From Cloud Storage to Real-Time Computing: A Deep Dive with Ather’s Mark Ryden

All right.

All right.

Good evening.

Good morning, everybody, wherever you are in the world.

It's Jeremy again with another episode of let's Chat to the.

Let's chat to the CEOs and the founders of some great coin projects.

And I have Mark Ryden from Ather.

Welcome.

Awesome.

Thanks for having me, Jeremy.

Pleasure to be here.

Right, can you please tell us you do cloud computing?

Can you tell us what's the difference between cloud storage, which we're used to, and cloud computing, which some people may not be used to?

Sure.

I mean, it's really in the name, right?

So cloud storage is something that I think we're very, very familiar with in everyday life, right?

We've all got a phone, right.

We're very used to kind of offloading our photos from our device into the cloud, right.

We realize that our phone or a local hardware device only has so much capacity for storage, and we can kind of take those files and install them on the cloud and they don't take up space on our local device.

Right.

This is cloud storage, right.

It's kind of a cupboard or a drawer that you can put files in and then you can kind of access them whenever you want.

And maybe you trade a little bit of the accessibility, right?

So maybe you kind of have to redownload a photo or maybe it takes a couple of seconds to access it, right?

But that's a pretty good trade for most people, right?

Because it frees up the speed of our phone and the capacity on our phone.

Compute is a little bit different, right.

Instead of kind of offloading a file or offloading something that you're not using, compute or cloud computing is where you actually offload the processing.

So let's say, for example, you're doing something on your device or you're a company that's doing something that requires a big kind of computing engine of some kind, right?

Whether it's a GPU or a cpu, but something that takes a lot of processing power.

Now, maybe if you're a.

A regular person, maybe for example, your phone or your local device is not as powerful as it needs to be in those cases and select cases, you can actually offload the requirement to power that program or that activity to the cloud in the same way that you can offload storage.

So you can essentially say, hey, cloud, can you power this thing that I want to do?

Because my local device is not powerful enough.

And when you start to kind of scale up the use cases, right?

And we start thinking about these very compute intensive applications, like AI training, right?

Actually Almost no company is going to have amount of compute necessary to train their own models.

So they all need to outsource that compute to the cloud because the cloud essentially allows us to access the quality and size of compute that you can only get from a data center or another place that you have this co located super power cluster of compute.

Those are the main differences.

We'Ve looked at previously at coins like filecoin, which is similar again to the cloud storage.

And we've looked at render without outsourcing computing, I guess for rendering videos, Hollywood productions and that sort of stuff.

Is that similar to what you're doing or are you doing something completely different, Better?

It's similar and I think huge amount of respect to those guys for kind of pioneering this kind of depend or decentralized physical infrastructure network narrative.

I think it's funny actually, the OGs of crypto Bitcoin, they were farming out compute from different people.

You could run the Bitcoin program on your computer at home and you could mine a Bitcoin.

You're essentially providing work that the network valued and then the network was rewarding you for that work, right?

And then a couple of cycles later filecoin and Render came out taking a very similar model and finding utility in a different form.

So filecoin, for example, they took a look at all of the storage that's out there in the world and the fact that you probably have a portion of your hard drive that you don't use, right?

And a company who pays a huge amount of money for secure servers, they might have a bunch of storage on their servers that they don't use.

And filecoin effectively developed this protocol that allowed them to go to these companies and say, hey, if you want, you can connect your service to our network.

We can take off a piece of that and we can then redistribute it to to our customer base.

And the customers then have this really cool opportunity where they almost have like immutable file storage, right?

Because let's say you have one server sitting in one data center, 99 times out of 100, that's perfectly safe.

But let's say a black swan event happens and that server gets wiped somehow.

That's all of your business data gone.

But if you use a service like filecoin or Arweave, you have potentially got your file stored all over a very secure network that shares those files across a number of different places and you have this really secure network for your files.

That's filecoin.

