DePIN: Hype or the Next Trillion-Dollar Market? - TOKEN2049 Singapore 2025

Good afternoon everyone. My name is Auggie. I'm an investment partner with CMT Digital. We are 28-year-old trading and investment firm based out of the US. We operate a series of venture funds and we're heavy investors in Deepen, including in Helium, Andrina, and and many others.

And it's my pleasure to be here to moderate this panel titled Deepin Hype or the Next Trillion Dollar Market. Uh maybe to start it would be great if each of you can introduce yourselves and your project and for those that are new to Deepen in the audience how does a user or contributor get started on your network and how long does it realistically take to see value?

Uh my name is Greg. Um in the realm of decentralized infrastructure I created Akash a uh truly uh revolutionary computing network. What sets Akash apart is its fundamental commitment to decentralization and uh Akash is fully open source entirely onchain non-custodial and genuinely permissionless uh users are not bound by traditional gatekeeping mechanisms like login or passwords uh stored in like centralized databases instead they maintain sovereign control through their own um cryptographic keys. uh the economic proposition of uh Akash is simple right uh it's uh you can purchase GPU uh CPU based compute for about 80% cheaper than what you would otherwise pay on the cloud and uh GPU resources are roughly 50 minutes 50% or or less uh on a cache and uh users is is uh in in many cases many documented cases publicly users can deploy on a cache within two minutes and some of them in under two minutes uh you can get a workload in AI workload from start to finish and deployed and ready to use in 2 minutes. So that's significantly uh less than what you would incur on something like AWS.

So um you know according to many people it's a significantly better user experience.

Hello everyone. My name is Frank Mong. I'm the chief operating officer of Helium. been with the company for eight years and I'm very happy to say the token is H&T. So look it up, bookmark it and do more research on H&T.

That is the token that drives everything that's in the Helium ecosystem. In the United States, in Mexico today, there is a massive network that's growing from zero to over 115,000 hotspots in the region, providing cellular connectivity and offloading. The technology, while it's Wi-Fi, is only usable in the US and Mexico today. I I wish I could sit here and say we can use it in Hong Kong, Singapore, Thailand, South Korea. That's coming.

But we're trying to play out the two-sided market in the US because of a lot of what I would consider barriers of entry that cellular and telecommunications company have set up. There's things called license spectrum that you have to buy which cost billions of dollars so that your cell phone can work everywhere. That's the barrier that we're trying to figure out how to break so that everyone in the world can have connectivity to the internet without having to worry about cellular, Wi-Fi or whatever. It should just be you have a device that you can use for compute and wherever you go in the world, it just works and you're connected to the internet at a fair cost. That's all we're trying to do.

And so that type of experiment that we're testing in the US market today is called Helium Mobile and it's a phone plan where we offer it for free. So people are signing up. We get about 2,000 new subscribers per day of signing up with our phone. You get three gigs of data, talk and text. No, there's nothing nothing else beyond that.

No credit card needed. Uh there's no fine print. It's just a straight simple phone plan with with data. As we do that, we're trying to learn and understand how we can add additional value. And the more we learn, the better we are equipped to work with other telos around the world.

So our objective is to really figure out how to bring value to customers, how to make sure that they can connect to the internet very affordably and very safely and cheaply, and beyond that, how we can continue to add value to that customer experience. So that's that's Helium in a nutshell. We're growing massively. We burn about $2 million in revenue a month now of of H&T equivalent tokens. And so please go do your research, check out H&T.

Uh let me know if you have any questions afterwards.

Uh my name is Mike Horton. I'm the project creator of Geonet. Uh Geonet is the decentralized location layer for robotics and we provide a hyper precise location signal. What does hyper precise location mean? So you can think of GPS as giving you kind of a rough idea of where you're at.

Sort of a couple meters in X, Y, and Z. We provide something that's a hundred times more accurate. So is pinpointing the location of a device to within roughly a cubic centimeter. And that is an incredibly high utility for things like drones and robotics. The network is powered by the community.

