Luis Oscar Ramirez - Beyond Hype: Mawari’s Real-Time Immersive Network - TOKEN2049 Dubai 2025

Hello everyone. I'm Lu Ramirez uh founder and co of Mari. So what is Mari? Mari is the immersive compute network. Uh what does this mean?

Let me start from from our vision. So today we actually have media in in 2D in frames like we see it on the TV on the smartphones. Our vision is that actually that media comes straight from just peeping right into your eyes. So something like this. So we imagine a world in which actually uh you are the UI and actually the whole world is UX.

My name is Vanessa and the reason I'm here today is to announce that Kreda and Mari have partnered to expand the Kreda metaverse into the augmented reality realm through the Mari network in [Applause] [Music] Korea. So okay so you may be wondering okay what has this to do with blockchain? is actually all this is running on a deep network. So I'll give you an example. Uh without actually a distributed network, this type of use case wouldn't be possible.

So let's take a look. This is a collaboration that we've done with uh T-Mobile. [Music] Hey, good morning. Good morning. How are you doing today?

[Music] Oh, I'm doing great. Slept in my cat bed. I woke up. Didn't even have to be anywhere. It's supposed to be in Dallas, but I'm next to my cat bed.

What is happening here? Uh, basically, there's a person that is getting their emotions, their facial gestures, and their voice being streamed in real time to a server close to the end user. Then that server close to the end user gets all this data and makes a 3D image that then it gets stream to the AR glasses digital world. All this happens in 32 milliseconds and that's it network. Yeah.

Let's go to another use case. Uh if you were wearing glasses like this uh probably by the next token 2049 we will do an experiment like this. This is the f world's first uh mixed reality uh discussion panel in which actually an AI agent was the moderator and the AI agent was standing here right in front of me. I think the planet earth will catch up very soon with the technology future and our mission is to push uh the boundaries and head very close. Thank you very much.

I would also want to receive a brief answer. Um, how do you handle the bandwidth and or latency issues that often arise with high quality realtime virtual experiences? Sure. Um, as I mentioned, you can see all the audience are actually looking at mixed media. They're looking either with the glasses or with a naked eye and AI agent interacting with the speakers in in real is a specialist.

So this is one of the use cases that we power today. Uh this is the second one. So this actually a commercial use case that we spearheading in in Japan. This is an AI agent which will be your customer service uh representative or your therapist or you call it. Let's take a look.

Thank you for your inquiry. Mari Corp. established in 2017 specializes in decentralized cloud rendering and streaming focused on interactive augmented reality AR experiences. Our advanced platform allows for real-time 3D content delivery to various devices including smartphones and AR glasses. By prioritizing seamless user experiences, we empower developers and businesses to create captivating AR applications that engage their audiences in innovative ways.

As we continue to evolve in this dynamic field, our goal is to enhance accessibility and interactivity in AR, paving the way for exciting applications across multiple industries. If you have any specific questions or need further information, feel free to ask. So, so as you can see, she actually is much better at explaining what our company does. So this is the next generation of our technology in which we're actually adding into the pipeline a full connection into an LLM and an SLM that is actually animating the virtual humans in time. Our vision is that actually for AI to become really relatable uh we actually need to interact with it like if it was our our best friend and he's we're giving AI a body and a face.

So what is what is Mari? Uh Mararyi is the world's first uh DP network for spatial computing. We've been building this since 2017. Uh we actually focus on uh orchestrating um a network of GPUs to render and in and streaming realtime 3D content to mobile devices uh close to zero latency. So now why special streaming matters and why I'm wearing these glasses?

Um I'm going to explain in in a second. So today when you press play on Netflix uh within less than a second you have a 4K movie in your phone, in your TV, anywhere. Uh as end users we we keep that for granted. But um when we actually watch the movies, we see that special content will become holograms, AI agents, everything that we're imagining in in the future. But bottom line is that for this to actually become a reality, the content needs to be accessible at the same level.

So it means that if I'm wearing a glasses, I just press play and I am interacting with an AI agent. So that's the elephant in the room that we have today that actually this type of content is not ready for prime time. This is because uh we have um technical bottlenecks and infrastructure limitations. So let's go over the technical bottlenecks. So when we think about uh 3D content um we are now talking about uh gigabytes or terabytes of space.

So when you are downloading uh something it will take you five minutes to an hour and we are in uh the generation of Tik Tok in which you swipe into seconds and you have new content. So end users are no longer used to to wait for for content. The second one is the format. Uh we have seen 3D content in 2D screens for more than 20 years. Game engines were designed that way.

