Latest / Crypto Coin Show / Aethir’s Web3 Cloud Powers AI’s INSANE Processing Demand | Blockchain Interviews
Transcript
- 0:00Speaker 1: I'm Ashton Addison from the Crypto Coin Show, and today
- 0:03Speaker 1: on Blockchain Interviews we have Mark Ryden, co founder at Aither.
- 0:07Speaker 2: Mark. Welcome to the show and thank you for taking
- 0:09Speaker 2: the time.
- 0:11Speaker 3: Thanks for having me. Yeah, great to be here, Ashton.
- 0:14Speaker 1: Yeah, I know Ather's been crazy busy with the launch
- 0:19Speaker 1: of the platform and the token, and it's a pretty
- 0:23Speaker 1: hot area decentralized cloud, graphic processing, AI and gaming, which
- 0:29Speaker 1: I'd love to dive into. I would love to start
- 0:32Speaker 1: off our conversation for those who aren't familiar just a
- 0:35Speaker 1: high level on what is everything that Ether is working on,
- 0:37Speaker 1: and then we can dive into all the latest details.
- 0:41Speaker 3: Yeah, for sure, we're an infrastructure project, so sometimes it's
- 0:46Speaker 3: better for us to start with the problem we're solving,
- 0:49Speaker 3: give people a bit of contact, and then I can
- 0:51Speaker 3: jump into explaining exactly what we do. So taking kind
- 0:54Speaker 3: of a problem for its approach. I think a lot
- 0:57Speaker 3: of people in your audience have probably heard the phrase
- 1:00Speaker 3: compute scarcity for GPU scarcity, Right, people like Elon Musks
- 1:05Speaker 3: have come out and said things like GPUs or harder
- 1:08Speaker 3: to buy the drugs. Right. You've got people like Mark
- 1:12Speaker 3: Zuckerberg and Meta for just saying, you know, just under
- 1:15Speaker 3: eleven billion dollars worth of Nvidia's top GPU chip for
- 1:20Speaker 3: for their kind of LAMA three AI ambitions. Right, it's
- 1:24Speaker 3: get increasingly more difficult for dot ups and small and
- 1:29Speaker 3: medium sized enterprises to access this uh critical GPU resource right,
- 1:35Speaker 3: critical for AI companies, critical for gaming companies. At AH
- 1:40Speaker 3: we kind of studied this problem quite deeply, right, uh
- 1:44Speaker 3: and and and we've broken it down really into four
- 1:47Speaker 3: distinct pieces. The first is obviously by pressure, right. By
- 1:53Speaker 3: pressure is driving scarcity in the market. The biggest tech
- 1:57Speaker 3: companies in the world Open AI, Google, my for Stoft,
- 2:00Speaker 3: Amazon and Trophic. They all essentially have an insatiable appetite
- 2:05Speaker 3: for GPUs, right. They know that for them to win
- 2:08Speaker 3: the AI rate, they need to purchase more GPUs. Essentially,
- 2:12Speaker 3: more DPUs means training AI quicker, and they tap train foster. Right.
- 2:17Speaker 3: So you have this sustained, ongoing BI pressure that makes
- 2:20Speaker 3: it really difficult to get access to these resources. Second,
- 2:24Speaker 3: we have the inability of Nvidia to produce chips FoST
- 2:29Speaker 3: enough for the market. Right. They can only produce a
- 2:32Speaker 3: certain number of tips per year. They're doing their best,
- 2:34Speaker 3: and you know they are the biggest company in the world,
- 2:37Speaker 3: or they swap places with Microsoft every couple of days.
- 2:40Speaker 3: They're doing a dance, but they can only produce so
- 2:44Speaker 3: many chips, and as I mentioned, it's not yet enough
- 2:47Speaker 3: to kind of meet global demands. Thirdly, we have what
- 2:50Speaker 3: we call kind of increasing LLLM complexity. Right. So LLM
- 2:56Speaker 3: stands for lags language models, right. It is the name
- 3:00Speaker 3: for the technical term for chat CDT three or CHATZDT
- 3:04Speaker 3: four or LAMA three. These are all versions of large
- 3:09Speaker 3: language models, right. And what we're seeing is that later
- 3:14Speaker 3: generation models are exponentially more expensive to training. And when
- 3:19Speaker 3: I say expensive, I mean from a compute perspective, from
- 3:24Speaker 3: a data perspective, and then also from a cost perspective. Right.