So anyone can purchase a geoet station and set it up on a rooftop. You want to look at the map and kind of understand where there are gaps in the network, where there open hex's because the network is already the world's largest network where 21,000 roughly nodes have been added to the network over the last uh almost three years now, making it the broadest most adopted network for hyper precise location in the world today.

Hi guys, I'm Gor. I'm CEO at.net. We are the biggest decentralized cloud platform in the world. uh we have the max amount of GPUs and the biggest revenue onchain trackable in the industry.

I've been from last 10 years in leadership engineering positions in booking.com and finance and in these 10 years of mine I've seen that AI was literally a superpower in the hands of somebody automating building and scaling but there was a fundamental problem the fundamental problem was on one side we are saying the future will be built on AI but on the other side the new builders were being set up for a failure right from day one. So when these big companies were going to AWS or Google or companies like these, they're giving 60 70% discounts on compute which is like 50% of the expenditures of these AI companies. But when these new builders were going to the same players, not even 10%. And the hardware which they need to compete in this market, they'll not even get the latest hardware.

So they the bigger players will become bigger, the smaller player were like not even able to compete. So we saw this problem and in my previous turn in booking.com I knew how to build this network aggregating these inventories like we did in hotels and creating a marketplace for consumers and we saw this opportunity in the crypto side using tokconomics which could be consumed by anyone web 2 or web 3 which is compute and software products on top of this compute for the next generation AI builders and that's exactly what we built in the last two years we built a platform Within a year we launched the platform. We're proud to say we are one of one of the biggest suppliers of GPUs by aggregating all the supply from different part of the world and then we just focus on the revenue in last 78 months and we could like aggregate more than $20 million AR revenue in no time. So yeah that's our journey.

Thank you for those introductions. Uh we obviously have a jam-packed agenda today. We'll dive deeper into topics around uh incentives interoperability but I think before that let's start with the level setting big picture question which is when does decentralized physical infrastructure networks i.e deepen when does it make sense and when does it not? uh and maybe for the audience's context, you know, when DeepPin broke into the scene uh or broke into the public sphere, it was really heralded as the the the next phase of the of of the shared or sharing economy, right?

So there was talk of Uber for X but with token types use cases that would transform uh every vertical. I think since that time uh at least for us here at CMT we have a slightly more sober view of the sector which is deepen works best with assets that are inherently fragmented and have a coordination problem to be solved through a token but that's just our view. I'm curious to hear from the panelist perspective in your view what categories of infrastructure are strictly better when decentralized and which are not how do you draw that line? So good. Okay.

So, it sounds like the the the hype is gone. The reality is are in. To be honest, people who are in this industry, the deep inside, they are more excited than ever to be honest. And I have a different opinion than than yours. I genuinely feel each and every asset could be decentralized and in long run it'll be for better.

But the thing is like we have to do things strategically, right? You think of the normal world also any startup when it starts it's always centralized there are few founders who have vision they have power they execute but when the same organization matures becomes bigger what happens

they create departments they give autonomy

execute in different departments right by default it decentralizes because with centralization when you scale bottleneck comes system becomes fragile we see it in the normal life very similar thing happens in almost everywhere right now. But there are some nuances. The nuances is not every technology matures at the same rate. But there are some industries where technology exist. The economics make sense.

The consumers are right there in front of you, right? And the builders are building the product. Like in our case, when we talk about compute and AI, there's a huge demand. The problem is there and technology also exists. So we ventured into this space.

But there are bunch of other assets like mobility and other places where if you focus on it right now the ROI may not be that high as builders. So whenever we build something we think of multiple aspects right that how much effort it'll take what the ROI would be and how it'll help the ecosystem and humanity as a whole right and we're seeing it already right like Bitcoin is literally doing to one of the most centralized things in the world which is money right so we we're seeing it we're living it right now so I think everything could be decentralized it's like strategically where as builders we focus our energy on so that the future of humanity is best in the coming days.

