Projects
![]()
No projects found
Investors
New
DAO Treasuries
DATs
Lend / Borrow APRs
Chains
Categories
Spaces
DePIN on Arbitrum (Open Spaces)
Recorded: May 8, 2025 Duration: 0:26:58Space Recording
Short Summary
Huddle01 and WeatherXM are leading the charge in decentralized communication and insurance solutions, respectively, with significant user growth and innovative partnerships. Huddle01's robust network has facilitated over 150 million minutes of meetings, while WeatherXM aims to empower farmers with on-chain insurance, showcasing the potential of DeFi in real-world applications.
Full Transcription
Music Thank you. sorry space has rugged us once again um these things happen
these things happen
we'll give Elon a call
right after this don't worry
alright
can we get Ayush
back up here I think that's where we left off
All right, Ayush, my friend, tell us why we should be using huddle spaces instead of
x spaces outside of the obvious reason that we just crashed out.
I think it's a perfect segue for my slot.
So yeah, essentially...
Almost too perfect.
Yeah, almost too perfect.
Yeah.
Yeah, I think a couple of things.
I think, first of all, in general, Xspaces uses something called as WebRTC protocol.
WebRTC is Web Real-Time Communication Protocol. This is something
on which Google Meet, Zoom, Xspaces, Clubhouse, Huddle 01, everything is built on top of it.
Now, the reason why we see a lot of times why Xspaces rug us, we get rugged is because
the WebRTC connection breaks. And this breaks because of multiple reasons. Sometimes it breaks because there is too much load.
The load balancing has not been done.
Sometimes it breaks because there has been too much dependence on hyperscalers
like AWS, GCP, Azure, and others.
And sometimes it breaks because Chase is in the U.S., I'm in India.
Maybe Alex is in some other part of the world and same with other people and the more
people are in different parts of the world the more difficulty essentially it is for the audio
packets to route cross-continently and the more distance it is with each other and the and the
servers are generally in one part of the world because of which there is too much load on the hyperscaler and because of
which x rugs us. So at Huddle we try to solve this by making sure that you don't depend on these
hyperscalers that means AWS, GCP, Azure and these bigger data centers. So we build a people-powered
network majorly for bandwidth.
So anything which requires huge amount of bandwidth,
that means low latency and high bandwidth,
could be powered by our network.
To do that, since I and my co-founder, Sushmit,
we had been building a WebRTC before even Huddle,
so we realized we were essentially good at knowing what problems are there
so we decided to build the bandwidth network especially for real-time communication which
we are doing here calls like real-time communication requires insane amount of bandwidth the call which
we are doing here we have already uploaded more than one gbps of data onto aws right now and the
more data you upload to aws the more you pay to these people to Jeff Bezos
and everyone who's running this you can solve this essentially by building a peer-to-peer network
that means if I use Chase, Karthik and Alex are on call and if Karthik is a good internet connection
he can become a node and he can power the call of all of us that essentially solves for a better geographical distribution of the network that solves
for lesser costing of the network and that essentially makes sure that there
is a better load balancing and things don't completely break and that's why
we are a deep in network for bandwidth And we realized that we need to build something
like X spaces ourselves, because that's how you technically
get demand on your network.
So rather than completely focusing on supply from day one,
we focused on building demand from day one.
And the more demand we started getting,
that's how we started scaling our nodes.
So right now, we have built our own app called Huddle01 Meet,
which has X spacesSpaces kind of
feature which had video conferencing kind of features and streaming kind of features
and what that does is that that that has done is it has clogged more than 150 million minutes
of meetings so far which has made sure that now we have more than 40 percent of our network that
means our nodes which essentially have been deployed by ourselves,
by node ops, by a couple of our friends,
and by people themselves as well,
leads to a good amount of utilization
of the network itself,
which basically makes a pretty good segue
to what I think USDI is doing,
which is pretty important for something like us
because our demand is surging
and we require new supply of nodes to be spun up and that's
where usdi can essentially help with speed composability and basically can bring crypto
capital markets to this real world bandwidth kind of deployment so yeah i think long answer
but this is what essentially we do and yeah that's about it thank you very much ay Ayush. I was on a space with, I was on a puddle space, which was broadcast to X months ago, and there were 1,100 people in attendance with probably eight or nine on the speaker panel, and it was flawless. You could hear everyone fine, nothing broke, nothing crashed.
And we just crashed with what, 70 people on the space,
maybe pushing a hundred.
So the network truly works.
