Unifying global cloud infrastructure with Impossible Cloud Network

Conor Svensson: Yeah, would you like to start by introducing yourself to everyone?

Kai Wawrzinek: Yeah. Perfect. Hi Conor. Thanks again for having me. Yeah. So, my name is Kai. I'm a serial entrepreneur. I founded a gaming company a long time ago and that was pretty successful, very large, like more than a thousand employees. We served a few hundred million players all around the world and we did a reverse merger to the Nasdaq, sold the company and then it was like a few more years on the board, doing lots of M and A acquisitions, capital market experience and so forth, from the listed entity there in the gaming space. And after moving out of that I was really thinking a lot about the cloud space because we had lots of exposure to cloud and infrastructure networks on the gaming side as well. Like thousands of servers. Later, we moved into the cloud. We probably spent a few hundred million just on AWS alone. And I always felt, this is something that we can potentially revolutionize and in combination with Web3 and incentivizing a lot of people to help us to build a decentralized version of the cloud. When this idea came three years ago, we founded Impossible Cloud Network. I founded this together with Christian, my co-founder and here we are. And it's exciting times.

Conor Svensson: Awesome. And so this, from what I understand, there's almost two components to it because you have the Impossible Cloud, a sort of company which talks more about the applications and so on of the technology and then you have the underlying network as well as. Is that correct?

Kai Wawrzinek: Yeah, that's correct. So essentially, maybe if you give me one or two more sentences on that. So essentially we started with a regular German based company. And what we did just a few weeks ago, we separated everything into a second part. So the protocol, the network, everything is being transferred over to a Swiss foundation, completely separate. They are the gatekeepers of the overall vision, the protocol, the connection to hardware providers, a connection to service providers and everything else. The German entity will just stay there with one service which is object storage and selling these services over to real world business enterprise customers and so forth. The reason why we did this is when we started in web3 we thought okay this is like super amazing what web3 has done before on the supply side, building large networks, really massive amounts of hardware connected into hardware and networks. But so far, Web three is in my opinion really lacking behind in terms of user experience, real world products, especially on the cloud space, compliant, performant, up and running and so forth. And this is what you really need in order to bring mass adoption to Web3. And that's the ultimate goal. We want to bring mass adoption, Web two mass adoption to Web three technology. And that's the reason basically why we started this way and the success and the scale of the product is really proving that's a good choice.

Conor Svensson: And so if we unpack the product a little bit more and the underlying problem that it solves as well. Because you mentioned at the start that with your previous work with your gaming venture that you had a lot of exposure to cloud technology and it just felt like there's a lot of problems there. And cloud, of course, encompasses a number of different technologies and so there's multiple ways in which you can be working with it so to speak. So yeah, I suppose, if we unpack the Web three part. So we're talking more about just general companies who are working with cloud technologies and the problem it solves for them.

Kai Wawrzinek: Yeah so the cloud space is a huge market. It's a trillion dollar market. It's by far the biggest profit stream of Amazon AWS and is growing super fast. It's extracting from all the verticals because all companies all around the world need to move into the cloud. And it's dominated by more or less three companies at least in the western space, which is AWS, Microsoft and Google GCP, Google Cloud Platform. So these three companies probably have 70% of the overall cloud market. And there's a lot of shortcomings because of that. So first of all the amount of services that they have is so big and it's so convenient for a lot of customers that they really feel this lock in effect and the price that they pay for it is really high. And more and more customers, more enterprises that we see around the world are having concerns about putting all their data and everything into these large tech companies’ clouds and there is a lack of differentiation, it’s lacking other opportunities. And when we started the service just a while ago, enterprise customers were running into our doors and really asking for more and more of these opportunities to put their data somewhere else. The problem is what I saw in web3 products so far, usually that other web3 products are having a hard time being on par in terms of performance, in terms of compliance. You need a lot of integrations. So building an enterprise grade product is super difficult and it takes a lot of time. And that's one of the reasons why you see this dominance of these large cloud providers. And essentially besides the price and the lock in effect you have other concerns. GDPR is one of the problems especially for European customers, because essentially you cannot ensure if you put the data with one of the large big tech companies based in the US there's always the risk that this data is handed over to US authorities even if this is going against European law. And this is one of the additional reasons why many, many companies nowadays either run their on premises data centers themselves which is kind of cost intensive or are looking desperately for some outside options.

Conor Svensson: And so by leveraging Impossible Cloud, what they get there is just more control. Would that be the best way of summing it up and not having quite that same vendor lock in that they have with the big tech platforms.

