X SPACES: AXONE X STRATOS AMAđź”—︎
Introductionđź”—︎
Hi, I am still waiting a few more minutes to start. So we are receiving today, Bin from Stratos. Bin, I guess that you are the CEO of Stratos, right?
Yes
So very nice to meet you. I'm Emmanuel. I'm the CEO of AXONE. I would like to thank you a lot first before starting, Mathieu, our Head of Partnerships for having prepared this AMA and this interview. BIn, I'm very happy to have you here.
Maybe as an introduction, I would be more than happy to hear from you, to let you introduce yourself, to tell us who you are, what is your background and how you become interested in Web3. I think it's quite a long time right now that you are working on the project. So your experience and your background would be very interesting.
Sure. Thank you and hello everyone. It's quite early in the morning for me here in Canada. My name is Bin Zhu, and I’m the founder of Stratos. Stratos was established in 2021, the same year we conducted our fundraising, issued our token, and began working on the project.
I’ve been in the technology industry for many years, having started coding in 2000. For most of my career, I worked in enterprise environments, primarily with large companies and fully centralized systems. My last role was at one of the world’s top traffic websites, serving approximately 100 million users daily and handling around 100 billion requests per day. That experience was incredibly valuable and provided deep insights into large-scale infrastructure and system design.
I first learned about Bitcoin in 2015 and later discovered Ethereum in 2017, which led me to experiment with mining and smart contracts. As we observed the evolving crypto landscape, we noticed that, much like the traditional centralized world, most companies relied on cloud services such as Google Cloud, AWS, or Microsoft Azure.
In the current decentralized infrastructure and projects, including NFTs and others, most only function as ledgers. Bitcoin and Ethereum primarily serve as transaction ledgers, recording and securing transactions on-chain. However, there aren’t many decentralized applications (DApps) beyond wallets.
We saw this as a great opportunity and decided to build a truly decentralized cloud infrastructure. With Stratos, we started by providing decentralized storage, and we aim to expand into decentralized computing and databases. This would make it much easier for developers to build decentralized applications and infrastructure.
Since 2021, I’ve been fortunate to have a fully dedicated team working on this vision, and we continue to push forward today.
This is pretty impressive. This is very impressive in the sense that back to these years in 2021, there were not so many projects dedicated for storage. You were quite... It's It was quite innovative at that moment. Today, everyone is talking about DePIN and proposing a lot of decentralized resources. But at that time, I think you were pretty one of the first, right?
Yeah, true.
DePIN vs Cloud Solutionsđź”—︎
And so I understand that the idea came to, okay, why not propose an alternative on a decentralized way? But could you explain from your point of view and your perspective, what would be, according to you, the main differences between decentralized storage and more traditional centralized and cloud solutions? What would you say in a few in few words, to to to mark disadvantages and differences?
Yes, there are several key differences. The most significant one follows the principle: Not your keys, not your coins. If you truly want to own Bitcoin, you must control your private key.
With decentralized storage, however, the concept shifts—not your keys, not your content or files. Many people store their personal or business data on platforms like Google Drive, AWS, or even social media networks like Facebook. But these services require you to log in, and ultimately, your data is controlled by the company.
Now, imagine one day you wake up and realize that all the information you rely on is suddenly inaccessible because your account has been blocked. What happens then? We've seen this issue affect many key opinion leaders (KOLs), influencers, and YouTubers—one day, their accounts get banned, and they lose access to their content entirely. Even though they created the content, they don’t truly own it.
The same risk applies to money in a bank. If, for any reason, your bank blocks your account, you may not be able to access your own funds. This is exactly how centralized storage works—if the provider decides to restrict access, you lose control over your own data.
With decentralized storage, you don’t depend on anyone—you only depend on yourself. As long as you control your private key, you can access your files anytime, anywhere, without worrying about restrictions or third parties interfering. That’s the most important aspect, and it’s the simplest way to understand the difference.