Render is more similar to ather in that they do actually provide a form of computer.

They actually do tap into the GPUs, the graphics processing units of generally consumer equipment, but also some enterprise equipment as well.

And they use those to render, usually CGI files.

If you're doing a movie or a game and you need a 3D image or a movie rendered, you can offload this to their rendering network.

The main difference between Render and Acer is that Render is a network where you send a job to it and it will do that job disconnected from the person who wants the work done.

Remember when I talked about compute?

When you're offloading a compute intensive task at Ether, you still need to connect the cloud to that task.

There needs to be this link where you're powering this task with Compute.

But with Render it's like you're ordering from Uber Eats.

You sending out the order and then someone's cooking it and you're not in the kitchen anymore, someone's cooking it and then it gets delivered to you when it's done.

That's the main difference.

Render does cgi, it's generally workloads that don't need to be connected to the user in any way.

They can kind of like farm them out to their network and then deliver it back to the user when the product's complete.

At Ather, where we live in this kind of real time compute world, where the compute that we aggregate on our network in a very similar way to filecoin, right?

We take idle compute from enterprise customers and we repackage it in a way that can be used by big AI companies or gaming companies, and that compute interact with their customer in a real time way.

So if you're doing an AI training run, if you're doing an AI, if you're running an AI application and you need real time compute to happen, Ather's network can handle that, but it's not something, for example, that Render has been built for.

Render is not real time when it comes to compute, more of a static compute system.

Okay, So I guess they're useful as long as we've got Hollywood blockbusters and that sort of stuff.

100%.

But obviously for you guys, and a lot of people don't realize that artificial intelligence needs a lot of computing power, need to process a lot of data and be educated.

So let's have a look at the AI before we jump into the gaming, because I know you mentioned gaming a moment ago, but where is the AI going?

Is it going to be all these smart fridges that buy milk for us, is it going to be self driving cars?

Is it going to be a virtual assistant that does everything that we want it to?

Where's that heading, mate?

All of the above.

All of the above and more AI is scaling at just this exponential rate these days, right?

Every benchmark that we had six months ago has been blown past.

Every benchmark that we had six months ago and six months before that has been blown past.

We're inventing new benchmarks to measure the level of intelligence of these systems every month, right?

They're just getting more and more capable.

So yeah, there's not going to be a segment of, of society that isn't somehow impacted by AI.

And I think it's a positive thing.

I think that people should be more excited than they should be scared.

I don't think it's necessarily an AI takes your jobs sort of scenario.

I think it's much more about how can you get more efficient, better at the work that you want to do, as opposed to a lot of the grunt work.

There's maybe a little bit less fun about the jobs and tasks that we have.

But when you talk about AI getting more capable over time, right under the hood, what does that really mean?

Well, it means that companies are using a huge amount of compute and a huge amount of data and they're essentially taking those two pieces and attaching a model.

I think we've all heard of like large language models these days like ChatGPT3, Chat GPT4 and that kind of trinity of pieces, right?

The model, the Data and the GPUs, which are like the compute engine, those are effectively what you need to increase the capability of a model, right?

Literally, if you have more GPUs and more data, you get smarter AI.

It's a linear, very simple relationship.

That's why you're seeing people like Elon Musk going out there and building these hundred thousand GPU multi, multibillion dollar data centers because he knows that if you throw money at this problem, you win.

It's that simple.

And that's what's kind of driving this outrageous demand for this GPU Compute.

And that's what's driving the demand for what Acer does, right?

Because at the moment you've got either compute coming direct from the source, right?

And you can go and buy from Nvidia, but Nvidia doesn't want to deal with a little fish Right.

They'd much rather sell $10 billion worth of GPUs at a time to someone like know Elon.

Right.

Most people are probably not getting their compute there.

You can go to a hyperscaler like AWS or Google Cloud or Amazon Cloud, sorry, or Microsoft Azure, but they have a large amount of the compute.