Kai Wawrzinek: Yeah. So essentially what customers get is a very high performance and more control of the data. So data sovereignty, no lock in effect, a better price which is much, much lower than what they have to pay with a hyperscaler. So probably it's 80% less. And this, in conjunction with all the integrations that we can offer is super compelling for them. And also especially in certain use cases you don't want to put all your eggs in one nest. You have these single point of failures if you just go with one large service or one large cloud. There have been outages over and over again, even on the largest cloud offerings because they essentially are built very central and for a very good reason. The Internet itself was not built central. It was more decentral to avoid unavailability or bottlenecks, single point of failures. This is what our service can offer in exchange. And so that's I think in the end, it's multiple reasons why customers really like our service and the approach. and if you look into the future, the long term future companies start to realize that also there's a lot of inefficiency, creating data at exponential rate all over the place, at the edges, everywhere and then moving that data to central data centers, and doing the computation there and moving it back. Right? So a lot of the future use cases are also going the direction that you need much more decentralized networks where you have more proximity between where the data is being generated and where the data is being stored and processed in the end.

Conor Svensson: This is fascinating to hear. You referred to, what in effect, are all very compelling reasons for people to actually want to work with the platform in terms of greater control performance, lower cost and more decentralized. So it's the natural follow on question from that is how is it possible with the platform?

Kai Wawrzinek: Yeah. So generally the beauty of that is like we are not building everything on our own, that's the whole idea of Web three. Essentially you are incentivizing others and building together. And when I was working in Web2 for a long time, there was a kind of common knowledge that it's almost impossible to create any kind of competition against the largest tech companies because they have such strong moats. Right? No way to compete, maybe except for demonopolizing them through something like a political campaign or whatever. But the beauty with web3 is no matter how large a company is, no matter how many talents they have, even if it's AWS or Google or whatever, there will always be more smart people outside. And if you're able through incentivization to let a lot of people all around the globe work together towards a common goal. This is the super strength, I think this is the superpower of Web3, one of the superpowers. And this is what we are essentially doing, connecting all the data centers around the world. At the moment, we have most of the data centers being located in Europe and the US throughout, next year there's many, many additional data centers coming. Just to give you a number, in Germany alone you have close to 10,000 data centers. So it's a lot of tier two, tier three data centers. They have a lot of underutilized capacity, which they are really having a hard day. They have to really keep this for peak usage times and so forth. And it's just much more efficient to connect their capacity into a network that can potentially serve a lot of customers on top. Right? And these approaches I think will come inevitably. So from my perspective web3 is maturing. So it took quite some time obviously but it's just about to mature. It's still early days in terms of mass adoption for Web3 but it's inevitably coming. And the efficiency gain of connecting these data centers is pretty high.

Conor Svensson: So in that respect it's very much utilizing this excess capacity provided by data centers that is one of the core angles for achieving this. I mean of course there's the actual protocols being built to make it happen and the token related incentives too for such a network. But the key thing is this idle capacity?

Kai Wawrzinek: Yeah. So that's the starting point. Right? So idle capacity, there's a lot of idle capacity out there. Essentially that's the same how AWS originally started. So they had a huge capacity to be built for the Black Fridays and Christmas times and whatever for their shopping services. And during other times a lot of their capacity was idling. And essentially this is the same with many, many data centers all around the world. And adding this capacity is just beneficial to everybody. Then on top, I would definitely say if you open this up and if you incentivize people to be part of the network and to also participate in future success, the amount of innovation will increase. You have seen this on many occasions also in the past, innovation is just a lot faster. If you decentralize, if you have many, many people working towards a common goal, there will be much more innovation than even in the largest companies. So I'm foreseeing that this is going to happen as well. You see this even today if you look into Bitcoin mining for example, one of obviously the oldest web 3 approaches there and how it started with mining on whatever, private laptops or so computers, and nowadays it's so optimized, people have their own hardware customized and are serving out of super large self built data centers, close to Hoover Dams where the electricity is super low, whatever. Something that could never be achieved by a central organization in that efficiency. And so down the line there will be a lot of efficiency gains. And then lastly in terms of use cases, I think a lot of use cases are just much better to be served from a decentral network. If you think alone about throughput, throughput is super important and that's pretty common sense already, since 20 years ago or 25 years ago when like the first peer to peer network started, that if you connect in parallel to multiple sources, your connection is just much much faster, your throughput overall will be much, much faster than if you just connect to one endpoint. Right? And that's also something that is super beneficial. Proximity. What I mentioned before, like CDN Network, started 15 years ago for good reason. Whenever you have a connection it's beneficial to have the connection as short as possible. Like physical distance and a decentral network on average the distances are also much shorter. So I think there are many, many reasons why decentralized networks are coming, and why the world needs decentral networks. And our goal is basically to build a platform where professional enterprise grade services are being built on top that everybody can use.

