[Music] With this, we're ready for our next talk. And for this one, I'd like to bring on Nandit to tell you all about tokenizing the data economy. So, please give Nand from Lighthouse a big round of applause. Cool. So, hey everyone.
My name is Nandit and I'm the founder of uh Lighthouse. Um just before we start, we would love to give a overview about what Lighthouse is. So at Lighthouse, we're building permanent decentralized storage. Um and uh we offer IPFS and Filecoin uh as the underneath infrastructure. Um and my journey um uh into lighthouse was even like you guys I was one of the global hackers uh and then after few years u now I got the opportunity to be like mentor and judge at couple of events and now glad to be speaking at this pragma event as well.
So before we jump to the main topic of tokenizing the data economy I'll share about what we are actually solving at lighthouse. So yeah there are a couple of things which some of which that s also touched in her talk. Uh the first one is high storage cost that storing important data be it AI NFT uh and even blockchain history per se long-term uh is expensive and most importantly hard to verify. The second thing is the lack of security. Most of the decentralized network right now uh they are um all public so they don't allow you to store private data or build token gated apps.
So what that uh um like hinders users is uh just imagine that if you want to store your AI data your private data um it's like a big hassle to you to okay how do you store your data on a public uh network especially like those big AI data sets and open-source like llama models etc. And third thing is the fragmentation. When we started building in the ecosystem, we realized that for developers uh all the SDKs are quite fragmented that I have to go to one SDK for storage, another one for uh maybe encryption, maybe a third one for verifying and setting up my own CDN retrieval things. So these were the problems uh that we saw uh while we we started building lighthouse. Um and how lighthouse solves this is first of all we have this concept of permanent storage that you can pay once and store your data long-term or forever.
I'll explain more in a while in upcoming slides. Second uh we have this access control where you can tie up any data to onchain logic be it by NFT ownership by token ownership by time based uh access control etc. And then we have one unified lighthouse SDK which allows you to do all of that and avoid any fragmentation during your um development journey. And for that we have like this web app uh NCL SDK as the popular tools which is trusted by right now thousands of developers to build apps and startups as well. So talking about the permanent storage part first uh we have this concept of endowment pool.
So what essentially happens is using the power of smart contracts what lighthouse is able to do is when you store your data you pay an upfront fee to our smart contract and over a period of time um the smart contract releases funds to the uh miners. These miners are primarily like filecoin miners but it makes that you don't need to worry about data replication data renewal or any of the data headaches that come with uh decentralized networks. Uh second thing is all our contracts for endowment pool they're deployed on Falcoin uh virtual machine. So um with that using things like XLR's squid router folks are able to build crosschain apps with uh other ecosystems as well primarily around D5 where these endorment pool of funds uh they are actively connected to uh things like a things like glyph for the staking part of the piece where your data which is stored um for the upfront payment it keeps on growing with the yield that it generates via D5 apps as well. The second most important part of our stack is encryption and access control.
So we use something called as BLS threshold cryptography. So what it essentially does is when you store your private data um like every file gets their unique encryption key and that encryption key is splitted across different lighthouse nodes and next time when you have to come and retrieve your data back you just give a signature from your wallet and that signature could authenticate your NFT ownership your token ownership it could be a custom onchain logic or it could be even a time based access per se uh and then when you give your signature like the lighthouse house nodes, they receive that signatures and individually on their end they verify that does this user have access to the file that they are claiming to be and if they have access to the file uh those lighthouse nodes share the key shards that they have uh and all of those key shards as are aggregated uh and decrypted on the client send so that the authorized user could have access to the file let's say it could be if I want to have some data to be accessed by only pudgy penguin holders or people who have thousand and unis swap tokens on uni chains those kind of logics um can easily be created so we've been fortunate enough to work with some of the top uh AI NFT projects out there one of uh which I would like to highlight is eternal AI so eternal AI is for example uh what they do is it's an open platform where anyone can run uh open-source models be it llama model quen other models and uh it's their agents are live on Ethereum but all the data of those uh agentic models but also the core foundational models which could be llama all of that is being served from uh lighthouse right now. Uh and what that allows us to do is um with eternal AI and their mobile app um anyone can access all of these models uncensored like although no political opinion but I was very much worried uh with the recent India US uh whatever the things were happening then I was like okay what if Trump tomorrow says that okay we are cutting access to open AI to India so already we have open models but there if there's a way to serve them in a more verifiable open way using eternal AI AI. I was now not much worried that okay even if open AI goes away we have eternal AI to uh read serve all of these models and have chat interfaces easily to uh run um all the AI apps and agentic things I wanted to do and some of the other applications to highlight are ocean protocol singularity uh who have integrated lighthouse to uh store data across their data marketplaces. In terms of NFTs, we have NFT storage.