They know that other people are really kind of committed to their ecosystem.

So their prices are very high, right?

Extremely high.

Anywhere from 40 to 90% more than the prices that we have within Acer.

Right.

And they have really limited capacity because so many people are trying to get access to that compute through those networks.

So you have this kind of like real scarcity out there in the market.

But similar to the filecoin example, there are definitely big pockets of compute that exist there within the the server rooms of big enterprise companies where they bought that compute for themselves.

And maybe they missed forecast or maybe they had a production delay, but this is really expensive compute that they now don't have a use case for.

And they're paying through the nose to keep that compute sitting in their data center requires a huge amount of power, huge amount of bandwidth.

So then what are they going to do?

Just going to burn a hole in their pocket.

But with they can kind of plug that compute back into our infrastructure network and then we can redistribute that supply to the global compute kind of marketplace, bringing this like net new supply back.

Because the average compute owner, whether you're a big enterprise company or a startup, right, the average compute owner doesn't have the infrastructure to resell their server equipment.

They just bought it for an internal use case.

So solving this problem for them is a massive benefit to their bottom line.

But it's also a really, really good thing to the overall compute market because it means people have access to compute again in a way that encourages builders to get out and build some cool AI stuff, particularly outside of just big companies like OpenAI.

Yeah.

Okay, so there's big companies who have spent a lot on infrastructure that may be sitting idle from time to time because it's not being used.

And you can tap into that and you can also, I guess, democratize computing power for local people.

I mean, I'm not using my computer from 11 o'clock at night until 6 in the morning.

So it's just sitting there and may as well be utilized by someone else.

So I guess the obvious question then, Mark, is how do we make sure that the data is secure both ways?

Because I don't want someone being able to access my computer and look at my pictures of my puppies or whatever.

And obviously the other person who is computing doesn't want me to see what they're teaching the AI to do.

So how do you control that both ways?

That's a good question.

I mean, just to be clear to the people who are listening, actually Ather doesn't touch consumer equipment at the moment.

So we don't use, for example, your computer or gaming PCs.

There are networks out there that do, and I'm happy to talk about the difference between kind of them and us at some point, but we don't touch that type of compute, we only touch enterprise compute.

Right.

So actually the biggest concern that enterprise customers will have when you're coming into their ecosystem, when you're offering a service to them, particularly when you're touching the boxes that they've used or they currently using.

Right.

Is security is the safety of their data, the safety of their kind of secret sauce that they may have for their business.

So we've had to spend a lot of time making sure that that's a very secure place for them to exist.

So yeah, we wouldn't be able to do anything unless we could prove to them that their system was secure by interacting with us.

And I think that's something that only really comes from a very firm decision to position yourself as an enterprise solution.

Right.

Because enterprise customers do have a certain standard.

Startups don't have the same standard.

Right.

Web3 companies don't have the same standard enterprise customers.

They do.

They need a certain performance for their compute, they need a certain uptime, they need the right security requirements, they need the right quality.

And you have to have all of that in a nice neat little package.

And you then need to overcome the fact that you're a Web3 company to get their business.

So yeah, a very important consideration and one that we've spent a long time coming.

People are used to.

There's different levels of security obviously when going on a normal website versus going to your banking website or your stockbroking website.

Does the blockchain enable you to have a whole different security to the normal Internet that we may be used to?

It does, but maybe not in the way that you, you think.

Actually we, we try and avoid having the blockchain in involved in any of the interactions we have with customers.

Right.

Because that's just a whole other X factor for them.

If we're working with a big telco, for example, and we're trying to get them on board to use our compute, we don't want to be saying to them also, how about, would you like using the blockchain as well?

Because it'll confuse them, freak them out.

But what the blockchain does allow us to do is create this trustless environment for our infrastructure participants, right?

Because if you think about it, building a decentralized network of compute or of anything, whether it's storage or GPUs within render, right?