- 3:28Speaker 3: So what's happening is as the companies that's trying to
- 3:31Speaker 3: make their models more intelligent, right, so that they become
- 3:35Speaker 3: more effective tool for us to use, they're finding that
- 3:38Speaker 3: it's also costing them exponentially more GPUs than it did
- 3:42Speaker 3: just a year before. Right. So we're seeing this exponential
- 3:46Speaker 3: ramp up of compute, but we're not seeing a similar
- 3:50Speaker 3: ramp up in supply or a similar ramp up in
- 3:54Speaker 3: the performance of chips. So we're getting this like growing
- 3:58Speaker 3: gap between what's required of later and the chips is
- 4:00Speaker 3: coming online. And then finally, and this is where we
- 4:03Speaker 3: get into I guess, really the sweet operator. Finally, we
- 4:07Speaker 3: also have the fact that these resources are really poorly managed, right,
- 4:12Speaker 3: as a resource that's out there, they're incredibly inefficiently managed.
- 4:17Speaker 3: And we did a lot of research on this. So
- 4:19Speaker 3: we found out that when you take a look at
- 4:21Speaker 3: enterprise GPU resources sitting in data centers in the US,
- 4:26Speaker 3: for example, they're sitting it anywhere from fifteen to eighteen
- 4:31Speaker 3: percent utilization rates, right, so they're underutilized often nearly eighty
- 4:37Speaker 3: percent of the time. Right. So we thought to ourselves,
- 4:40Speaker 3: if we could create this infrastructure layer, right that would
- 4:44Speaker 3: allow us to onboard compute from a variety of different
- 4:49Speaker 3: owners and then surface it really really simply to the market,
- 4:54Speaker 3: we could connect a new, very meaningful amount of supply
- 4:58Speaker 3: to this global global compute global demand, compute demand ecosystem.
- 5:04Speaker 3: Uh So that's effectively what we've built, right, We built
- 5:07Speaker 3: this enterprise focused GPU compute marketplace where data centers, companies, enterprises,
- 5:17Speaker 3: you know, even crypto mining organizations that have a large
- 5:21Speaker 3: amount of LATE and GPUs they can connect their gpuspon
- 5:25Speaker 3: network surface that surface them within our infrastructure, earn rewards
- 5:29Speaker 3: for doing work within the ecosystem, and provide those supply
- 5:33Speaker 3: uh to to the global compute demand market. And obviously
- 5:38Speaker 3: that demand market can then access compute within our infrastructure
- 5:41Speaker 3: and and h and access in the way that keeper
- 5:45Speaker 3: and also more available than our big solutions.
- 5:49Speaker 2: Mm hmm, No, it's a great overview.
- 5:51Speaker 3: Mark.
- 5:52Speaker 1: I appreciate that explanation, and it definitely seems like there's
- 5:56Speaker 1: a dis equilibrium in the demand versus supply that needs
- 5:59Speaker 1: to be solved with the amount of underutilization from all
- 6:04Speaker 1: these enterprises, and it's just not clicking properly. I'm curious
- 6:08Speaker 1: on you know, I know bitcoin miners and crypto minors
- 6:12Speaker 1: have a lot of GPU power as well, but what
- 6:14Speaker 1: about individual people or that want to contribute to decentralized network.
- 6:20Speaker 1: Maybe they don't have as much significant GPU power as others,
- 6:24Speaker 1: but is anybody able to contribute to GPU power and
- 6:29Speaker 1: then earn something through decentralized market.
- 6:34Speaker 3: Yeah, that's a that's a great question, and it's it
- 6:39Speaker 3: leads to to kind of a big differentiator that ether
- 6:42Speaker 3: has in the market. Now when you look at deepens
- 6:46Speaker 3: the decentralized physical infrastructure network, which is kind of the
- 6:50Speaker 3: broader classification for for projects to build infrastructure that's contributed
- 6:55Speaker 3: by others. Right, the big coin is effectively a deepen
- 6:59Speaker 3: people company to their compute to the deep and algorithm helium.
- 7:03Speaker 3: People buying these boxes contributing it to the helium network,
- 7:06Speaker 3: all the deepens within the deep in sector. That's compute deepens, right,
- 7:11Speaker 3: and this is where ether lives. It's where we have multiple,
- 7:15Speaker 3: multiple different compute providers contributing their compute to a broader
- 7:19Speaker 3: network objective. Now, when you look at how these deepened
- 7:24Speaker 3: projects have generally been built, right, what the traditional model
- 7:29Speaker 3: is is effectively taking your tokens and incentivizing consumers to
- 7:36Speaker 3: contribute their DPUs, right, contribute their gaming PC, DPU, constitute
- 7:41Speaker 3: their idle you know, MacBook GPU when they're when they're
- 7:45Speaker 3: not using it, because it's a huge amount of like
- 7:47Speaker 3: latent compute sitting it that sitting out there in people's homes.
- 7:52Speaker 3: And traditionally compute deepens have focused on this form of
- 7:57Speaker 3: latent compute. Right, because you can take your token and
- 8:00Speaker 3: you can pretty easily incentivize users to join your network.
- 8:05Speaker 3: Right because that compute that GPU. It's essentially a stunt
- 8:08Speaker 3: cost to them, and anything that they can can kind
- 8:12Speaker 3: of do to earn from that compute its beneficial to them.