Taking it a step further, let’s say you're using a public cloud service like YouTube or any other platform to store your data. You may have created the content, but once you upload it to a centralized platform, you might unknowingly lose ownership. For example, if you create a song or any other type of content and upload it to YouTube, you’ll soon realize that YouTube is profiting from your work. In some cases, you may even lose your copyright or ownership rights without fully realizing it.
Most people don’t think about this until something goes wrong. But if you take the time to carefully read the terms of service, buried within all the legal language, you’ll often find that your content no longer fully belongs to you.
In contrast, a decentralized solution never takes ownership of your data. You remain the sole owner, and your copyright is always protected.
In fact, decentralized storage can even help prove ownership of your content. When you create any piece of content, it automatically generates a unique hash code. As long as you can verify that this hash code was created by you first, it serves as proof of ownership. This is another key difference between centralized and decentralized storage.
However, at this moment, I believe the most important topic is AI.
AI and DePINđź”—︎
I wanted to discuss about that with you because definitely decentralized storage is a narrative that is coming back into the fashion with the conversion between AI and deep in. But for sure, what is very interesting to get from you is that because you are here, you are like a a pioneer on this field and you have been working on this topic for many years.
Could you give us a brief recap of how this field has evolved since 2021 when you started And the progress has made, and also the obstacles that has been encountered and that you had to face with?
There are two parts to this discussion, but let’s start with AI.
Over the past two years, AI has made incredible progress. Before OpenAI’s advancements, many people weren’t even aware of AI’s potential. OpenAI changed that—it was a turning point that reshaped how people think about AI.
I’ve been working with AI since 2018. Back then, we implemented AI models for facial recognition, helping businesses identify returning customers using their cameras. For example, if a store recognized a loyal customer, they could offer them special promotions. We also used AI-driven cameras to analyze customer behavior—whether they visited alone or with family—so businesses could tailor their services accordingly.
In 2018, AI was already a hot topic, but very few end users truly understood its impact. At that time, most AI components were complex and, compared to today, quite limited in capability. What seemed impressive back then now feels incredibly basic. Today, people are amazed by AI’s advancements, but in hindsight, the AI of 2018 was quite primitive.
For many users, AI was just a vague concept—something they had heard of but didn’t fully grasp. Now, it has become a transformative force.
In the past, most people didn’t have the opportunity to use AI or truly benefit from it. But now, everyone has access to it. I personally use OpenAI and ChatGPT a lot, and even my kids use AI to help with their homework—they ask it questions just like anyone else. Today, AI is accessible to everyone, and in many cases, it’s free to use. This is a huge step forward compared to 2018, and it’s exciting to see how AI is benefiting so many people.
The major difference now is that AI isn’t just a paid service anymore—it’s smarter and more widely available. It continues to improve rapidly, getting faster and more intelligent every day.
Why has this shift happened? The main reason is the advancements in GPU technology and computational power. If you’re familiar with OpenAI or ChatGPT, you know they rely on natural language processing. Similarly, companies like Tesla are leveraging AI for recognition and computer vision. Many of the core concepts and models behind these technologies already existed in 2018, but what’s different now is the sheer computational power available to process massive amounts of data. This has made AI significantly smarter and more efficient.
Another key factor is the massive amount of data being collected. Data and computational power are interdependent. You can have all the data in the world, but without the computing power to process it, you won’t achieve much. On the other hand, if you have computational power but lack data, you still won’t get meaningful results. The combination of vast amounts of data and powerful computing is what’s enabling real AI today—AI that everyone can use. This is the shift I’ve seen from 2017–2018 to now.
This brings up another question: companies like NVIDIA, AMD, and others are producing GPUs that make all of this possible. But where does all the data come from? I drive a Tesla Model 3, and I use it every day. I also chat with ChatGPT, use Google, search the web, and upload files. Everyone is doing similar things—talking to people, scrolling through TikTok and Facebook. This is where the data comes from. AI gets smarter by consuming the data we contribute, whether it’s from public clouds or the public internet.
