particularly infrastructure that drives the internet. And I'm going to talk to you about three things over the next 20 minutes. First of all, I will give you a little description of how

messed up the world is right now. We all sort of know that, but I'm going to dive into it in a little bit more detail through the lens of connectivity. Second, I'm going to talk to you a bit about Giga and the work that our team has been doing over the past four or five years with the Ethereum Foundation and some of you in the room on building infrastructure to get ready for what we heard from Vitalic and Audrey and others this morning is definitely coming. And finally, I'll just take, you know, one slide to ask you for some help because what would it be uh what would a talk be if we weren't building it together? Uh, I saw this headline yesterday as I was getting ready. France is uh voting to put a tax on incoming internet traffic. I I don't really know how that

is going to work. I know a bit about how the internet works. That one seems a little harder to enforce. Um, but I put a picture of it on the talk because it expressed to me how little we know about how the internet is actually built. um how difficult it is for governments to understand what this asset what this railway track that they have running through their countries is doing and how maybe as we think about the spread of AI we're missing a big important piece of the puzzle around the world right now there are about two billion people who don't have access to the internet which is maybe a little bit hard to imagine when we think about how connected Mumbai is and most of India and how much we take advantage every day of apps that

are kind of in our pocket and connect us to the world. But it's true. So, we've got about a third of the world without good connectivity. And we have an asset, the the gigabyte, the mysterious gigabyte, which is priced completely differently in every country. I wanted to show you three pictures here. This is from a live map. So, if you want to take out your phone and scan the QR code, you can open up maps.g.global. This is a map of every school in the world. We've spent the last five years building uh one of the larger deepin projects in the world in terms of thinking about how decentralized physical infrastructure and networks work. You need to know where those networks are first. And so in 2020, we started trying to map all the schools in the world. It turns out that schools are a very good proxy for

human networks because where there is a community, there is a school. Whether that's a big school with lots of floors or it's a small school in a in a little wooden house or if it's a school under a tree, parents want their kids to learn and so schools are a part of almost every community in the world. Um, interestingly enough, countries don't always know where their schools are. So, we built the machine learning models in 2019 2020 to help find schools from outer space. But this map is the result of that. So, if you go to the site there through the QR code, you will see a map of the world. You'll see about 54 countries with green or red dots on them. They'll look something like Brazil or Botswana or Sierra Leone here on the board. And what this is is a live map of

global connectivity. The schools with the green dots are connected to the internet. The schools with the red dots are not connected to the internet. And the ones that are blinking are giving us data every four minutes. So if you look at Botswana on the live map, you'll see a bunch of blinking green dots. You see countries with a lot of connectivity like Brazil. You see countries with some connectivity in wealthy areas like Batswana down in the south and there's a desert and then there's no connectivity at all up in the north. And then you see countries like Sierra Leone, one of the poorest countries in the world where if you go to the map and you zoom in and zoom into Freetown, you can see about 18 or 20 connected schools.

This is the difference for countries and for people between being involved in a digital future and not and infrastructure is so important because if you are trying to build access to edge models or even localized non-frontier models uh not on edge you need somehow to have those agents talk to each other to have the models talk to you and to have infrastructure connect. We also know that there's a huge overlap between connectivity and the uplift of a country economically. So this is one of my favorite maps in the world. It shows the un how unfair the world is. Something like 7 65% of the world's money is in the green area and something like 15 16% of the world's people are in the green area. And if you look at the map of connectivity, there's an

interesting correlation. We actually did some research recently that shows if you connect all the schools in Sierra Leone, you raise the GDP by something like 20%. This is all to just describe a problem. The problem is that we're living in a world where parts of it are accelerating incredibly fast and pdoom is whatever it is and then you have parts of it which are so disconnected from the network that it's really hard to say how they will interact with the future. This is the one kind of like lesson slide. So sorry about this but it's a quick refresher if one doesn't know exactly how the internet works. There's no shame in that because we're so used to using it through a phone that we forget there's some fundamental infrastructure before it.

