Hi everyone, it's Rachel Peter for FinTech TV and we are live at Salana breakpoint here in Abu Dhabi.
I am joined by Jack Collier, who's the Chief growth Officer of Io.net.
Jack, thanks so much for coming on the show today.
Thanks for having me.
So let's start with some overarching basics.
What is IO.net?
Tell me more.
So IO.net, it's a Salana-based decentralized GPU net marketplace.
And what we do is we essentially get GPU providers from all over the spectrum.
Whether that's just somebody, you know, with a 4090 or a GPU chip at home running a server at home all the way through to large data centers.
We kind of pull that supply and we serve that to customers.
And so what that means is we're able to offer a diverse set of GPUs at on demand, so people can come in and flexibly spin up a cluster, spin down a cluster.
We offer for a cheaper pricing than, say, compared to AWS and Azure and GCP, and, and we can do that in a way that is in keeping with all the crypto values.
Like it's transparent, it's fair, and it's sort of a new way of being able for people to get access to GPUs without paying the big price tag.
Incredible.
And you've, you've also had some big announcements recently.
Can you talk us, talk us through those as well?
Yes, of course.
So.
We use a blockchain obviously to uh support the whole network.
So uh an example being um every time someone purchases a GPU or rents a GPU on our network, all of that record is kept on chain chain.
So you're able to see the amount of revenue that goes through the chain.
You're able to see what type of devices they're buying, what, you know, whether they, what the up time and things are like in relation to those devices.
And the way we encourage suppliers when we first launched the project back in July was um we did a supplier-based incentive program with our tokenomics.
So there are emissions that come from that token, and we use those emissions to incentivize people to come and join.
So great, every deep pin in um basically in existence has done this to to sort of bootstrap their network.
So it's encouraging lots of suppliers to come and join.
Now it gets to a point where that actually is no longer a sustainable model for a, uh, a deep end network because as price goes up, those supply of those, of the token goes up, the suppliers are getting a lot of value out of the network.
And as the price goes down, um, obviously the, the value that they're getting goes down.
And so they, they're not protected by the economic shocks of the token itself.
Um, and obviously with it being an inflatary system, we're constantly giving out, uh, rewards for supply, it's not really sustainable in the long term.
So what we've done yesterday was we launched the IDE, which is the incentive dynamic engine of how we treat uh tokennomics.
And what that does is it flips it on its head.
It goes from a supply in supplier incentive to a demand incentive.
So as we sell GPUs, then we emit to be able to reward suppliers for the participation in the network.
And we've made a few other changes such as we denominate the rewards in US USD.
So we still give IO token as a reward, but that's in USD.
And what's quite novel is we've also created um what we call uh volt a volt structure.
So as the price of the IO token goes up or down, um, we keep a reserve within those volts that we can use to basically make up any rewards, um, any reward difference or incentive difference to be able to ensure that a supplier gets a consistent payout for their device.
And so.
What that means is it's still encouraging suppliers to come and join because we're able to sell, we're at the point now where we're, we're selling a lot of our demand to, you know, web 2, web 3 clients.
We have over $20 million in revenue already on our, on, on the platform.
Um, and it, it means that they're gonna, they're gonna be able to predict the amount of rewards that they get at the end of the day.
So it's good for suppliers, it's good for the network, um.
Uh, and I guess the last thing I would say is within those vaults, we're also able to decide to burn tokens should we wish.
So if the, the size of that vault gets, gets too big, we can just burn a lot of that supply, which will reduce the inflationary pressure on the network that we, that I mentioned earlier, um, which is just a great, more self-sustaining way of running the token as opposed to relying on an endless supply of sort of inflationary tokens.
Yeah, 100%, and you spoke about the supply demand dynamics and, and inverting that.
I'd like to break those down into two pieces.
Where are you, where are you seeing most of the supply coming from?
Yeah, great question.
So when we first bootstrapped the network back 22 years ago, um, a lot of our supply was from consumer grade GPUs.
So people just running, you know, one or two, chips or nodes at home.
Um, and over the first sort of 345 months we were also working with larger data centers to get more enterprise grade GPUs onto our network.
So we, we've done a really good job of being able to attract a large and diverse set of supply.
We're alive over 138 countries across the world, so wherever you need to spin up a GPU, we kind of have that on our network.
And whether that, like I say, is a high grade premium H200 device, say, or a, you know, RTX 4090, you're able to do that.
Um, and so, yeah, whatever use case you have to be able to, to solve for, I feel like we have that, you know, whether you're price sensitive and, you know, need to use a lower grade chip set, which are often aren't available with the large, uh, hyper scales anymore, um, or whether you need a cheaper enterprise grade device, we have that.
So that's on the supply side then, so moving to the demand side, is it mainly on the retail side, institutions?
Tell me a bit about the makeup of that.
Yes, so the majority of our customer base are actually like non crypto based projects, so.
And they're mainly AI fights, right?
So in AI, there's, especially as it's taken off now, um, you need a large amount of compute power to run the models, train the models, whatever it is.
And they're, they, they don't care whether you're a web, like whether it's a web 2 or web 2 solution, they just want to make sure they have cheaper, more flexible access to GPUs.
And so, yeah, we, we kind of market to that audience, uh, primarily.
Um, not to say we don't market to Web 3 customers, we have a large group of webre customers as well.
But what we're really trying to do is break down those barriers and bridge the gap between web 3 and web 2.
I, I mean, I, my personal conviction is that I think Deepin will be the next big breakthrough in crypto.
You know, we've, we've broken through on, on the finance side, on the payment side, and I think, you know, being able to.
Use blockchain to be able to uh solve this complete crunch and this, this crisis we have um on that side, I think, uh, I think will be the next narrative because like I said, we, we already have $20 million in total net worth earnings on our, on iO.net and that's only, that's only growing.
And, and it's because like I say, the, the technology of blockchain is able to solve for this problem, you know, of where people can are looking for GPUs and are having to pay too much or they just can't get access to them.
I love this, the narrative around this of using web 3 to solve a real world problem.
I almost feel like oftentimes it's the other way around, whereas we try and create something out of web 3 or we create it in web 3 and hope that it solves the problem, but this is just a perfect example of, you know.
Yeah, and the founder's story is like they came across this issue first, this is why they wanted to solve for it, you know, they had a need, a large need for compute power.
And they couldn't get it, you know, I imagine you're like a, a startup with, you know, you've got very, very, you're very short on funds, you know, you may seed round or whatever it is.
You know, we're looking at most AI startups spending 60, 70, 80% of their, of their total cost on just compute power.
And often that doesn't need to be the case, you know, you go to AWS, you're looking at paying $20 per hour for an H200 device.
Whereas on iO.net you're looking at around $2.
So you're making an incredible saving.
And that's, and that's because these big providers have had, had it too good for too long, you know, they, they're trying to dominate the market and they often, um.
Are biased towards certain types of companies versus others, you know, so they can manipulate who's trying to succeed and who isn't.
And so we need to come up with these novel solutions.
I mean, one of the good things about, um, uh IO itself is that we're able to, instead of us owning lots of data centers and building out more capacity uh across the world, which is obviously energy intensive, climate intensive, that type of thing, we're, because people can opt the redundant supply onto a network to get it utilized.
We're actually increasing utilization rates globally of GPUs, you know, which can only be a good thing, you know, and it's, uh, it's just making things more efficient and uh and if we're able to pass on all those benefits to, to customers at the end of the day so that they can innovate more, then that's, then that's fantastic.
Well this has been such an interesting use case.
I mean, I love the story, I love the ethos, I love the evolution that you've had, so thanks so much for coming on the show and telling us about it having me.
Cheers.