- 8:18Speaker 3: The challenge here, however, right is, even though you can
- 8:22Speaker 3: build a network of GPUs really really quickly, right, and
- 8:27Speaker 3: even though you can build a really strong community because
- 8:29Speaker 3: you have a bunch of people that are very interested
- 8:32Speaker 3: in your token doing well because they're providing the resources,
- 8:36Speaker 3: they're minding that network, it's a really difficult, if not impossible,
- 8:44Speaker 3: compute network to sell to enterprise customers. Right, Because let's
- 8:49Speaker 3: say you take a bunch of gaming PCs, you know,
- 8:52Speaker 3: one thousand, five thousand gaming PCs, and you take that
- 8:55Speaker 3: to an enterprise customer. The first thing they're going to
- 8:58Speaker 3: do is off you, well, action, how can you gu
- 9:01Speaker 3: guarantee to me that those GPUs aren't just going to
- 9:05Speaker 3: be switched off at night? Right? How do you know
- 9:08Speaker 3: that the game is not gonna, you know, just switches
- 9:10Speaker 3: PC off when you go to sleep, Or how can
- 9:12Speaker 3: you guarantee me that the bandwidth of that user's home
- 9:17Speaker 3: is not just going to be throttled because they have
- 9:19Speaker 3: a zoom fall or because they're watching Netflix right. You
- 9:26Speaker 3: really struggle to meet the service UH and quality and
- 9:30Speaker 3: performance and uptime requirements of enterprise compute buyers when you're
- 9:35Speaker 3: building a compute network on the back of consumer equipment. Right,
- 9:41Speaker 3: So I guess your your your question was, can consumers
- 9:45Speaker 3: contribute to this type of network? Yes? They can traditionally
- 9:50Speaker 3: and they contribute to ATHOS network. No, they can't not today,
- 9:53Speaker 3: and the reason is because Ether has built a one
- 9:58Speaker 3: hund enterprise focused enterprise grade deepens right. And that's a
- 10:04Speaker 3: massive differentiator for us. Right. We have two forty three
- 10:09Speaker 3: thousand DPUs within our network. One hundred percent of them
- 10:12Speaker 3: are enterprise grade sitting inside a data center or connected
- 10:17Speaker 3: to enterprise grade network infrastructure. We have a global compute
- 10:23Speaker 3: network that's perfectly suited to meet the requirements of the
- 10:27Speaker 3: biggest enterprise customers in the world. Right. And that's something
- 10:30Speaker 3: that no other defend can say. It's put us in
- 10:34Speaker 3: a situation where we can onboard the biggest AI companies
- 10:37Speaker 3: in the world. We have the largest LM training contracting
- 10:43Speaker 3: on the back of the deep in Right. This is
- 10:46Speaker 3: fed mL or as a rebranded recently to to tend
- 10:49Speaker 3: to Opera. They're training their own open source model on
- 10:53Speaker 3: the back of a thres eight one hundred. We have
- 10:56Speaker 3: the biggest seller communication company in the world galing out
- 10:59Speaker 3: our structure. And you couldn't do any of that if
- 11:03Speaker 3: you had a network built on the back of consumer GPUs. Uh. So,
- 11:08Speaker 3: to be honest, it was a harder thing to do, right,
- 11:11Speaker 3: I just spent the first half of this uh yea
- 11:15Speaker 3: telling you have to build it was the GPUs and
- 11:18Speaker 3: it is. But if you can aggregate them, right, if
- 11:22Speaker 3: you can create a meaningful amount of supply and convince
- 11:27Speaker 3: and show to your customers that that this apply is
- 11:29Speaker 3: available at the highest level of quality and performance, uh,
- 11:34Speaker 3: you have a You have a massive advantage because obviously
- 11:37Speaker 3: the unit economics of compute within our network are very
- 11:39Speaker 3: very different to a network where you have to purchase
- 11:43Speaker 3: every GPU yourself. You have to own those servers yourself. Right,
- 11:47Speaker 3: So we can we can out compute compute traditional compute
- 11:51Speaker 3: providers uh in almost every way because where we're built
- 11:55Speaker 3: in this in this c centralized way. Now that being said,
- 11:59Speaker 3: and sorry for rambling, but that being said, do we
- 12:03Speaker 3: intend to open the door for consumer compute in the future?
- 12:08Speaker 3: First of all, yes, right, will it be within the
- 12:11Speaker 3: next maybe six months, we'll see if we do open
- 12:16Speaker 3: that door, it has to be for meaningful product that
- 12:22Speaker 3: we can sell, right. Ather is very much about real business.