The internet that we know of, the way that agents communicate with us, that we communicate with LLMs, that we communicate with each other, starts with these subc cables with big deep dug infrastructure, it goes from there to these kind of infrastructure companies in the middle and then somehow it goes to the very end, usually through wireless radio frequency or through smaller cables. And I put this slide up just to show that there are right now a bunch of different pieces that are all speaking differently together. And this is why I' I'd ask if it's the new barrel of oil. It's really interesting because if you were buying oil in the same way, somebody might buy a I don't know like a barrel of oil. Somebody might buy a an hour of running time for their car and somebody else might buy a kilo of

plastic derived from petroleum and that might all be considered oil. That's how weirdly priced the internet is because each of these three areas talks about their asset differently, trades it differently, and moves it differently resulting in huge price disparities. So, if you've worked with connectivity or if you've traveled anywhere pretty much uh across a road that looks like this, you'll know that where things are better arranged, where houses are more connected, where the the fiber cable runs through the road, you have cheaper and better internet. In places that are more informal, where there isn't that connectivity between all three levels of that system, you have spotty internet, it goes out and it's much, much more expensive. And I think of this fundamentally as a problem of

coordination and of describing an asset. Right now, the way that internet is purchased is insane. It's as if one of you wanted to rent a flat in a city. And instead, you were told, "Just buy the entire city." And you're like, "No, I'd like to spend $15,000 to purchase a flat or $100,000 to purchase this house." And somebody said, "Oh, well, you can have the entire city of Tokyo for 10 trillion dollars." That's the type of packaging that the internet is sold in now if you're trying to talk to the big subc cable providers and the other people on that diagram that I showed. And so what we started thinking about in the last 2 three years with the help of the Ethereum Foundation is whether or not we could slice up

these assets. We could take this thing that exists that we all understand at some very basic level, but we really don't know how it works and try to make it something smaller, something that was packaged in a way that there could be more liquidity. It could be moved around and it could be pushed into places that it wasn't right now into all those places with the red dots. It turns out, and this was a little bit before like we were in AI crazy time, but it turns out that if you are an agent in a network and you're trying to move resources around, you also need them to be described in some way that you can manipulate them. And so, it's become very interesting to us as we've seen the rise of this AI moment. And we've seen the fact that we have some of these

assets described in a way where agents can take them and manipulate them that we think that this may be part of building a set of railroads that we weren't even expecting before. I'll give you one more educational slide on how the internet works. Uh if you have a phone in your pocket, it's running on radio frequencies. Those radio frequencies are defined by a government within the confines of the country and each country has the ability to sell its own radio frequency. You may know this if you've, you know, been walking inside a big football stadium. You can't you can't suddenly call anybody anymore.

It's because that frequency is all blocked up by all the other users. The way this is currently executed is that the government sells or rents those frequency blocks to a few big phone companies for like 10 to 20 years at a time. That asset is locked in. It belongs to that phone company for the entire country. And you really can't use any other services within that which means it's up to that one phone company and there might be five or six or three in a country to provide services that are dynamic and ready for people to use at the edge of the network. It is not often in the interest of an incumbent to do that. Often, if you're a big telephone company, you've already built your infrastructure. You've built it where wealthy people are. And you really don't need to include poor people

because to you, poor people may be worthless. It's hard to figure out the value of them. It may be more risky. So, you stick with your big block and you don't let anybody else use it. This poses an interesting problem because if you had some way to take an asset to describe it digitally to fractionalize it and to share it with some common ledger where everybody could agree on what was what you might be able to create more liquidity in this system. You might be able to create a system where you have dynamic allocation of resources. Uh luckily we have that thing. I think everybody in the room understands how easy it is to execute the idea, but how hard it is probably to operationalize uh you know the work with

governments and and all the stuff on the back end that you need in order to get something like this to turn into something that's more uh active. So what we've done over the last year is started to slice radio spectrum. Um we've started to treat it as one of the network primitives in the telecom world. And that means that we can identify a piece of that spectrum. We can allocate it dynamically and we can let the government even resell it or rerent it which creates a certain mobility inside of that system and creates a lot of money potentially for the government or the person who bought it originally. But most importantly being able to take that big asset and break it into a smaller thing allows for new actors to come in and provide connectivity. People who were blocked out before by the big