And my understanding is it works by utilizing latent compute
my understanding is it works by utilizing latent compute or latent bandwidth in the nodes nearest to the relevant locations of the speakers themselves.
Is that accurate, Aish?
Yeah, absolutely.
So any kind of communication which you do is very bandwidth intensive and not compute intensive.
So the node sales also, which we did, had a pretty high specifications on the bandwidth side.
That means if you have a 250 Mbps of internet connection, that means uplink and downlink,
and you have, then you are eligible to essentially run these nodes themselves.
And that's where essentially even our friends from NodeOps really helped us
because they had pretty awesome, node deployment mechanisms and also pretty good level
of nodes in different parts of the world similarly our other partners similarly people who had their
own essentially established deployment centers at their own homes. So essentially it uses your latent bandwidth.
And if you have an unused bandwidth,
you can use that to basically give to the network.
And the network's orchestrator,
what it does is that it sees the geography
and it sees the bit rate.
And based on that, essentially it routes the call
to the closest, nearest possible node operator which is there so
for example for caller like this it will make sure that it's in between different countries because
we are all in different parts of the world and it performs way better latency than x spaces or any
other because it doesn't rely on one server generally x spaces is in a server based in north
virginia because that's where the cheapest servers are because that's where the largest server farms are. This model completely
flips on its head in a bandwidth-based network because now you can utilize unused bandwidth
at different locations because a lot of people now have latent bandwidth which is high uplink
and downlink. And interestingly, this network could be used not only for just real-time communication but anything which requires
low latency and high bandwidth which essentially means that if you want to do
a CDN kind of job which means if you want to make essentially hit do any kind
of job which requires streaming require any kind of job which requires gaming
also essentially hitting sequencer faster essence more towards even high
frequency trading.
Whatever requires a low latency and high bandwidth could now be done via this network at a fraction of a cost.
Just to give an example, right now hyperscalers charge, basically pay to these AWS and GCP,
they pay $1 to every internet service provider to get the internet and it charges zoom $80 so essentially
charges an ATX markup to zoom. Zoom technically doesn't make the money when it charges you
$80 it has $20 it makes major money goes to Jack Bezos which essentially could be stopped if zoom
comes and uses the network of Huddle 01 and And that's what we are building. And that's where something like what USDA is doing
is pretty interesting because this can really lead
to people just spinning the supply set really fast,
seeing the humongous amount of demand
which is there in the bandwidth side of the network.
That's really well articulated.
I can totally see why people would switch to this.
And the Huddle team has been heads down building this product since 2022, maybe even 2021.
Yeah, it's been four years now.
I've been at it for a lot of years and finally we're seeing some fruit to it because we are
starting to see network effects happening on the app side of our network.
That means more demand on the supply side of the network.
So people in, for example, Nigeria, people in APAC region where there's not good connection,
people in Russia where the connection is not that great because wherever the servers are not there people really have a tough time uh to uh to essentially
communicate coordinate and to collaborate which changes uh with uh with the network like this and
with apps like this and the best way to essentially feel the the impact of a network is to build the
app ourselves that's what we believe and that's where we just bootstrapped and dog fooded our own app on top of it.
And now I think after three and a half, four years,
we are seeing a good result of almost 10 million minutes of meeting every day.
That's incredible numbers and insanely good pace by you guys.
incredible numbers and insanely good pace by you guys.
And I think the nose sale was a great way to bootstrap the network
and make it far, far more robust and handle more minutes at a time,
more consecutive meetings without inhibiting the quality of said meetings.
Yeah, I think we've gotten through each individual team now. Wanted to open it up for a bit of a
broader conversation. What would you all say are the biggest challenges facing D-Pen, InfraFi, WeatherFi today?
And how do you hope to see those evolve and be fixed over the next,
you know, let's call it one to three years?
Yeah, I can share just some thoughts for us
because we've obviously been spending so much time on building out the protocol design.
And, you know, we, Metastreet is the original brand.
We rebranded to USDAI, but we spent so much time on, you know, kind of illiquid asset financing. And, you know, with the stuff that's represented on chain, a big challenge is what's represented on chain and what that actually equates to in the real world.