- 12:25Speaker 3: That's why we have, you know, probably the strongest revenues
- 12:29Speaker 3: in definitely in deep end, if not in the broader
- 12:32Speaker 3: sort of crypto community. So we're very, very focused on
- 12:36Speaker 3: making sure that the apply and the compute within the
- 12:38Speaker 3: network has utility and that we're able to sell it
- 12:41Speaker 3: and we actually have customers for it. So we do
- 12:45Speaker 3: intend to bring consumer compute into our network in the future.
- 12:49Speaker 3: What form that'll be, I can't tell you today. However,
- 12:53Speaker 3: I can just and I know we'll touch on this
- 12:55Speaker 3: in a second, but I can mention that we specifically
- 13:00Speaker 3: did a node sale, right because the specifications for running
- 13:07Speaker 3: a node were such that anyone with a computer, anyone
- 13:11Speaker 3: with a consumer grade device could run a checker node,
- 13:15Speaker 3: and that checker node is functioning doing work within our ecosystem.
- 13:19Speaker 3: So for us, we were like, do we downgrade to
- 13:23Speaker 3: the requirements of our enterprise compute so that we can
- 13:27Speaker 3: let retal in or do we give them this give
- 13:31Speaker 3: them access to a different role within ecosystem, a very
- 13:35Speaker 3: very important role. Right The checker and let them in there.
- 13:39Speaker 3: And we chose to keep the standards for our GP
- 13:42Speaker 3: compute extremely high, but open the door for the checker
- 13:46Speaker 3: role and decentralize that to kind of onboard consumer into
- 13:51Speaker 3: our network.
- 13:53Speaker 1: It's a lot of very great points, Mark, and I
- 13:56Speaker 1: agree with you that you know this is business and
- 13:59Speaker 1: you want to go for the highest quality and make
- 14:01Speaker 1: sure that you have ninety nine point nine to nine
- 14:04Speaker 1: percent up time. And it's great to have a decentralized
- 14:07Speaker 1: network where anybody can contribute. But who knows, you know
- 14:10Speaker 1: what old graphics card Jimmy is using as a hobby
- 14:13Speaker 1: trying to connect and if you're trying to train the
- 14:16Speaker 1: next chat GPT five on that, it might just not
- 14:19Speaker 1: cut it. So I see that vision and it's it's
- 14:24Speaker 1: nice to see that that's possible in the future. But
- 14:27Speaker 1: also the checker node, because I did see on the
- 14:29Speaker 1: site there there's a node sale.
- 14:31Speaker 2: With with deep in.
- 14:32Speaker 1: But that's how people can contribute and not just through
- 14:36Speaker 1: their their graphics card, but they can be a part
- 14:38Speaker 1: of the network.
- 14:39Speaker 2: So that's really good to know. Now. At the beginning, you.
- 14:42Speaker 1: Spoke about the huge gap between and there's a lot
- 14:45Speaker 1: of demand and the supply and and how you're filling
- 14:47Speaker 1: that what It's hard to quantify, but are you guys
- 14:54Speaker 1: making a significant gap being filled there? You know, how
- 14:59Speaker 1: much more demand needed with what you guys have, Are
- 15:02Speaker 1: you guys able to sort of supply what everyone needs?
- 15:09Speaker 3: Yeah, that's a that's that's so a great A great question.
- 15:16Speaker 3: Short answer, short answer, No, right now, we're not. We're
- 15:20Speaker 3: not making a meaningful dent. Right. The the the massive
- 15:25Speaker 3: amount of compute that is needed, right, the gap between
- 15:29Speaker 3: global compute demand and global compute supply, it's just so iman.
- 15:34Speaker 3: The really, unless you're unless you're an Amazon, right, unless
- 15:37Speaker 3: you're a Google, you're not. You're not making a dent, right,
- 15:43Speaker 3: I Mean, it's crazy to think about the scale of
- 15:49Speaker 3: what's being built in this space m HM to service
- 15:54Speaker 3: the compute sector. So for example, I you get try
- 16:00Speaker 3: and find the exact quote because I'm going to butcher it.