internet service providers or the big mobile network operators. um you can go to that link and see more about spectrum certificates. But what we're uh what we're starting to understand now is how quick and dynamic you can be you can be if you're a country that doesn't have a great uh you know a huge incumbent network already. The other thing that that allows us to do when you have something and you've sliced it up into you know a small box and you can you've made it mobile and usable and sellable is you can create a marketplace. As I said before, there's no spot price for the gigabyte. A gigabyte is not even a real thing. In many countries, you buy a gigabyte on your phone. You buy like, you know, a five gig plan or a three gig plan and it expires at the end of the

month. Has anybody ever bought a a gigabyte plan? Yeah. Okay, we've been there. And it expires as if it's like an avocado and the thing goes off at the end of the month. That's not true, right? That's not how the internet works. But it's a it's a company making money on the fact that you have to reby that that thing later on. By identifying the asset and slicing it and segmenting it um we are able to create a system where you can trade this asset. And so what we've done now in a few countries is built a connectivity credits marketplace. This is the first time in the world that we're getting close to a price for connectivity. And this is an asset that varies in price. In Zimbabwe it's about $43 per gigabyte. That's what it would cost in neighboring Mozambique.

It's about 73 cents per gigabyte for the same avocado. You're paying, you know, a huge difference in in money. And that's crazy to me. And it's just because of inefficiencies in markets. Often when these spectrum bands are allocated, it's a bunch of people in a room with wooden paddles bidding on them for 15 years. And it's crazy how old the system of slicing purchasing defining connectivity is for us. We've created a marketplace where anybody can come online, a small internet service provider, a large mobile network operator, and start to put their contract prices on. And it's all onchain. So they can see what's happening and they can trade their contracts with each other, which is very important because it also gives us, all of you, anybody who looks at the ledger, it gives you an idea of the price of

this asset in real time. And we've seen this drop the price of connectivity, the ability to have this kind of clarity of price by up to 55% in some of the 54 countries that we're working in right now. Um, I just wanted to give a quick case study, an example of what what's happened in Brazil. If you look at this small map, this is Brazil in 2019. Uh, you'll see that the Amazon Amazon region is pretty red. It's pretty disconnected. Uh, and now if you look at the map and you can see it, you can check it live because you'll see all those dots are blinking. You'll see that the country is pretty well connected. And a bunch of stuff came together to make that happen.

One thing is obviously the rise of satellites and LEOs as part of the as part of the internet equation, internet infrastructure. Um, but three things that we helped with were really really important and they're all part of this marketplace that I was just describing. First of all, in Brazil, interestingly, there there about 11, 1100 different internet service providers that can bid on government contracts. Uh, we were able to work with the government and to open up uh the contract pricing. We've done the same thing now across 50 countries in Africa. Um, so you can see the published wholesale prices of connectivity across the country.

The other thing is that once you have real-time information on the delivery of something, you can pay for it in real time. If you have a school or a health center or another government building that is an Oracle that tells you what's happening, uh, you can start to create a mechanism for paying those service providers not on a yearly basis but monthly. And third of all, when you have software installed in the school, and now we have that little blinking dot coming from 140,000 schools around the world. And if you look at Usbekiststan, there's something like 12,000 schools in Isbekistan that are all blinking green.