Uh they not just have uh Lighthouse as the storage back end, but now they have an onchain endowment pool that anyone who is storing NFTs uh using the NFD storage interface through their latest product. All of that funds is going to uh NFT storage pool and which is like kind of like a public good where um other NFT collections, some top marketplaces, openc etc. they can easily contribute for the long-term storage um of the NFTs and some other um like apps like okay contract Alexandria they are also doing some cool like Alexandria is preserving all the uh important archives and books on top of lighthouse which is quite cool um and then coming to the deep end as well uh we have some very cool products being built like web hash is one of those products which now have 10,000 websites deployed um and connected to the ENS domain. So they now also serve I think SNS which is the Solana naming service and some other uh naming services as well. But more than 10,000 websites are daily indexed and retrieved through lighthouse uh on web hash platform and aether and stack OS they have also integrated lighthouse to access filecoin storage easily uh using the stack and um all of these things that I'm talking about they are in production hundreds of these uh apps are using lighthouse already and right now we have around 30 tips of uh data that's being stored across miners and more than 7 million um unique data objects which itself are have like a further uh like tens of thousands of objects per um this data object that I'm talking about.
But I think while building lighthouse um another thing that we realized that uh while lighthouse solves a lot of problems around data storage and encryption um the data problem in the overall ecosystem on how do you get the data where the data lies who controls the data was still not solved. uh for example at the moment as well most of the data is siloed and undermonetized. Most of the consumer apps, enterprise apps, they control all the data and maybe they want to use decentralized storage, maybe they don't want to use decentralized storage but consumers are on the behest of all of these um consumer apps or enterprise apps be it Spotify, Uber or whatever that we are using on a daily basis that uh we have to go to these social platforms to view and access the data. Apparently what we saw was um like from last six months as we were working with eternal AI, singularity net and bunch of other AI apps that yeah AI data hunger is real like like all these models which are running on our phone like Gemini uh chat GPT perplexity etc they are like very hungry for data interestingly most of them have already eaten all the public data that is out there like everyone has read I think all these AI models have read YouTube they have like scraped the out of uh Reddit and other platforms as well. So where does the next data opportunity lies?
So probably that data opportunity lies um in not the public data out there but where the private data is being stored which is like in our consumer apps in the enterprise applications and the other part of the picture is synthetic data like not just AI consumes the data that humans produce but using uh human input. Now AI wants to consume the data which AI itself is producing. So uh that's why we see that web3 technology um and like our evolution from lighthouse uh fits really well into this picture on not just storing and encrypting the data but how do we feed this AI data hunger demand especially around the private data and synthetic data there that these foundational models are very hungry for right now and using that insights and another insight that we had at lighthouse was uh we increasingly saw that people want to not just store data but also retrieve the data on a very frequent basis. That was very evident from our bandwidth cost going up and a lot of retrieval requests like tens of millions of retrieval requests that we have been getting at lighthouse across NFT AI and deepen platforms and using that inside uh we built this platform called as 1 MB.io MB is like everyone knows the uh unit of storage which is megabyte.
So 1 MB.io IO is your gateway to not just store and encrypt data on lighthouse but also to monetize, manage and move your data uh in the age of AI. So what 1 MB essentially is is now using Lighthouse you cannot just store and encrypt data but you can tokenize and monetize the data that you are storing underneath. So the basic flow would look like that you as a developer you build a application like a front end could be a mobile app. Uh you integrate our SDK and you upload and encrypt data to lighthouse or you enable a subset of users uh like which is like your application users to upload data.