I would just add I think all digital infrastructure is better and more resilient when it's decentralized. But there may be an ordering problem that relates to how quickly a deepened network can generate demand. In our case, the problem that we've been solving has been a problem that's been around for 20 years and no one has ever been able to build a largecale high precision location network because it's inherently very decentralized. You have to have these nodes operating in all corners of the world. So our 20,000 nodes are probably the most geographically diverse and dispersed nodes in all of deepen.

And the coordination that the tokconomics have provided had allow us to bootstrap this in in a time where the traditional networks in web 2 have been built up over 20 years and a roughly 1/4th the scale of geoet and in two years we built something four times bigger. So I think it's a real testament that decentralization can really work and as a result we've been able to unlock demand very quickly and have scaled to about six million in ARR in just just under a year of of working on the revenue side as the network has scaled.

Yeah. Let I kind of address that question with first a question for the audience. Who who here has Wi-Fi at home at at your house wherever you live in the world? Some some of you do not have Wi-Fi. I I don't believe you.

Basically, everyone has some form of Wi-Fi at home and when you're home, you're connected to it. That and by the way, all of you h own a decentralized network. Your Wi-Fi at home is decentralized. The the question is, how do I convince you to share your Wi-Fi with robots, devices, sensors, other phones, if it's secure and it's safe for you, how do I how do I make you do that? Make you share?

And that's where crypto I think was the answer for Helium. H&T as a token allowed us to ask everyone in the world to share their Wi-Fi or internet at home so that we can create a massive sharing network where anything can connect anything. It started with internet of things little sensors like tracking your bike or tracking your car keys your dog your pets stuff like that. Now, now governments use that IoT network for smart city applications. So what we're the way we think about this is we created that massive 1 million node IoT network in 2021 in just two years.

We launched in 2019. By 2021, we were in 800 cities, 120 countries. That's that's like phenomenal. the largest decentralized network that ever existed. And so I think from my perspective when it comes to infrastructure like what what works and what doesn't work, what's the problem?

I think I think the idea is like things that naturally are decentralized are the best candidates similar to industrial, you know, applications around location sensing and so forth um that Mike was talking about. But in this case for consumers, you guys are just average people. you guys have your internet needs at home. If if I gave you a way to share that and you get ba basically paid in H&T, I think you would do it. And that's what we're doing.

Um, I think deepen shines when it mirrors its own design. Like countless small parts like uh trees in a forest or neurons in a brain or stars in a galaxy working together to form something resilient and uh greater than the sum of all its parts, right? uh in our world that means uh fragmented assets like GPUs or rooftop solar or or Wi-Fi routers or city sensors right uh these are abundant uh underutilized and hard to coordinate with uh with traditional institutions um so tokens act as the the protocol layer that reduces entropy and turns that chaos into order but not everything can be decentralized or want to be decentralized rather uh a nuclear a plant um a hydro dam right uh a submarine cable is indivisible like it's a monolith you can't really decentralize it adding decentralization there only adds only creates friction so the line is simple uh decentralize what is abundant and idle and centralize what is scarce and indivisible that's where uh deep end doesn't just work it unlocks new networks uh we couldn't build uh any other Okay, makes sense. Maybe a quick followup if I may for for both Greg and Gorov just to better understand the value proposition of decentralization for your category. And I guess the question is compute and inference have traditionally been centralized domains because centralization equals performance, right?

There there are centralizing forces at play. Moore's law on on GPUs, hyperscalers. How do we think about the benefits of decentralization in in your domain? Uh I can take this. So inference wasn't very centralized.

Inference is a lot more flexible when it comes to uh you know deployment. I would argue that it's better when it's at the edge. So getting access to edge resources uh uh with a common set of rules in a protocol is significantly better than having a centralized coordination layer. Right. Um when it comes to training, training has been traditionally uh or even now as very centralized because there are three challenges with training.

The first challenge is uh training requires I mean AI training requires heavy communications. So you know it requires lots of nodes talking to each other uh or rather sending uh data to each other lots of data with with high frequency and second big problem is training is not fall tolerant meaning if one node goes down it it stops the entire training run. So when you have such like fragile uh topology uh it's very hard to like have them distributed because distance means problems in generally speaking in computer computing. Um and third problem uh is it's heavily serializable heavily serialized. So it's not parallel.