And so, you know, I think if you say, hey, you know, you can get kind of mid-teens yields on
stable coins and there's no risk, that's great. A lot of people would want that and that should
be a pretty easy product to sell. Of course, there's risk to what the actual backing is of
those loans. And so what we spent a lot of time on with USDAI was figuring out how do you get something natively on chain, be the primary source of
trust and reputation as opposed to off-chain legal processes. Like if you think about,
you know, Goldfinch and some of the RWA protocols of like the 2021 era,
what happened when there were defaults was ultimately, you know, it was very challenging to
unwind those positions and people lost money. And so when we were designing USDAI, we came up with
this sort of framework for tokenizing these GPUs or solar panels or weather stations or whatever it is. And then we call it caliber, but it basically
is just a way to have the representation of the asset be fully on chain and have full legal rights
to the asset. So when we do these loans with GPU operators, for example, data centers, whatever,
they actually sell the GPUs to us. And then we
tokenize the assets and we get a master insurance policy that covers the value of those assets.
And so when you're looking at the actual collateral to these loans on chain, which is just
this NFT, that NFT is actually called an electronic document of title, but it's basically
like a deed to that asset.
And by doing that, that means that if there's a default and you have this master insurance
policy, even if somebody goes to a data center and like takes a hammer to every machine,
or there's a tornado that takes out a data center, or someone goes in and steals them
all and tries to sell them on eBay, you have this trust because you have this master insurance policy that's kind of backing everything. And so that process
was really challenging for us to figure out. And I think fundamentally, for deep end networks to
scale, you really want to be able to plug all of this hardware value into DeFi networks. And so
really, like the best way to do that is have sort of an
ironclad setup where when you tokenize something and represent it on chain, it's very, very,
you're very confident that that asset is actually represented on chain. And we've seen this now
progress across a variety of other projects like 375 AI is doing this and a few others.
And really, I think like the outcome here is you keep the decentralized design of the
Deepin network in place, but the decentralization can happen at the economic level.
And then you keep the operator responsibility maybe more centralized.
operator responsibility may be more centralized. So, you know, maybe huddle is better off running
all of the equipment itself. And then you sell the rights to that equipment on chain,
but the actual operations are maybe more centralized. And so I think this is like a
good way to scale because again, you can access these debt markets then because you have confidence
in what is like being represented on chain. And I think these debt markets then because you have confidence
in what is like being represented on chain and i think the debt markets are like a big missing piece
for scaling these networks generally so um that was something that i feel like has been a big
challenge for us to figure out and i think maybe mirrors some of the challenges that deep end
networks have had over maybe since 2020 era um we think of it we call it like a deep end networks have had over maybe since 2020 era.
We think of it, we call it like a deep end trilemma,
which is basically price, reach and scale.
And you kind of give up one or the other
if you go like lower cost hardware
to try to get the distribution in more hands
than you have worse data fidelity and so on.
So that's just something we've been thinking a lot about.
I think maybe rings true for some of the other folks on here as well.
I think that makes, I think that that's a, you know,
very clearly educated take.
educated take you guys have spent a lot of time on this problem um i really like the
You guys have spent a lot of time on this problem.
comparison to you know huddle running its own infrastructure while
the community has rights to the cash flows of of the assets themselves and maybe if it reaches a
point where okay huddle has a ton of infrastructure to run and they're trying to build out a chain and new products and, you know, maintain their community and ecosystem, et cetera,
et cetera. Maybe they delegate some of their hardware to an operator like NodeOps or, you know,
people step in, um, in the same way that Gauntlet, RE7, and Steakhouse
have stepped in as vault providers
in different DeFi protocols,
maybe there's a similar comparison
for people who are operating compute hardware
and different node operating systems.
I see a pretty big opportunity there potentially. I know the USDA AI team is working
with Dacouse and RE7 on the credit side of things, but would be really interested to
see what that looks like sort of in the long run. Whether XM or NodeOps, I don't know if
you guys have thoughts on any of that.
Do you want me to go or do you want to go, sir, from Weather?
You can go ahead.
Yeah, definitely. I mean, to further broaden the scope, right,
like from my perspective,
and I think Node-Op's perspective as well,
just to steer the conversation a little bit,
but the opportunities there, I think, in terms of the compute market in general, right?
We had the cyber report come out recently, and it was written by Jake, very well written.
And an interesting take he had was that he,
I think when you're so focused on the intricacies
and the ins and outs of your protocol,
sometimes you miss out.
I wouldn't say you miss out,
but you're not able to look at things from a bird's eye perspective.
And the market is huge, right?
Like I think it was quantified at 750 billion market.