Conor Svensson: And so what kind of starts off as this initial thesis, around these data centers and so on could really, really expand to any sorts of cloud services longer term. And do you envisage that actual software services that are provided, could be provided on a cloud scale as well. So once you move beyond the actual physical infrastructure and you have whether it's thinking about some of the virtual, you've got storage but also like virtualized compute, the database and so on and so forth, do you envisage a world where all of these could come in, where the software angle becomes a big part too?

Kai Wawrzinek: Yeah, absolutely. So again on the hardware side, but as well on the software side, there is no way for any startup in the world to do everything on their own. On the service side especially we said okay, we need to demonstrate first one service at least to be running, to be operational, being used with a lot of customers. And that's what we have proven. So today we have more than 1,000 enterprises like typical Web2 enterprises. They have no wallet, no crypto and so forth. They are just using our service and they essentially are storing data on web 3 hardware. Right? So this is what we have today. Then in terms of timeline we are opening this up, the system. We are launching a token which will facilitate token launches planned for Q1 next year. So pretty soon we will open the system and with launching the token there will be a lot of improvement on the supply side. So the hardware side will grow quite quickly from there. So there's probably 20 or 30 different locations joining the network globally very, very soon. And then in terms of service, what you just mentioned, on the software side we are already partnering with quite a lot of other DEPIN projects and are in preparation for second services. So there's still of course some stuff that needs to be built, some interfaces and APIs and so forth. But the customers that we have today are already requesting compute on top of their data, so it makes a lot of sense. Obviously, they store a lot of data with us and they want to compute the data on top. The question is just do we start rather with an inference, sorry with a serverless computation or do we rather start with some kind of GPU computation, which is also doable. There are multiple different options, and we are just evaluating them and 2025, next year, will be the year where the foundation is more and more opening this up for other service providers to provide more services on top to the benefit of the customer.

Conor Svensson: In terms of, I know you touched on like use cases here, but also thinking in terms of verticals mentioned GPUs just now and with AI being such a sort of crazy growth industry right now, no doubt there's some natural synergies there. But what are the verticals that you're most excited about, for the technology at the moment? I know it can apply to everyone who uses cloud these days, but by the same token there must be those specific ones which really excite you.

Kai Wawrzinek: Yeah, so we started for good reason with storage because essentially every vertical, every company, every market, needs storage. Right? And what's also super interesting is what is always called or referred to usually as data gravity, or data locality, meaning wherever you store the data, usually it makes sense to move any kind of service nearby. So moving data is very costly. Usually, you want to store the data in some place and then you want to move any kind of computation, CPU databases, GPU computation and so forth, close by. And there are a lot of these options. We just need to investigate, kind of, in evaluation what other services are making sense. The problem of course here again is developing an enterprise grade service is not an easy task. Just the object storage alone if you like. Again, I very much like numbers. So there's like one Web2 competitor that we are referring to usually. And obviously if you provide enterprise grade services for two customers, you are in competition with Pure Web 2 products. And just to put this in perspective, like this Web2 competitor who just built object storage, their claim is to be the fastest of the fastest and the cheapest of the cheapest. They had to raise around 500 million as of today, obviously because they are running all the data centers on their own and it took them more than 20 million USD to provide the MVP object storage product. And that's not even including any initial marketing. So providing these kind of enterprise grade services is time consuming and cost intensive because the requirements in terms of performance, uptime, integration features and so forth is quite rich. For that reason it's not an easy task to move into multiple additional services. But anyways we have very good relationships to companies like 8year and some very good relationships to others also in the deep end space that are already today offering other services that could potentially be quite neatly integrated. and then we open access to the customers that we have already.

Conor Svensson: The system and kind of related to that point is this this vendor locking that happens a lot with a lot of these services for enterprise especially now. How is it that the service can prevent that and provide that ongoing optionality for your customers?

Kai Wawrzinek: I don't know if I got your question 100% correctly. So in terms of. So maybe you just rephrase it?

Conor Svensson: So yeah, sure. So just on the vendor lock in. I know it's something that we mentioned at the start but it is a challenge for companies that want that optionality with the Impossible Cloud Network. How is it that it helps provide that ongoing optionality so that companies if they start working with it, they're not going to be tied to it in the same way that they would necessarily with the status quo so to speak.