You can't really do you know uh u training in a asynchronous or a parallel manner. It has to be one after another. So uh that means uh and that's another big problem is you require homogeneous uh uh GPUs meaning if you run the the speed of a training run is only as good as a slowest node in the in the run. So meaning if you have H100s which are traditionally uh you know which are advanced GPUs and you have A100s which are older GPUs that the batch is only as good as the oldest GPU or the slowest GPU. So everyone wants latest and greatest chips.

That means you need there's high demand for the chips. Every time Nvidia comes out with a new model, the older model demand goes down or rather because everybody wants the new models. So coordinating new models and coordinating purchases and and the utilization is fairly high for the for the newer models, right? That's why it's traditionally centralized. Now in the last one year or so and this creates another big problem.

One big problem right now is energy, right? The amount of energy we need is doubling every two years. Um and the problem is so pervasive now that Congress invited me to testify on energy and I did in in in May. Um the the amount of energy is doubling to a point where we'll need a nuclear reactor by 2030 to train the state-of-the-art AI. So we have to figure out how to distribute AI and train them and there's a lot of work that happened in the last few years uh on low communication algorithms to asynchronous algorithms to fall tolerance all coming together.

I believe in the next 6 to1 year you'll actually see a truly decentralized model or a model that's trained on truly decentralized networks. Uh we still haven't figured out the incentives how that looks like. uh primarily what would what's incentive for me to provide my compute for a model run right that that's still I think uh an area of opportunity but but the fact that the AI training will become decentralized because of the energy crisis we have uh and we cannot build new supply of energy faster in America anymore uh so to answer your question yeah it it's not decentralized yet but it will be very very very soon

makes sense

long answer but

from from my side Like when you say decentralization right the same term could be envisioned in a very different way by different builders also right uh for example when we build the our our decentralized cloud platforms and that's the first thing which people think as consumers when they come to our platform and we have to correct them is let's say I imagine my platform as a straight line which has two extremes so let's say you're an enterprise player and you want to consume a deep network but for your use case you need thousand GPUs in the same data center. We as a network provide that. We have aggregated all this supply from hundreds of cities, hundreds of data centers where if you have that use case as an enterprise that you need thousands of GPUs in one data center, we can do it.

If you're a web three player that you need hundreds of GPUs each one at a different data center, you can do it. We as a platform provide a technology the way you want to consume and we have aggregated this way. Right? the use case is very different. Now coming to the to the decentralized part like when we said earlier I think everything does get decentralized on its own.

For example the example of energy was given in all honesty if you take a look at the history energy was centralized in past slowly and steadily the energy generation in each country has been decentralized. Money never thought nobody thought that money will be decentralized ever. We're seeing it right. We're living it soon. Each and everything which you think is powerful is critical for the world to like live in and supported.

You will slowly and steadily see it getting decentralized. The ROI part is there. Some we are ready to have some we are not ready. I think that's a piece. So same decentralization concept could be could be envisioned in a very different way.

So I think that's why you see very different flavors of the platform like I think Akash the way they built it was very different. We did it very differently and we have a kind of a different business model also. But all of us are doing well. So like different user base.

No, thank you for that. And and so I think now that we understand broadly where deepen fits uh and where it doesn't. The next question uh is how and and whether these networks can sustain themselves economically. Maybe touching on you know the supply and demand issue that that that Mike alluded to earlier. I think each of your networks uses token incentives to bootstrap contributors and it's it's well understood that token incentives can help to jumpstart a network but sustaining usage and long-term economics is obviously more difficult.

Uh I'm curious to get the updates for each of your projects. What's the clearest proof you have today that shows your network can sustain long-term growth with real unit economics where supply and demand are naturally imbalanced? Maybe we start with Mike or Frank.