If you don't own the equipment, it's really hard to kind of do two things.

It's really hard to ensure that everyone that's contributing to your network is telling the truth.

You say you're this type of equipment, right?

You say you've got a one terabyte hard drive, right?

Oh, actually you only have a one gigabyte hard drive.

You're lying.

But you want to get access to more rewards by lying, right?

So that's one difficulty within decentralized networks.

The other difficulty, right, is how do you convince someone that's contributing hardware to your network that they're going to get the rewards that you say you're going to get?

Because you're building things in a, in a decentralized way.

You're, you're trying to remove the human element, step out of the way of, of the scalability of your ecosystem.

So you're not cutting contracts, you're not sending SLAs, you're not doing things in like a web two way with, you know, lawyers and paper trails and things like that, right?

You're trying to do things quicker, more efficient.

So the blockchain helps those two things in a really, really key way, right?

You couldn't build this type of network without the blockchain.

From a trust side, it allows us to.

You probably heard that we did a really big node sale, right?

I think we raised somewhere in the region of $130 million by selling licenses to run what is effectively a decentralized digital police force, right?

So we gave people licenses to operate as these checker nodes in our network.

They received salary for doing that work.

But they're effectively out there on the blockchain and it's passive, right?

They don't need to do anything.

It's just they need to run a program on their computer.

But that program goes and confirms.

Oh, you said you're a one terabyte hard drive.

Let's confirm it.

Right?

Oh, you said you're this type of gpu.

Let's confirm it.

So the blockchain allows us to kind of confirm that through an autonomous network of checkers or agents.

Right.

So that's really cool from the kind of decentralized management side, but then from the trust side, we can build smart contracts, right.

That we effectively don't control.

Right.

Whereby if someone does a job that the network recognizes as being done, they don't need permission from us, they don't need to trust that they'll get paid.

They know they can read on the blockchain, oh, here is a smart contract that will execute automatically when I do the job that the network wants.

So they don't need to trust us, they don't need a contract.

They can literally onboard their compute to the network, they can do the work that the network wants them to do and they know that they'll get rewards.

And that's all through the blockchain.

Again, stuff that you just couldn't do without Web three.

Very exciting stuff.

And solving a lot of problems for a lot of people because there's always people who have the power and then people who want to use the power and usually not the same time, 100%.

So let's jump back and go into the gaming because that's the fun part, right?

This computing to train the AIs and teach them to take over all of them.

As you say, the grunt work that we don't want to do.

But how does ather work in the gaming sphere?

Yeah.

So this is pretty interesting.

I think for anyone that's followed the tech sector for a while, Nvidia would be a very familiar name, but not because of its AI connections.

Right.

Nvidia would have been familiar to most people as the producer of the chips that were used for gaming.

Right.

That's how they started.

They produced graphics chips that helped you run games.

It just so happens by pure fluke that the processing on those chips that is really effective for graphics also happens to be more effective for AI than your Pentium or Intel cpu.

Right.

Just coincidence, right.

A coincidence that Nvidia has taken a huge, taken huge advantage of.

Right.

In an incredibly predictive and well thought out way.

But long story short, Nvidia was a gaming company first because they made these GPUs.

Well, our cloud, our decentralized GPU cloud is a GPU cloud.

So we have AI relevance and we have gaming relevance.

So effectively what we do.

Remember that example I gave earlier about how if you have a really compute intensive task right, on your local device, you can offload that task to the cloud?

Well, what are some of the most compute intensive tasks that the average person would ever touch?

Games.

Right, games.

We're so familiar with the idea that, oh, my computer's not good enough to play this game, or, you know, I've got a PS4 instead of a PS5 and the PS5 is obviously better, so it plays better games, right?

Well, hardware has always been this massive barrier for access to content.

You know, you need a newest, you need the newer computer to play the newer games.

You need the latest PlayStation to play the latest PlayStation games.