That's in only in the last few months. When you have that metering, when you have that ability to understand what's going on, you can start to uh make the network shaped in a way that provides more connectivity for people who need it. And that's really an interesting part of an agentic network because suddenly you have all these volumes of data and you need some way to sort through it and to um kind of define the pathways for information to flow and we can do that now. So Brazil is a wonderful case study of of how that can work. Um I wanted to give a quick update because since uh since ETH Prague I'm I'm kind of timing my night life now by like Ethereum related events. So there's a little timeline on the bottom. U but since ETH Prague uh we've now got four countries which are working with the

radio spectrum certification project which means that there are four governments uh where we have access to the regulatory and the ministries regulatory folks and the ministries of IT to allow us to in their sandboxes where they test out this technology um tokenize radio spectrum which is pretty crazy. Now our project giga works in 54 countries around the world. So we're connecting schools and health centers at scale in a lot more than four countries. Um but to have four governments that are really looking to understand how these assets can be shaped and moved in real time is quite exciting to me. A second thing that's happened in the last few months is that we have uh more than 70,000 kilometers of fiber that are being traded on the connectivity credits

marketplace. Most of that is happening in Mexico. There's 6,000 small internet service providers there that are uh starting to move connectivity around on this marketplace. But we're also running a similar system in Malawi and South Africa and in Kenya and a marketplace needs a buyer. So what we've done now is connected companies whose interest it is in to get more connectivity. Companies like Equinex for example where a lot of the internet cables land in a country. There's usually an Equinex building which connects the cables. Countries like Equin, companies like Equinex are starting to pay to buy credits to connect more rural places in the country. They know it's in their business interest to get the last

quintile of people, the last 20% of people, two two billion people online. They know it's in their interest to make sure that that's fair, that connectivity is fairly distributed. They're not the mobile network operator or the one in the middle. And suddenly they're playing a role in this marketplace. And so we've watched this ecosystem start to develop. Now, one of the big questions for us is like, can that marketplace connect multiple different countries? We don't know that yet. That's going to still to be discovered. Um, but we do know that within a country, at least we can start to see the oil price, the barrel price for this asset that nobody was really talking about. And it's interesting because if you are building a data center, you price in a lot of things. If you're building a data center, you price in the GPUs and the land. You price in

the stuff you need to cool it, whether it's fans or water. you pricing the people, of course, the electricity. Uh, but right now, connectivity is usually not on that pricing schedule. Not really. You kind of take a guess at how much it might cost. Um, but it's not really a firm line item there. And I believe one of the contentions of this project is that over the next few years as you get a lot more traffic moving over the subc cables and through the satellites and through the networks uh we you me are going to be displaced in favor of traffic that can push itself onto networks and that traffic is going to be agentic. that traffic is going to be AI that's able to understand how the

networks work and move itself around them faster. And that's interesting because we hear that there's going to be like a hundred billion and then a trillion and then a few tens of trillions agents uh you know all around us in in the creepy future. Um they are going to want to talk to each other. They speak differently than we do. By which I mean they use internet protocols differently than we do. They have different needs for speed and latency and all the things that networks are measured on, but they're hungry and they're going to need this infrastructure that is so badly designed right now that you cannot get from the barrel of oil to the hour of driving time without jumping through 15 different names and auctions and owners.

Uh, and the work that we're doing with this connectivity marketplace is starting to allow us as humanity to understand that asset. Um, and so I think and I think that's why countries are coming on board. So, I said I would end with an ask. Uh, I have three asks. I'm a greedy person. Um, first of all, we're working at significant speed with governments in 54 countries. We work with the ministry of finance, with the regulator, with the network operators. Those aren't the governments, but there's a huge connection, close connection between them. um and with the parts of the government usually office of the prime minister or others who are making these connections between uh you know the gross domestic product and the ability for people in their country to be connected to modern infrastructure.

We'd welcome further connections and contributions from other governments and asset providers, the people who own the cables, the people who own the satellites uh are often interested in being part of this because they know that they can spread uh more of their network capacity downstream. Secondly, if you're interested in getting involved, I've put four fresh jobs here in that middle uh that middle box. Please share them with your networks, with your friends. Feel free to apply if that's something that's interesting to you. A lot of these jobs, and you can look through them later, are about how you do the go to market for this weird technology which somehow crosses blockchain and AI, but I think I've gotten through the entire presentation without saying blockchain once, maybe. Uh how do you sell this to a government? How do you sell this to uh