- 16:04Speaker 3: But Microsoft is deploying at the moment, or they're undertaking
- 16:10Speaker 3: the deployment of the biggest infrastructure project in the history
- 16:15Speaker 3: of modern man Right, They're deploying more capital to stand
- 16:20Speaker 3: up compute data centers than has ever been deployed on
- 16:24Speaker 3: any infrastructure project. Ever, right, it is massive. Like we're
- 16:29Speaker 3: sitting at the point now where the world is looking
- 16:33Speaker 3: down the barrel of so much compute demand that they're
- 16:39Speaker 3: estimating that the world can't produce enough power. Right. We
- 16:43Speaker 3: literally won't have enough power supply as the planet right
- 16:47Speaker 3: with current infrastructure to satisfy the compute demand just five
- 16:52Speaker 3: years from now. Right. So is if they're making a
- 16:56Speaker 3: meaningful dan there, not yet, right, But I can guarantee
- 17:01Speaker 3: you that nothing scales quicker in a compute ecosystem than
- 17:08Speaker 3: a decentralized compute network. Right, And as more and more
- 17:13Speaker 3: compute is coming online, right, and as Ather refines the
- 17:19Speaker 3: the onboarding mechanism, as we get better and more visible
- 17:22Speaker 3: to traditional compute providers, do I think that we have
- 17:25Speaker 3: the ability to make a very meaningful debt, right, Because
- 17:31Speaker 3: you have to remember, the companies that are buying compute
- 17:34Speaker 3: are not just companies like Amazon or AWS in this context,
- 17:40Speaker 3: and not just companies like AWS or Microsoft Azure or
- 17:44Speaker 3: or Google Cloud. Right. Those companies have the ability to
- 17:48Speaker 3: buy and on sell, right, But the average tech company doesn't. Right.
- 17:53Speaker 3: They buy and it sits in their own data sense
- 17:56Speaker 3: or assist in the data centers. The renting faith in
- 17:58Speaker 3: and if they stuff up, they're forecasting if they overbuy, right,
- 18:03Speaker 3: or if they delay production that expensive compute. And this
- 18:08Speaker 3: is the most expensive compute of all time, right, there
- 18:11Speaker 3: have never been shipped this expensive, Right, it's burning a
- 18:16Speaker 3: massive hole in their pocket. So you have this resource
- 18:19Speaker 3: that's incredibly rare, right that companies know they need to
- 18:23Speaker 3: compete in this market, and they also know that it's
- 18:28Speaker 3: probably going to be more difficult to get to get
- 18:30Speaker 3: this compute in six months, right, So if I have
- 18:33Speaker 3: the opportunity to buy now, I'd better do it. And
- 18:36Speaker 3: then they have the flip side of that, which is
- 18:38Speaker 3: if they stuff up how much they need or if
- 18:41Speaker 3: their timeline is a little bit off. It's so expensive
- 18:45Speaker 3: to hold this compute idol. And that's where Ather has
- 18:50Speaker 3: this really really incredible opportunity to de risk that compute
- 18:56Speaker 3: infrastructure for all of the companies out there that don't
- 19:00Speaker 3: necessarily have the ability or the BD teams or the
- 19:03Speaker 3: infrastructure to on sell their compute through the cloud.
- 19:09Speaker 2: That's a great point. I'm looking forward to it.
- 19:11Speaker 1: And you know, I've been following Web three cloud a
- 19:15Speaker 1: little bit over the years and it's growing, but it
- 19:18Speaker 1: seems like the majority of you know, cloud infrastructure, you know,
- 19:22Speaker 1: I don't know the exact percentage, but it seems like
- 19:24Speaker 1: Amazon or AWS you know, runs like eighty percent of
- 19:27Speaker 1: the Internet, and it's hard for Web three to make
- 19:32Speaker 1: a dent. You have to either compete on price or quality.
- 19:37Speaker 1: What do you think about that transition of hosting, processing power,
- 19:42Speaker 1: everything data into Web three? Do you think that Web
- 19:47Speaker 1: three will gain more market share than the traditional cloud?
- 19:51Speaker 1: And traditional cloud will go down or they'll both grow,
- 19:53Speaker 1: but you know, Web three will it be able to
- 19:55Speaker 1: keep up as much?
- 20:01Speaker 3: Yeah? I think I think anyone that's out there kind
- 20:09Speaker 3: of thinking that that Web three will will overtake kind
- 20:12Speaker 3: of Web two in any sort of major sector is
- 20:15Speaker 3: probably a little bit delusional today. Right. It's particularly when
- 20:23Speaker 3: we're talking about the core competency of the biggest companies
- 20:27Speaker 3: in the world. Right, So, is is Web three going
- 20:32Speaker 3: to to overtake? I would say probably not right, But
- 20:37Speaker 3: like I, like I kind of mentioned in the previous question,
- 20:39Speaker 3: do I think that we can have a very meaningful
- 20:41Speaker 3: position within that, right, Because if you think about if
- 20:49Speaker 3: you think about the ecosystem, right, just comparing as the
- 20:52Speaker 3: two AWS for example, GPUs are the same. Right, we
- 20:58Speaker 3: both were selling the same compute. It's not like we
- 21:01Speaker 3: have special Web three GPUs and they have Web two
- 21:04Speaker 3: gv NO, same GPUs.
- 21:05Speaker 2: Right.