Well, what if instead of having to purchase new hardware every time you wanted a new gaming experience, what if you could just offload that entire computer to the cloud and stream in real time an interactive game to your PC, Right?

So effectively it's called cloud gaming, right?

But it's where you install a game on the cloud instead of on your computer, and then any computer, any phone, any device can play any game instantly.

It's the easiest way to visualize it is imagine you install the game on a gaming PC in the cloud, right?

And then you set up a camera and you take a video of that game being played, right?

And then you just have a really long wire that goes to person's, person's home.

And they're just controlling the game through the controller, but they're watching through a video feed.

It's effectively like that because everyone knows that you can watch YouTube on any phone or any computer.

There's no heck or hardware limit to watching YouTube, right?

And that's effectively what we convert a game into.

We convert a game into an interactive video feed, right?

So you can just watch.

It's like watching YouTube, but you're interacting with a game, right?

So it's super cool.

You know, companies like Google, a service like this, it was called Google Stadia, they actually shut it down two years ago because they couldn't scale it.

It's too expensive to scale in a centralized way because it's so compute intensive.

You have to purchase all of this compute, you have to pre buy it, deploy it in a location, then you have to kind of sell subscriptions because you have to offset the cost of all that upfront purchase.

But the challenge here is that it's 20 to 25 bucks a month.

That's what kind of Google charged.

But where do the people benefit most from this technology?

Right?

It's in places where they can't afford expensive equipment, right.

This tech is not best suited for people in the US or Australia or Europe, Right?

Because we already have, you know, probably an Xbox or a PlayStation, maybe we already have a gaming PC.

But Southeast Asia, Latin America, India, these are places where 90% of the world's gamers live and they all game on low end devices and they can't access high quality games.

But this technology was never able to reach them because it was too expensive.

The people that can afford a gaming PC or can't afford a good phone, they also can't afford 20 bucks a month for a cloud gaming subscription.

But as soon as you can deploy this tech in a decentralized way where the network doesn't need to own and pre purchase all of that hardware to begin with.

Right.

We don't need to buy a massive data center and put all these GPUs in it just to buy the service.

We can aggregate underutilized equipment all over the world and do the same thing.

All of a sudden then the unit economics become like, they become scalable into the regions of the world where this technology has the most value.

And you know, it's 3.3 billion gamers out there.

Roughly 50% of the world are gamers.

80% of 2.8 billion are gaming on low end devices.

They don't have access to the best gaming content that's being released every day.

Right.

But also biggest gaming companies in the world don't have access to them.

So this is a huge opportunity if you can unlock those gamers.

For the biggest gaming companies and publishers out there.

It's the single largest kind of untapped market left within gaming, which is, you know, arguably one of the biggest industries in the world.

That's extraordinary.

You're giving me flashbacks to many years ago when I used to play Tomb Raider on my PC and the first Tomb Raider that came out was very boxy and you know, triangle graphics and that sort of stuff.

And it played okay and it got me hooked on the game.

It was a great game.

And then a couple of years later they bought out the sequel, which much more curvy and you had drips of water and blood and that sort of stuff.

Yeah.

But I couldn't run it.

It would just lag.

It would be really, really slow on my computer.

I'm like, I've just spent $90 on this game and now I'm going to have to spend $800 upgrading my computer so I can actually play the thing 100%.

It's nuts.

And it's a dance that we're so used to doing, but it's one that I don't think we should, I don't think we should have to do.

I think abstracting away that consumer hardware barrier is 100% possible.

The tech already exists.

We just have to scale it better.

And that's exactly what Ace is doing on the gaming side.

Fantastic.

So in order to, I'm imagining that's a slightly different section of the business and you've already quoted somebody else was charging $25.

So you got to tell us how much you actually charge for this and how people access it if they're into gaming.

Yeah.

So this is actually free for users, right?

So yeah, the way that you, the way that we think about it.

Right.