telephone companies who might fear it without knocking the table over? And finally, if you'd like to learn more, in May, uh we'll be hosting ETH Geneva, uh obviously in Geneva, where we're going to have a focus on Agentic Networks and the tokenized assets that run them. Um so we'd welcome you to contribute in the variety of ways that you can contribute to an an ETH event. Uh and I hope to see some or many of you there. I'd love to have any questions in the last few minutes. Thank you so much for being here today. Um, I want to do them in this like fire. I want I want a a yellow mask and to do kind of fire popcorn questions, too. Uh,

let's see how fast we can go through these. How do you make sure whales don't buy up and constrain supply? Uh, whales already have. That's why the internet doesn't work very well right now in poor places. I think the first thing that we can do is to start to slice and itemize this thing in a way that you can see it publicly. Um, and then we can use pressure from a policy and regulatory side to get governments to create allowances for those smaller fish to play. And we've seen this happen. So that's why in several countries we've seen the price drop by like 50% for connectivity to schools and health centers when we do work like the work that I've described. Is Ethereum a primary network? Uh, so I don't I no longer work at UNICEF innovation. I loved those days. I can't speak to it.

There's some wonderful presentations by the UNICEF innovation colleagues here. Are you guys in the room there? So, uh you can see those presentations over the next two days. Um at Giga Ethereum is the primary network that we built on the spectrum certificates are built. We've built our own chain of blocks for spectrum certificates because it's in early early stages, but it will be ported over to L1 and we believe firmly in open source and permissionless public networks. uh are we exploring partnerships with Starlink direct? Uh we they are a provider in the marketplace that I described. So if you look at the Mexico marketplace, Starlink is one of the providers. They provide actually backhall. So it's not the direct to phone uh but they provide a backhole

issuance of access uh to us in Mexico. Why don't you test these concepts in developed worldwing in we're working in Spain and Switzerland. So our we're I'm based in Switzerland. Uh we're working with the regulators in those countries. The uh you know we talked recently to the FCC in the US, the equivalent organization, federal communications organization in Canada and in the UK about this and what they said is that the ability for governments to be innovative and to be ahead of uh technology in smaller countries is tremendous and those countries are often even more locked in by the incumbent big network operators.

So, it's harder for them to break through what's really an unfair uh unfair system. And if you look at the price of gigabytes, it's actually much higher in wealthier countries than it is in in poorer countries. But we're building this uh globally and we're it's and you can see on the map that it's, you know, there about 54 countries where we're where we're working closely. Uh why not invest in your own connectivity infrastructure? uh as soon as you try to grab the whole thing I believe you become compromised uh and we you know on Vitalik's triangle one of the things was like dominance and domination of the world as one of the bad things that you can get to. I really believe and our team believes that the way that you get

better access to these resources if is if there is a network of different providers who have their own incentives. Um, so we're working with companies that have the subc cables, companies that have the satellites. Uh, but I think that as soon as you try to grab too much, uh, it it becomes very it becomes very scary. If a becomes a tradable asset, um, there are no competitive telecom markets right now. That's basically the problem. Uh, telekcom is an area that is, and it's not, you know, that's not exactly true, but I'll stand by that statement.

Telekom is an area that is has been built in the last 50 years. The telecom stuff that we know really in the last 30 years. It was not built in a way to promote competitiveness. Uh even in countries where there are anti-competitive tools like universal service or universal access funds, it's not competitive. And the first thing that you need to do is figure out how to shift that into a place where there is cooperation either by force or by need. uh and I believe that the idea of creating a marketplace where the price is available is one of the best ways to force or create that cooperation. Um we saw that actually in about 2002 2003 in the space of vaccines the vaccine market is another one that's tremendously

anti-competitive. It certainly was 20 years ago. Uh UNICEF put out an editorial and said there you know in 20 2002 there were only four or five companies selling key vaccines for children. Uh they their prices were all different. they were all hidden and nobody had access to the pricing list and UNICEF was able to force that pricing list into the open because of that vaccine price dropped 80% uh India and Brazil and South Africa produced their own vaccines and that market became open and I believe that access to fair safe connectivity at an individual level at a national level at a sovereign level is going to be one of the vaccines for the future for this scary future that we