- 21:06Speaker 3: Infrastructure, right, physical infrastructure, same data centers. Right. So this
- 21:12Speaker 3: this equipment is sitting in the same places that that
- 21:14Speaker 3: Amazon's aws' equipment is sitting in. Right. We're linked up
- 21:18Speaker 3: to the Web through the same subs cables, through the
- 21:22Speaker 3: same network infrastructure. Right, It's all it's all the same, Right,
- 21:28Speaker 3: So what's different? Well, Web three has kind of two
- 21:32Speaker 3: really unique advantages, right, or two levers that they can
- 21:35Speaker 3: pull that I think are very useful levers. First of all,
- 21:40Speaker 3: we have the ability to onboard other people's compute, right.
- 21:44Speaker 3: I think that is a meaningful scalability advantage from a
- 21:50Speaker 3: cost and unit economics perspective, but also a speed of
- 21:53Speaker 3: reaching market perspective. Right. And then we have the token. Right,
- 21:59Speaker 3: the token, it's it's something that these companies don't have,
- 22:03Speaker 3: and the ability to take that token. Use it to
- 22:07Speaker 3: incentivize certain behaviors within your network, use it to to
- 22:12Speaker 3: subsidize use it to transmit currency across borders into regions
- 22:18Speaker 3: that you don't have a presence like the there are
- 22:20Speaker 3: just great like there are just some really strong network
- 22:23Speaker 3: advantages that that a web bree provider in the space
- 22:28Speaker 3: has that allow us to to compete in some specific
- 22:32Speaker 3: areas in really really interesting ways. Now, I think whenever
- 22:37Speaker 3: you look at a Web three company coming in and
- 22:40Speaker 3: trying to compete with with Web two, generally, generally we
- 22:45Speaker 3: will try and compete on products, right because we'll we'll
- 22:48Speaker 3: take the token and will subsidize and we'll we'll have
- 22:51Speaker 3: like a lower cost product in some sense. And it's
- 22:57Speaker 3: they're going to do that, yes, right, I think we'd
- 22:59Speaker 3: be still not to take advantage of the token in
- 23:02Speaker 3: ways it can affect the price that we're able to
- 23:04Speaker 3: provide to our customers. But I think the trap that
- 23:09Speaker 3: a lot of projects fall into within this space with
- 23:13Speaker 3: their token is using the token as just as effectively
- 23:20Speaker 3: only a discount and also using the token to heavily
- 23:24Speaker 3: subsidize the scalability of your network. Right. So, creating these
- 23:28Speaker 3: incentive programs that allow you to onboard a huge amount
- 23:32Speaker 3: of users really really quickly, right, but then put you
- 23:36Speaker 3: in this situation where you've attracted a lot of people
- 23:39Speaker 3: that have a short term interest in your network, right.
- 23:43Speaker 3: And then as an enterprise service provider, right, and you're
- 23:48Speaker 3: going to these customers saying, hey, I've got ten thousand GPUs,
- 23:51Speaker 3: they're perfect to you, right, And the compute provider says, okay,
- 23:54Speaker 3: I want a three year contract. Then you guarantee that
- 23:58Speaker 3: all of that compute's going to be here in three years.
- 24:00Speaker 3: If you've created a network that's really kind of volatile
- 24:04Speaker 3: because people need rewards to be there, then you've done
- 24:09Speaker 3: yourself with real disservice. Because actually the biggest thing you
- 24:12Speaker 3: need to be optimizing for as a web free company
- 24:16Speaker 3: is stability, because that's where we get picked apart by
- 24:20Speaker 3: every websuite company out there. Right, they rip us to
- 24:22Speaker 3: shreds on on our credibility, they rip us the shreds
- 24:25Speaker 3: on on the state of our ecosystem as a whole. Right.
- 24:29Speaker 3: So I fundamentally believe that the most important thing you
- 24:33Speaker 3: can do with your token is actually ensure that it's
- 24:36Speaker 3: incentivizing the stability of your network. Right. And if you
- 24:40Speaker 3: could do that first and then also use your token
- 24:43Speaker 3: to create great price incentives, right, to create great compute incentives, like,
- 24:47Speaker 3: all of those are really important pieces and they give
- 24:49Speaker 3: you a great competitive edge. But unless you're solving for
- 24:52Speaker 3: the stability piece, you're not going to be around long
- 24:55Speaker 3: enough for someone like Amazon to even really care about you. Yeah.
- 24:58Speaker 2: No, that's a great point.
- 25:00Speaker 1: It's traditional business is so much about how can we
- 25:06Speaker 1: make it ten years down the road so that this
- 25:09Speaker 1: can be a company that is here for decades and
- 25:12Speaker 1: it becomes one of the biggest companies in the world.