Is in gaming the main metric for a publisher is the annual revenue per user.

Right.

If they onboard a user into their game, how much do they expect to earn back from the user?

Right.

Because a lot of games these days are free to play, particularly on mobile.

Right.

So they earn through microtransactions or buying digital assets like a skin or an in game item, things like that.

Right.

But they have an idea.

They'll know, okay, one gamer is worth $5 over a year or $20 over a year, whatever it is.

Right.

If they come to a service like ours and we're the only ones that can kind of do this, but if they come to a service like ours and we can say something like, and I'm not going to give you the real numbers because these are a little bit kind of secret sauce.

Right.

But let's say they came to us and we said okay, on Acer it costs you 50 cents to host a user on our network.

Well then can do a pretty simple calculation, right?

Okay, one user is worth $20.

If I host that user for 50 cents then I can cover that cost pretty easily.

And then it gives me access to 80% of the market I can't touch because my game's too, too intense or too difficult to run for the most, most of the hardware in that region.

And then it just becomes a decision of how, how much do you want to turn the tap on, you know, how many users do you want to onboard and can our network scale fast enough to help you do that?

So I can play the most advanced game possible for free in high risk?

Yeah.

I'll say this ather doesn't provide a consumer facing solution for this.

We work as the infrastructure behind other solutions that do the kind of the direct to consumer angle.

So some customers and clients with will charge a subscription to their users, some won't.

But certainly I think, and this is why we started the company, I think the best way to get value out of this tech is to offer it free to users.

But then you do have to make money in order to keep running the company you do.

But you still make money through the gamers the same way that you would normally.

Right by them buying in game assets.

Or maybe it's not a free game.

Maybe they pay $80, $90 for the, for the game itself.

You still make money in the normal way.

You're just choosing to host the user in a different way.

So it's usually just a slight added cost.

So is this the technology that you're calling the Aether Edge?

No, this is our cloud.

This is our kind of ather atmosphere service.

Right.

Play on words like ather atmosphere cloud.

No, the Edge is actually our physical kind of mining device.

It's this thing here, super cool.

Actually looks like a short fat PS5 to be fair.

But this device, what we realized is, remember how I said before, we don't use consumer equipment, we don't touch your computer or another computer and we don't allow you to contribute compute to our network in that way.

We wanted to reduce the barrier to entry.

We wanted to give people the ability to operate as a compute provider in our network.

So we set up this Aether Edge device.

So we actually partnered with Qualcomm, you know, one of the biggest semiconductor manufacturers in the world.

And we have this device available for purchase.

People can buy it, plug it into the power, plug it into their network and it will then act as an enterprise capable node within ACES infrastructure.

And it will mean that companies enterprises, they can push work to this device that's using your power and your WI fi.

Right.

And it will do work, process it on that chip and then provide that completion of work to the ather network.

And as a result you guys will receive rewards.

Right.

So it's actually a very.

You know, when Bitcoin had Bitcoin mining machines that you could buy like specific ones.

Right?

It's kind of like that or helium, you could buy a specific like helium WI fi router.

It's just like that.

This is the AI version of a Bitcoin specific mining machine.

It just allows you to own a piece of AI infrastructure and get rewarded really, really well for contributing that to the network.

Okay, Are we allowed to give away pricing structure and reward structure?

Yeah.

So this not reward structure.

Legally I'm only allowed to say so many things with relation to the token, but you can see what the reward structure is and you can access these devices for sale there as well.

So you can actually go to MyEdge IE so MyEdge IO you'll be able to purchase a device directly There you'll be able to see what the price is and see what the average rewards have been over the past months since the product was launched.

We actually did a big $20 million airdrop to all of our edge and check node holders just a couple of weeks ago.

So we regularly expose our ecosystem to these really cool rewards that come through our network.

So, yeah, it's a great, it's a great place to be for sure, as a.

Owning one of these nodes because there.