But in order to really scale 3D, we need to keep it in its native format and stream it to the devices. And lastly, um you've seen uh Meta Quest, the Apple Vision Pro. It's super heavy. Now, I'm wearing the Meta Ravens that has AI but doesn't have a display. But in order to really get to this limb of of a glasses, we need to offload the rendering and the AI compute to the edge.

And that's what malware network does. So lastly, infrastructure bottlenecks. So in 2021 when we were actually going to scale our network, we had a lot of traction with some Fortune 500 companies. Uh we went to AWS and GCP and and we asked them for a very specific request. Okay, so we need distributed GPUs in many cities and and we need it at scale and you may be surprised but actually AWS and GCP were not able to tailor our uh requests and that's when we decided okay let's let's go web three.

We got inspired by render network by helium we saw how quickly actually the supply could be bootstrapped and we didn't look back. So how did we break the the stream bottlenecks? First of all, um since three content is very different, we had to reinvent the the rendering pipeline to solve the latency and bandwidth problems. Second, uh we knew that we needed to create a streaming format that is native to augmented reality, not not make an modification of 2D content. And as you see, we rebuilt the rendering pipeline and we actually rebuilt that rendering pipeline in a way that can be streamed via network.

So that's what we created the MARI platform. Uh three layers. Uh we have the Mware network which is the centralized orchestration layer, the MAI engine which is our core technology. Uh we've been called the MP3 of AR. Uh we have four patents uh already granted and 10 more on the pipeline for this core technology.

and then the application layer which is what developers uh use to connect our core technology and the decentralized resources. So this is our vision today. We have very clunky headsets and very simple content. So Mari's mission is actually to bring this into very slim glasses and rich content around you. That's what we call a spatial streaming.

So what what is special streaming good for? Special streaming is good for delivering sense of presence. So when you need that next level of content. So the metaverse, you have a Roblox, you have Minecraft, you have a sandbox, very simple avatars, but in order to really engage with an AI agent like in um Blade Runner for for example, you need that level of quality. And that's what we bring um as a end user benefit.

Uh we achieve this through special streaming and split rendering. That's our core technology. Now you may be wondering why DPIN. Uh well deepin um is a way to solve a business model problem and a scalability bottleneck. And this is exactly the how we are addressing that actually there were not enough realtime graphic rendering GPUs distributed across the globe and now we're working on on changing that through deepin so in that sense mi network uh has been a deepin in inception in 2021 we started working on the architecture and it was way before the term was actually coined.

So also just to emphasize uh did you know that today actually we have more than 50 million uh XR headsets uh installed in the world and and they're being used and Meta is actually very aggressively projecting that by the end of 2027 we have we will have 100 million devices. Bottom line is that the amount of compute that these content and devices will require is exponential and right now we are not actually having enough compute for AI. So when we add XR which is a full special computing industry, it will become a a ve a very imminent problem and why now um there are investors uh like Samsung that are backing us because they believe in our vision and how content will be distributed in in the next few years. And this just tells us uh basically uh borderless 1KX and other web 2 investors. We're a very unique company in that sense in which web 2 and web 3 actually do believe in our mission and we're moving forward.

And just a quick summary of why Mari we've been in the space for eight years. Uh we started as a web two company that actually has real demand and revenue and we're moving into web 3 to to scale. We have a patented technology and a very experienced team and we are the only network uh or DP network that is actually orchestrating uh three type of resources bandwidth uh storage and and compute in real time. So this is the scheme of the Mari network and we have more than uh 50 for 14 500 clients that actually are actively using Mari's technology and in the press if you Google Mari you will find a lot of articles in forbs in venture vid we've been called the aami of the xr cloud mp3 of AR and the $20 billion codec so now let's go into success stories um this is at the mute festival this is another exhibition in which we actually uh enabled an artist. I'm not going to play this video because I'm running out of time.

Mard, I have another two announcements to make. So, actually, we are the first DP network going into a commercial deployment uh in Japan. So, we we signed a deal with uh Nankai Dendetsu, which is a railway company. We will set up AI agent stations like the one that you see here in which they will be assisting tourists on how to buy uh train tickets and other stuff. And lastly, um we're doing a decentralized uh infrastructure offering.

So if you want to be part of the network, please join soon. Uh we have a promo video for for this. So please take a look. [Music] [Applause] Heat. Heat.

[Applause] [Music] All right. Thank you everyone.