And even if decentralized infrastructure is able to let's say convince
some of the users like one percent that's like 7.5 billion right so the i'm terribly excited to
a be part of course the the note of steam and i think all the folks here and all the core
contributors from this stellar protocols it's it's to me it's the hardest vertical right and that's
what we're all here for so i'd like to end on a positive note. So yeah, that's what I wanted to contribute with.
I don't know whether I'd like to add to that.
Yeah, yeah.
Can you refresh me on the other questions, Chase?
Yeah, it was related to...
Yeah, it was related to, well, the original question was,
Well, the original question was...
what problems are you guys looking forward to solving in the next one to three years,
whether D-Pen broadly or more WeatherXM specific?
And that got us into a rabbit hole around debt financing for compute, et cetera.
But let's make it WeatherXM or DPN specific here.
Okay.
No, it's all good.
I think I missed the beginning because I think I saw some comment on pricing.
I looked at the prices and everything's up.
So that was great to see.
But no.
So what are we excited about?
The thing I'm most excited about and where I've been spending a lot of time is bridging the gap between D-PIN and DeFi.
The sense of, you know, in many sense, WeatherXM is, we're a deeply scientific project.
In the sense that we work with meteorologists, climatologists.
And so they're trying to solve some of the biggest, world's biggest problems in the sense of, you know, emergency management, keeping population safe,
solve some of the world's biggest problems in the sense of emergency management, keeping
population safe, making sure that their energy is continuing to go, or when it goes down,
fixing it as soon as possible.
And so, from my opinion, we're trying to partner with a lot of government and large organizations
to get them better data in a more affordable way, as well as helping underserved areas
get access to high quality data that they
might not otherwise be able to access. Because, you know, in a lot of senses, people think of
like weather data, well, that's free, right? And it's free if you're, you know, kind of from a
wealthy nation, right? And even then, if you're in a remote part, you won't have access to that.
And so WeatherXM is essentially enabling people to get
access to high quality weather in a very affordable way. But so the thing I'm most excited about,
though, is in that D-Pen and DeFi partnership, is the prediction market and on-chain weather
derogative side. Because if we can get that going, the end goal of WeatherXM has and
probably will always be to be able to sell insurance to a farmer in Columbia who would
normally never be able to get insurance for his crop, right? The idea is he stands up a weather
station and then can buy a small insurance contract through WeatherXM. And then WeatherXM
can take that insurance contract and sell it into the weather derivatives market on chain.
So that's the end goal.
And that's where we're trying to get.
And we have a couple of pilots to do that,
but that's why I'm really pushing them
and doing so many demos related to our weather data
and how it can be used in on chain DeFi.
be used in on-chain DeFi.
I think that is a wonderful, wonderful note to end on.
Being able to provide data to farmers that otherwise wouldn't...
Being able to provide data that allows farmers that otherwise wouldn't be able to get insurance
is proof that gambling is not the only use case.
I don't know.
Thanks for coming, everyone.
This made me super, super bullish on the future of this space.
I think D-Pen is somewhat of an unsung hero when it comes to real world use cases in the crypto space um
definitely you know want to give you guys your flowers every now and again um yeah where can um
where can everyone learn more about your guys's projects Oh, usd.ai.
It's a domain name.
Yeah.
Sorry.
Go ahead, Alex.
Yeah, weatherxm.com.
And our Twitter handle's right here.
You can follow us now.
And my Twitter handle is 0xATD.
Yeah.
I missed what I used to say.
But for our end, it's NodeStar Network.
And for everything else, as you can see, our Twitter handle here here and we also have another one for all um nodops ecosystem updates which is build on nodops
so yeah if you have any questions or any kind of comments don't hesitate to reach out and once
again chase thanks and the off-chain labs team thanks for setting this up it's been stella
yeah you can find out about Huddle on two handles.
One is huddle01.com, which is the network,
and the second is Meet on Huddle01,
which is our Huddle01 Meet, which is our app.
So if essentially you want to use the app,
that's where you can essentially use it on that handle.
And if you want to run the nodes,
you can be a part of the network.
I'm at Ranjan3118.
That's my handle. And yeah, it was a great space. It was great to be a part of the network. I'm at Ranjan3118, that's my handle.
And yeah, it was a great space,
it's great to meet a lot of faces again,
and yeah, it's been awesome, thanks Chase.
Thanks folks, and I'll shout out Aether as well,
they're at AetherCloud on Twitter.
Thank you very much guys.
It was great chatting with you all all hope to do it again sometime
soon thanks a lot bye see you guys like
Host
Arbitrum
from
Arbitrum IOU