May be some people out there who have invested in like a little bitcoin miner, particularly people who have got excess solar power or excess power or something like that, they can whack the bitcoin miner, miner in the corner.

It also functions as a handy room heater because it generates a lot of heat.

This one, this one is, is no more power than a light bulb, or at least that's what my tech team tell me.

Very low draw, super efficient.

It won't heat up your room, you won't even hear it.

It's not like it's got a big buzzing fan.

It's super, super silent and really passive.

Plug it in and it just works.

I think that some of the bitcoin miners, they say if you buy this one, it's going to do so much processing and it'll pay for itself within five years or it'll pay for itself within three years or something like that.

So do you make any sort of suggestions about roi?

Yeah, that's what I'm not allowed to talk about, but it's on the website, so my Edge IO, you can, you can check it out there.

Is there any limits to these machines that I can have?

I wanted to have 20 of them in the corner of my.

Let do that.

Yeah, absolutely.

I've.

I've.

We've got people in our network that have hundreds of these.

We've got people setting up warehouses.

We're actually even producing a commercial version because obviously this type of compute is really.

People and enterprises want it and it's a really kind of exciting place to be as a provider, being able to say, hey, I'm contributing to this sort of compute network and I'm earning rewards for doing so.

Passive income's massive.

It is quite small a unit that you showed me and I'm imagining myself I might want to buy 20 of these or 50 of these, but then I need 50 PowerPoints to plug them in.

So can I buy a really, really big one instead of buying 20 little ones?

You can.

That's coming.

Actually, I don't even know if anyone knows that yet.

I don't think we've made it public.

I wonder if I'm even allowed to say it.

But yeah, it is.

We have a bigger one coming, effectively more of a.

More of a server version.

Right.

Where it's kind of like 20 of them in an array.

Brilliant, brilliant.

I think that could be very attractive for people.

Obviously, you know, the gaming thing is going to be a game changer, as you say.

There's many people without dated equipment and they want the fancy new software to run on their devices.

I mean, I've got an iPhone that I think is two steps behind what everybody else is carrying and some apps don't run on it.

So, yeah, I think that's absolutely extraordinary.

We're just about out of time.

I appreciate your time so much.

Is there anything else that you want to tell us about why everybody should buy your devices?

Why everybody should buy your token?

Yeah, I mean, these devices are.

The Aether Edge is just really cool.

Right.

I think it's an opportunity to participate in absolutely booming sector.

Right.

You're providing very useful infrastructure for businesses and companies building in the AI space.

You'll be rewarded for doing so.

I think that is just a very good decision in general for a number of different reasons.

From a broader kind of token and product perspective, I'd just say keep an eye on our socials.

Some really, really cool information that'll be coming out maybe today or tomorrow that will kind of highlight a little bit of a step change in what we're doing at Aether, kind of expanding the scope of what we deal with on the compute side, which I think is going to be very, very cool and just signals that we're taking this next step in our evolution as kind of an AI and infrastructure project.

And that's.

I think people are going to get really excited about that.

Well, you seem to have a lot of bases covered, obviously.

I mean, I'm not a real big gamer.

I gave that away when I was talking about a game that's like 15 years old.

But I know from one of my friends who's high up in the music industry and he said that people spend more money on games than the amount of money they spend on music streaming and movies.

So it's huge.

It's absolutely monstrous.

Trillion, trillions and trillions of dollars all the time.

So.

And AI obviously is just expanding like nobody's business.

Everybody's into that.

So you're on the cutting edge.

You're doing the right things and you're democratizing it for us ordinary people who want to be able to capitalize on that.

I don't have a server that I have access to, but I can certainly put a few devices in the back corner of the office.

Absolutely right.

Perfect.

Perfect use case.

Fantastic, mate.

Really appreciate your time.

Mark is going to give me all of the links to all of the socials and things like that so I can put it in the description box.

And Mark Ryden from Ather, thank you very much.

No worries, Jeremy.