Sure. So, um, Geonet has a token GOD and that token has, uh, started with a total supply of a billion tokens and as the network is utilized, the number of tokens goes down through a permanent burn mechanism. So, when any customer subscribes to our network to get precise location services, they're paying a fee um, which is used to buy back and burn God. And that rate of buyback and burn now is about 70% and we expect it to exceed 100% um sometime in early 2026. So I think we've established that you do need to use the token initially, you're not going to have that kind of buyback and burn power because you're building out the infrastructure, but on the once the infrastructure is there, if you have a good business case for the network, which we certainly do, our network is better coverage factor of 10 cheaper um fully modernized.

So it's really outperforms any of its web two peers on all aspects. that's able to drive that demand, buying back and burning the token, decreasing the total supply of geode all the time.

Yeah, we we focused on building the network initially with with our token H&T. Um, as I mentioned earlier with basically everyone sharing their network, sharing their internet from home and what what we discovered was we didn't spend a whole lot of time on the usage side. And so in this network that we're building now that's all about what I call Wi-Fi offload. It's a very technical term basically means your phone just works everywhere. What doesn't matter where you are the Wi-Fi you can your phone connects automatically and you don't have to think about it.

Uh, so that means for you, you just have coverage wherever you go. Indoors, outdoors, US, Singapore, doesn't matter. That that concept for us really where we focused our efforts there was really on the users on people sitting here. You guys all have cell phones. You have a phone plan.

If you're in Singapore, maybe it's Singtel, right? If you're coming from the UK, maybe it's Vodafone. Or if you're coming from Germany, maybe it's Deutsche Telecom or or um or T-Mobile. And so my my point is we we really focused on what is that user trying to do? What do they need?

They need the internet somehow. They need to connect. They need data. So how do we get them the data they need the fastest possible? And that and that's really how we zero this in and not focus on crypto token economics or no crypto.

We started with some crypto with the phone plans. it actually didn't perform the way we thought it would perform in that we weren't getting the traction we wanted once we started to hide the crypto and we only talked about the benefits of our phone plan and how it worked that was using partly a sharing network it was also using partly our partner T-Mobile in the United States when we explained that that actually became more interesting to the average person not you sitting in this room you're here because you know crypto, you're interested, but think like your relatives, your brothers and sisters, your your mom and dad, and your aunt and uncles. They they think you're crazy for even looking at crypto, right? So though they're the ones we went after, th your your family members now are our customers at Helium Mobile because we give away a free phone plan. It's very simple.

The value proposition is you connect to the internet, you get a decent amount of data. It's not a lot. It's three gigabytes. You can still use your phone anywhere in the world, all right? But it's free.

That helped us grab a ton of people to come in and use the network. We were able to use that to get people to use our network. Once we got that, we introduced other P plans. There's a $15 plan. There's a $30 plan.

And people naturally will self- select and upgrade themselves. Bottom line is that's driven us to about 400,000 subscribers in the last 18 months, which is pretty good when you're talking about starting up a a phone plan, a phone service. And because we were able to demonstrate that, we got AT&T, T-Mobile, Google to look at the Helium network, and now they're offloading their traffic to our network. So, they're using it. And so the Helium network itself now sees about 1.

9 million users a day. So we have 1.9 million daily active users on the Helium network. So we've figured out the supply side, I would say, and that's growing rapidly. We're we're adding about 150,000 new subscribers every quarter to the Helium network.

So I think we've figured out the supply side of our equation. Now, what I got to do is figure out how do I get the Helium network here in Singapore, in Hong Kong, in Thailand, right? I really want to do that. And so, that's that's really how we think about this. But it's but we've been pretty methodical because we've built a a network once already.

Now, we got a second network here with Wi-Fi.

So, really figuring out usage and supply uh demand on that side has been really really our focus.

We exactly have the same journey like very very similar. I can relate to it. So when we actually created our network like we literally had one time like more than $20 billion worth of hardware on the network as supply and then we did multiple experiments. So supply side of things we were sorted in the first six months we knew that like we had did multiple experiment that if you pay only in cash or USDT or IO token or combination of all supply side is sorted on the consumer side we took couple of months and what we realized was we're going after enterprises and the reality today is if I accumulate all the web three AI consumption that in totality would be covered by couple of medium-siz startups on web two side that's the reality Right. And when we going after these web two players and the moment they come to your platform and they see crypto tokens, they take a step back.