- 25:14Speaker 1: And on the crypto side, you're also sort of fighting
- 25:17Speaker 1: against in some sense bitcoin as a gatekeeper and the
- 25:22Speaker 1: market fluctuations of crypto being such a small industry that
- 25:27Speaker 1: although you have a great traditional service with the Web
- 25:31Speaker 1: three twist, you're sort of thrown into the wild waters
- 25:35Speaker 1: of bitcoin going up and down. And a lot of
- 25:37Speaker 1: people are there thinking shorter term, although we are starting
- 25:41Speaker 1: to see a lot more institutional investors, especially with the ETF,
- 25:44Speaker 1: coming in thinking long term with ten year plus investment horizons,
- 25:49Speaker 1: which I think is great as that trickles down into
- 25:53Speaker 1: Web three services and infrastructure. That long term investment horizon
- 25:58Speaker 1: is hopefully going to affect these kinds of companies more
- 26:02Speaker 1: where they're thinking long term. I think it's a great
- 26:04Speaker 1: point about stability and because it's not going to be
- 26:10Speaker 1: You know, it will scale very quickly with AI, as
- 26:13Speaker 1: you said, in the next year or two the compute
- 26:16Speaker 1: power will we be needed. But you need to be
- 26:19Speaker 1: thinking five ten years down the road, otherwise you're not
- 26:22Speaker 1: going to make it, because it's going to continue to
- 26:24Speaker 1: scale exponentially faster each year.
- 26:28Speaker 3: One hundred percent. Yeah, totally agree. And infrastructure. You know,
- 26:33Speaker 3: people in crypto love the term infrastructure, right, but infrastructure
- 26:38Speaker 3: is something to take time, right, Infrastructure takes time. And yeah,
- 26:47Speaker 3: we're definitely in for the long run, right. That's why
- 26:51Speaker 3: we chose to be enterprise first, even though we knew
- 26:54Speaker 3: it would take us longer to get the market. It's
- 26:56Speaker 3: why we have such a strong focus on real business
- 26:59Speaker 3: and on boarding real revenue and real customers into our
- 27:03Speaker 3: network because you know, five years is going to be
- 27:07Speaker 3: a blanket of an eye for us. Right, We're going
- 27:09Speaker 3: to be here for decades, and I think we'll be
- 27:13Speaker 3: a pot of building some really significant pieces of the
- 27:18Speaker 3: AI economy that we're all going to start to really
- 27:22Speaker 3: be seeing over the next kind of a couple of years.
- 27:26Speaker 2: I'd love to see it.
- 27:28Speaker 1: And now I was just looking at you know, the
- 27:30Speaker 1: launch just happened in the in the recent weeks here
- 27:34Speaker 1: and also the ath token.
- 27:37Speaker 2: I saw the note sales going on.
- 27:39Speaker 1: Is there any other way that people can contribute, say
- 27:42Speaker 1: they don't have an enterprise warehouse of graphics cards in
- 27:46Speaker 1: their backyard, how else can people sort of get involved
- 27:50Speaker 1: with this if this is the kind of thing they're interested.
- 27:54Speaker 3: For sure? Well, just to clarify for for any of
- 27:59Speaker 3: the the listeners, the nodes sale itself is now closed, right,
- 28:05Speaker 3: so the noales closed. That was a fantastic campaign. We
- 28:13Speaker 3: ended up selling just over seventy two thousand checkernodes, raising
- 28:19Speaker 3: around forty two thousand easts, right, so depending on east
- 28:24Speaker 3: price anywhere from one hundred and twenty five to one
- 28:26Speaker 3: hundred and fifty mil rightest biggest nodes sale in web
- 28:30Speaker 3: pree history was it was totally insane and really, to
- 28:34Speaker 3: be honest, just really a such a strong indicator of
- 28:42Speaker 3: the power of a community, right and the conviction that
- 28:46Speaker 3: that community has for what we're building. And I think
- 28:50Speaker 3: having the nodes sale and giving people the ability to
- 28:53Speaker 3: come in and run those nodes and do truly meaningful
- 28:56Speaker 3: work within our ecosystem. Effectively, the checkers are the police
- 29:02Speaker 3: force of the network. So it was this really clear
- 29:06Speaker 3: separation of police and state. Right we were effectively handing
- 29:10Speaker 3: over the keys of the police force to our network
- 29:13Speaker 3: through this decentralized checker ecosystem, and by running those checker nodes,
- 29:19Speaker 3: by conducting those checker pass within a network, they're keeping
- 29:23Speaker 3: the network safe, They're keeping the network free from bad access,
- 29:31Speaker 3: and that's a really powerful role and it deserves a
- 29:33Speaker 3: lot of rewards as well. So there's a significant amount
- 29:36Speaker 3: of at this token supply that's distributed every year to
- 29:40Speaker 3: all of those seventy two thousand note operators, So that
- 29:45Speaker 3: was super cool. Unfortunately, for those that missed out, we
- 29:49Speaker 3: closed that sale a couple of weeks ago. There is
- 29:55Speaker 3: a really cool staking product that's available. So if you're
- 29:58Speaker 3: really interested in committing and and earning rewards, but you
- 30:02Speaker 3: don't necessarily have any physical hardware to contribute, you're more
- 30:05Speaker 3: than welcome to go to stake dot athor dot com
- 30:09Speaker 3: and contribute your aph to the staking pools there, and
- 30:13Speaker 3: then you'll also earn reward as a result of that contribution.