Once they upload data um they could mint these data coins or data tokens um whatever the term you would like to call it um and then you can easily trade them in the marketplace. Um so essentially 1 MB is a platform where you can not just create these data coins but also curate the data coins and also monetize the data coins um as well. Um primarily how do we verify that who to give mint uh who to mint data coins to uh can be done through technologies like ZKTLS. So at Lighthouse we have a way on uh if you use ZKTLS to bring in real world data um we can verify those ZKTLS proofs for you. So you can imagine an application uh maybe something like Uber kind of application where you get 10,000 users in next 3 days to connect to your web app uh front end.
uh use ZKTLS uh to enable them to um import data to your app, export from Uber and then uh we let you verify that data using uh ZKTLS that you know that the data which is coming from Uber is correct and then these users can easily mint uh the data coins uh and the inherent demand of the data coins uh could come from monetizing it further to these um AI models as well. So right now we are still building that part where you can um like easily have a playground on okay how do you fetch this data and monetize the data but for the hackathon what we are proposing to users is to uh create these data coins and data do application where you connect it to real world data verify using zkls monetize from day zero and let it trade in the open marketplace and for that uh we already deployed it mainet on ethereum polygon on base and world chain because we saw like those are some of the hackathon other participants um bounties out there. So how does the current platform architecture for uh this 1 MB looks like at the bottom layer we have the perpetual storage back end uh and also the encryption access control that I talked about um and in the middle layer is this data coin discovery and uh datacoin launchpad which is 1 MB.io IO and on top of that we foresee hundreds of data coins which can be created in various domains like it could be uh users giving their AI data still privacy preserving that you need not give your full information using ZKTLS. It could be NFTs like we already have NFT storage with their endowment pool live.
So we can foresee that NFT storage could potentially have their own coin where anyone who wants to fetch that data uh they could potentially pay in that NFD coins and then easily retrieve and stream the data as well. And deep in networks I think they have already proven that um they are the economies where people connect hardware into these GP networks. generally they have to buy the hardware but then uh whatever the verification mechanism is like in case of high mapper helium etc people contribute their mapping data their IoT data their maybe wireless uh kinetic data whatever that could be so there's already deep in proving the way that yes if we enable uh crypto rails uh to allow people to curate data maybe in deep in that is because of the hardware but we see this data coin platform can enable more ways for people to uh create data coins and also potentially monetize them and bring real data from day zero uh from your hackathon project launch itself and already we are seeing like data overall um the AI data hunger growing a lot like if you see most of the big deals like with this open AI and Wall Street Journal like $250 million uh deal that they had uh last year that is like over and over proving again that uh data hunger is real and the data monetization market is already growing like more than 40 50% perom it's expected to be like tens of billions on of dollars by 2030 and as you can see in these graphs like hivemapper and helium they are able to very easily curate um with their hardware mechanism verify and then allow further uh marketplace where that data uh can be sold to enterprise clients but now with this platform 1 mb.io IO it becomes like a one-stop platform where not just existing deep in data could be accessed but uh developers could create new data coins uh with things like ZKTLS etc. And about the hackathon itself, uh we had like some of these ideas posted along with filecoin in the bounties where I think Uber as a right sharing dow could be one interesting idea like can we have 10,000 users uh to not share maybe their emails preserve that on a privacy preserving manner but then share all their 30-day ride history every month and on some UBX tokens and some people would argue that okay why would these people why would these companies allow that Again I'm not I'm not like a big fan of regulations but like for example article 20 of GDPR does say that uh consumers have the right to data portability that anyone who wants to port the data from one controller to another they have the right to easily port data and you know most of these uh companies they are allowing nowadays even Facebook allows you to export delete whatever data that you want.
So you can create easytouse web apps where uh you can bring real data from day one for these data coins. You can using zkls like zumato could be another example very famous in India where can I have let's say maybe gura sector 33 10,000 people to contribute their data and potentially like monetize it in the future as well or like some other applications are also possible here. Cool. Awesome. Thank you.
[Music]
Automatic transcript — names and jargon may be misspelled.