So you have to like rebrand yourself. We have to like change the strategy also in the beginning that we started doing partnership with universities, antler programs in the early startups try to tap into the usage there giving them credits and when they started to use the platform honestly from the word of mouth itself we started to get the usage. So I think many times even if you have a product uh because we are coming from the crypto side you have to change your brand you have to change you have to take that token away from the discussion so that it's not about token it's about the product and if your product is there like things just get sorted out and that's exactly what we saw in the last 6 months first two months we had a challenge then we did a rebrand we cleaned everything up not talk about the token once we did that people tried product everything sorted

yeah so we took a little different approach we didn't bootstrap a cash with tokens uh because we believe there's too many dimensions in the cloud to come to a common commoditiz commoditized uh you know uh uh instrument to to incentivize and second big challenge with incentivizing uh such highdimensional uh you know asset is verification. So one might say verification you know there's quite a lot of networks that do buildtime verification but build time verification that means once you know you verify that you have the resources and then you know then you then you don't uh and so you can one might say it can be easily gained where you verify when you when you're building but you can take away those resources and you can give those resour right so that's why continuous nondisruptive verification is is key. That means without affecting the workload that's deployed on the on the GPU or on the on on the node, you got to continuously verify and no one has solved the problem and I think we're close to solving the problem and doing it in fully open source and decentralized manner, right? So um uh but we were able to achieve hyperutilization like on a because we didn't bootstrap using tokens. I mean, of course, the supply is not going to be similar to a bootstrap network.

Uh, but utilization is fairly high. So, our utilization right now uh is somewhere north of uh 70% uh utilization. For comparison, AWS has about 40% utilization, right? And most decentralized networks have super high supply and less utilization. I think Filecoin has 0.

5% utilization. Um so we took a lot of those lessons and we said look there's enormous supply less demand let's look let's let's try a reverse thing where we have more demand than less supply so that we can understand how to incentivize and now we're getting towards an incentive structure uh we call utilization guarantee incentives because we discovered most providers on the network they're happy if they have 80% of their utilization basically 80% of the time they're actually leasing the resources so we discovered we only need to incentivize that delta That means if H100s have typically on the network about 70% utilization, we only need to incentivize or guarantee the extra 10%. Right? So uh you know to answer your question yes you know tokens are rocket fuel like you know once they help you launch and bootstrap a network but once you're in orbit you have to run without the fuel right or like very less fuel. So um so we are in that orbit phase where we don't like we can literally use less fuel but somehow we got to orbit using hot air balloons instead of rocket fuel.

Makes sense. Great. Uh maybe second to last question before we wrap up and I know we're just coming up on time here. Uh I do want to spend just a minute talking about abstraction because I think one of the other objectives uh for the industry and I imagine that's top of mind for many of you uh you founders is you know how do we make on the supply side the onboarding process more seamless uh and then on the demand side perhaps how do we make it such that the user doesn't even know that they're using deepen somewhat analogous let's say to DeFi applications that uh abstract away the key management and the wallet setup process. Um I want to for the interest of time point this question at Mike just because Geonet I think has evolved in quite an interesting well way when it comes to abstraction from a more uh heavily involved hardware interaction to something that's much more seamless and so my question for you is how do you think about abstraction for your product and have you seen any measurable impact on churn or retention?

Yeah, I think that um traditionally setting up a what we call an RTK base station, which is the core of our network, has been something that really only large huge web two companies and governments try to do. A node would run $40 to $50,000 to set up and $10,000 per year of operating cost. We've simplified that down to a device you can buy for $700 and run on your rooftop. But there were a lot of lessons along the way and how to get those installations to be the same quality that a traditional geodetic node has done. And we've done a lot of that through UI improvements, through a lot of improvements in the firmware, understanding the huge diversity of networks that are out there across the world because we're deploying in essentially all the countries of the world except for uh a few countries with unusual geospatial regulations such as China and Russia and denied countries.