- 30:17Speaker 3: I think for those people out there that are really
- 30:20Speaker 3: interested in the hardware piece and being a part of
- 30:25Speaker 3: the physical network infrastructure, we have a really cool product
- 30:31Speaker 3: called either Edge right, so you can check out that.
- 30:34Speaker 3: It's my edge dot io right, and it is a
- 30:39Speaker 3: physical mining box about this big. It looks like a short,
- 30:45Speaker 3: fat PlayStation five. It's super cool. It's a very very
- 30:51Speaker 3: premium device, right, you buy it. It's bosked incredibly well. Right.
- 30:57Speaker 3: It just feel awesome. Anyway you buy one of those,
- 31:02Speaker 3: You plug it in to the power in your home.
- 31:05Speaker 3: It connects to your network and it will passively operate
- 31:10Speaker 3: as a net as an enterprise grade network node within
- 31:14Speaker 3: as ecosystem. We actually partnered with Qualcomm, right, so it's
- 31:18Speaker 3: their latest AI chips in that service. Uh. And it
- 31:23Speaker 3: functions as an AI capable node within as enterprise node
- 31:28Speaker 3: networks right, so we will actually be pushing work out
- 31:32Speaker 3: to the eighth or ecosystem. It will be doing network
- 31:36Speaker 3: and receiving you know, really really you know, well thought
- 31:42Speaker 3: out rewards for those that are that are the owners
- 31:47Speaker 3: of the device. So super cool. You just you purchase it,
- 31:50Speaker 3: you plug it in, you forget about it. Right. It's passive,
- 31:54Speaker 3: but it takes advantage of the fact that we're putting
- 31:57Speaker 3: compute close to the user. Because big AI companies, big
- 32:03Speaker 3: tech companies, they care about what we call the edge, right,
- 32:07Speaker 3: that is compute that can happen closer and closer to users,
- 32:11Speaker 3: because the closer you can do some costs, the quicker
- 32:15Speaker 3: they can then distribute those costs to the user who's
- 32:18Speaker 3: requested it. Right, And it's traditionally been a very very
- 32:21Speaker 3: difficult part of the network infrastructure puzzle. Right if you
- 32:27Speaker 3: try and solve it with consumer compute, like you mentioned,
- 32:31Speaker 3: you don't know whether Jimmy has a different PC from
- 32:34Speaker 3: someone else, and you've got so much heterdineity and variety
- 32:38Speaker 3: on the edge that it's really difficult to like unify
- 32:41Speaker 3: and sell it, right. But what we've done with the
- 32:46Speaker 3: eighth edge is homogenize that layer, right, because if we
- 32:50Speaker 3: sell the box, the box is the same everywhere, and
- 32:55Speaker 3: then we've got this unified edge network that we can
- 32:58Speaker 3: start delivering really cool enterprise workflow too, and people can
- 33:05Speaker 3: earn reward from owning it.
- 33:07Speaker 2: Wow, very cool. I'm going to check that out for sure.
- 33:11Speaker 1: I can leave a link in the show notes to
- 33:12Speaker 1: the to the edge and also to the staking platform
- 33:15Speaker 1: for those who are interested. It sounds like you guys
- 33:18Speaker 1: are on the cusp of something huge, because I know
- 33:22Speaker 1: the processing power is going to be ridiculous in what
- 33:25Speaker 1: is needed in AI, and I don't think I can
- 33:29Speaker 1: even fathom how good it's about to get. Even just
- 33:32Speaker 1: the incremental versions of the l ll M models over
- 33:36Speaker 1: one year is crazy to see the expansion, and I
- 33:41Speaker 1: think the next few years are going to be mind blowing.
- 33:44Speaker 1: So I'm wishing all the best for you and your
- 33:47Speaker 1: team on delivering the processing power to the right areas
- 33:51Speaker 1: to make all that happen.
- 33:55Speaker 3: Yeah, it's not going to be easy, but this is
- 33:56Speaker 3: what would grade that right. So whenever we've got a
- 34:01Speaker 3: really strong community, we have a fantastic infrastructure already out there.
- 34:07Speaker 3: We're in you know, twenty five different countries, and that
- 34:10Speaker 3: number goes up, you know, almost every couple of days.
- 34:14Speaker 3: So we're extremely well positioned and I'm very very optimestic amazing.
- 34:20Speaker 2: Thank you so much, Mark.
- 34:21Speaker 1: I would love to have you guys back on in
- 34:22Speaker 1: the near future and wishing you all the best until then.
- 34:28Speaker 3: Thanks very much, and thanks for having me. It's been
- 34:29Speaker 3: a pleasure.