Kai Wawrzinek: Yeah. So one very easy thing is usually if you store your data in AWS, GCP, Microsoft Azure and so forth, putting the data there is completely for free. But getting your data out is super cost intensive. Right? So that's part of their lock in. That's something we don't have. So you can put your data, you can retrieve the data. It doesn't cost anything. Additionally, just if you have a certain amount of data that you store then you pay on a monthly basis per amount of storage being used. Right? So I think it's a very fair and transparent model which is much easier to do than much less cost intensive and also like the hidden costs are not there. If you look into your AWS bill for example, it has a tremendous amount of cost, line items. There are even calculators out there trying to estimate your calculation. And what we see really is a lot of companies all around the world, they start with very low cost because it's so convenient to start something off in an AWS or GCP cloud platform. And by now their builds are tremendously high and everything is really, really entangled and it's very hard to move. And nevertheless a lot of companies started to move away from the clouds, from the hyperscale clouds because of cost and hidden cost items. And so I think this is a non fair monopolistic approach that they play. I think down the line we can also revolutionize the space.

Conor Svensson: Super helpful. And so we've got five minutes left and I know that we're probably not going to be able to get into a lot of detail on this but thinking in terms of the token and how it factors in, you mentioned already how you have customers who are using the service but they're not exposed to the crypto part of it. But obviously the token economic model is no doubt going to incentivize participants on the network. So how will the token be used from the perspective of the actual network itself, but then also for potential customers and users as well? How does it play into that side?

Kai Wawrzinek: Yeah, so with a token a lot of different participants are able to take part in the network. First of all you have the hardware providers providing hardware and earning token rewards at the same time. Hardware providers have to use some tokens as collateral. This is in order to ensure that they behave in the right way, that the services are up and running and so forth. And so the collateral might be partly slashed if a hardware provider is not having a very good service. So let's put it this way. This is in order to ensure the network operational ability. Then of course you have customers who just buy any kind of services and they pay fiat money USD or Euros or whatever and that money needs to be transferred by service providers into tokens. So there's created a lot of buy pressure , probably from that perspective. So that money is transferred to tokens and the tokens can be taken to use the services and to access the hardware. Then there is an additional layer which is super important is what we call the SLA Oracle nodes. So people who are providing some kind of hardware, they can buy NFT licenses. and with these licenses, these node licenses, they are able to participate in the network. Even if you don't have a huge budget or whatever, you can buy one of these licenses. The sale is going to happen in just a few weeks and also with providing these services and having bought this license you are rewarded with some tokens. Right? So the token is basically a utility that aligns the interest of all the participants in the network and which really helps the network to grow and deliver a scalable and performing service.

Conor Svensson: I really like hearing that sort of analogy of referring to NFTs as kind of being like licenses as well. That's the first time I've heard it. But I think it can be a useful mental model for some people who aren't looking at this from a web 3 centric lens as well. In this capacity.

Kai Wawrzinek: Yeah, absolutely. And again for us the ultimate goal in order to make this happen, the goal of building a decentralized version of AWS which is beneficial in the end for everybody. The ultimate goal will be to onboard as m many people as possible to work jointly towards this common goal. And that's also one of the reasons why we are building up on the community side, encouraging people to take part in this, to potentially buy an NFT license in order to run a node and participate in the network. and all of this combined will be super strong and rewardful for everybody.

Conor Svensson: Absolutely. And so your testnet has just launched, and you mentioned about the upcoming token sale. What else is on the horizon and if people want to learn more, how best should they engage with the community?

Kai Wawrzinek: Yeah, so right now we have around 200,000 people in the community, mainly on X, Discord and Telegram. Just join there, keep an eye on all the updates coming. You can still join the testnet. The testnet is using some air drop mechanisms so people are rewarded for helping us test the stability, the functionality. and also we want to continue to build community engagement with the Testnet. So please join the Testnet and test things out. You can also just go to the website ICN Global and everything is connected on there. You can also of course test the object storage product, if you want to store some data, that's also doable, just to get a feel on how this works and then in terms of timeline, after the node sale like four weeks from now roughly. There's the token listing as mentioned before, upcoming in Q1 next year. From there on there's a lot of scale coming to the network in terms of additional hardware providers at massive scale being onboarded, additional services coming next year with ongoing partnerships, especially in the DEPIN space, and from there, kind of to the moon. So that's the goal, to make this larger and larger. And with 1000 enterprise customers that we have already today, it's just a starting point. We see more and more customers coming and I think this is also super beneficial for anybody who is running some kind of service, company and so forth. If you want to build with us, there will be some grants, there will be opportunities to work jointly together, provide additional services, investigate with us what you can build on top of the platform and then everything is open also to all the customers that we have and any customers that we will win in the future.

Conor Svensson: Awesome. Well Kai, just to say thank you so much for your time, it's been a really enlightening space and I look forward to hearing more and more about Impossible Cloud as it grows in the next year.

Kai Wawrzinek: Thank you Conor, so much, much appreciated. Thanks.

Conor Svensson: Bye.

Kai Wawrzinek: Bye.