Um and I think making it very simple has absolutely improved the not only the retention but also just the performance the SLA of the network. And the next thing that we did that I think has helped a lot is we started to combine multiple projects into one node. So rooftop space is inherently a very valuable asset. So if you own a rooftop whether it's a home or a business and you have stable power and stable internet that's a valuable asset. And by putting multiple deepen projects together like we've done in a node that we have that both provides coverage for geoet but also provides coverage for wing bits which is aircraft tracking network.

We're able for a single person to get a a device up on the on the roof and and earn two different projects rewards. So I think that's another type of abstraction this multi-RO abstraction that's super useful.

Makes sense.

Great. Um so we're coming up on time here. uh very last question and I imagine the question that everyone in the audience has top of mind which is you know very clearly after this invigorating discussion deepen is not just hype but can it be the next trillion dollar market and I'll sort of wrap things up here with a quote from one of the leading deepin focused funds escape velocity according to data cited by escape velocity deepin's onchain revenues are still relatively small today at around $und00 million according to their estimation but the future remains bright as Escape Velocity wrote quote Deepins onchain revenues today are less than 1% of the overall crypto market but are not representative of the industry's promising long-term future potential. End quote. For our panelists, if Deepen is to become a trillion dollar market, what's the specific consumer change uh or technological unlock that will get us there and how close are we today?

Yeah, I think I'll I'll kick it off because we're going after a market that's been around since the beginning of the internet. A lot of you here don't remember what it's like before internet was accessible. I I I was around. I I know what it was like. It was a simpler time.

But that market, the telecommunications market is worth I don't know three to four trillion dollars globally. And I think if Helium can convince just 10 or 20% of them to work with us and use our network as well, then you can imagine how we can carve into that overall macro market size. But that's that's really what we're going after. And I think if Helium can succeed there, we will continue to offer services and value beyond just connectivity. Once once we're in your phone and once we are the ones providing you access to internet, I think you would trust us with it with other additional services that are highly valuable.

So we can go after adjacent markets that are worth trillions of dollars. But ultimately, I think one thing you guys should remember about Helium is as I as I kind of put my mic down, we we really do focus on consumer interface, consumer user experience. And so, think of us as more like Phantom wallet than the really difficult to use wallets out there. So, if you guys have you hold crypto and stuff, we're we're the ones that are super easy to use and that's that's our endgame.

So,

super easy to use.

Yeah. So I think in my view I think to become a trillion dollar industry the the consumer mindset and the builder mindset both have to change over a period of time and it is changing which is you can only become a trillion dollar industry when it's not about decentralization itself it's like the consumer is coming because the products is cheaper faster scalable and stable then only you can reach that scale per se right um there's a take I'll tell you like people will not like it honestly out here but you'll see In few years that'll be the reality right there. A lot of projects get launched in crypto. There are only handful number of projects which have a real revenue. Within that handful number of projects, majority are the ones where a particular chain becomes bigger.

Maybe a meme token getting whatever and the whole ecosystem earns a lot of revenue. But when the market goes down, the revenue goes down. You will see deepen as only handful number of players among which which you can count on hand will be those players who will earn revenue irrespective of crypto cycle market goes up market goes down does not impact the product the revenue will keep flowing in and as we said like we just built the platform within last two years and within 8 months we built like a 20 plus million business right and that's just exponentially increasing and I think so is the case with everyone out here so all of us have proven that the business model exists. We have done it. There are few places where what needs to be done is take a step back, let the builders build, be patient.

The market is there, the builders are doing the right thing. Just be patient and things will happen. And I think deepens in aggregate are at that place.

I think it's super super simple actually. AI is the market. It's really the mega market. And within AI, there's physical AI. And the only way that physical AI can come to real fruition is through deepen networks that are providing the infrastructure for that.

And physical AI is posed to 70% of