This is a two-way exchange: we feed AI with data, and in return, we get smarter AI that helps us in countless ways. While this is undeniably a positive thing, it’s important to also consider the bigger picture: all of this data is being contributed by us.
We need to start asking an important question: Who owns the data? Today, most AI models use our data for free. But we should also be mindful of the fact that, as we interact with people and share information, we’re storing a lot of personal data in publicly accessible places.
In the past, this data was more difficult for companies or others to access. But now, with AI’s capabilities, it can link together a complete profile based on your online presence. It can determine who your friends are through Facebook, gather personal and professional information from LinkedIn, and even track your financial situation through credit card transactions. If they can access government data, AI could also learn where you live or what property you own.
While this level of insight can be useful, it’s also a bit unsettling. Right now, it may not be a concern, but in two years, the landscape might be very different.
Given this context, what advantages and services can decentralized projects like Stratos offer to people?
AI and Stratosđź”—︎
Because on one hand, you have the capability to be the owner of your own data. On the other hand, you have a lot of needs to train models, and you need a lot of data, and probably that there is a bottleneck with the data because now there is a threshold that as the big company scrapped all the web already and they will need more data, and this data will come from private data.
How are you playing the game in that context with Stratos? What is your position? How do you see that?
I believe there will come a point, a "red line," where AI will cross and cause people to realize they need to protect their own data. Once that happens, people will begin to feel uncomfortable with AI using their personal information. No one wants their personal thoughts or ideas to be used by AI to train models and then shared with others.
We all have friends and family, and we wouldn’t want them talking about our personal lives with others. The same goes for AI—we don’t want it spreading our private information. Yet, sometimes, when we seek help from AI, we unknowingly reveal a lot about ourselves in the process, just to get answers.
As people become more aware, they will start to recognize the need for a balance. They’ll want to use AI, but without sharing everything about themselves. They’ll want AI to assist with certain tasks but will set clear boundaries about what information they’re willing to share. Each person will have their own line—things they’re comfortable sharing with AI, things they want to keep private, and things they don’t want AI to access at all.
Everyone has their own boundaries, their own "line" that they don’t want to cross when it comes to privacy. This is where decentralized storage offers a unique opportunity.
In the current landscape, AI, public clouds, and public websites are like the vast ocean. Everything is interconnected, and information can flow freely between platforms. In this ocean, it’s easy for AI to link various pieces of data together—who you are, where you're going, and where you came from.
But I envision decentralized storage as a small, safe island amidst this ocean. It’s isolated and secure—protected from the vast, interconnected web of the public space. Here, you can have control over your data, knowing that it’s not exposed to the risks of the larger, open ocean.
Stratos Architectuređź”—︎
To keep this images of the small island of freedom preservation and privacy preservation. Can you give us more detail about Stratos and specifically how it works and what is its eye-level architecture I mean, how did you build Stratos and what are the tools that gives the people the allowance to preserve their privacy and to own their own data by their own?
The structure of Stratos is a bit complex, and that’s because, when we began working on Stratos, we saw that existing decentralized storage solutions, like Filecoin and Arweave, were already in the market. However, their performance was lacking, and there was a lot of unnecessary data on their networks. We realized that a complete redesign of decentralized storage infrastructure was needed.
As of now, Stratos has three main components:
-
Stratos Chain: This component stores all the hash codes, proofs, transactions, and Stratos tokens. It’s a separate entity focused on managing this foundational data.
-
Meta Node Layer: This layer acts as the bridge between your data and its accessibility. It holds all the metadata of your files—essentially, it’s the guide to locating and accessing your data quickly. When you need to find your files or transfer information, the Meta Node layer is where this happens.
-
Resource Layer: This layer consists of miners who contribute their servers to the network. These miners provide the storage resources for everyone who wants to store their data. Whenever someone wants to retrieve their data, they first access the Meta Node layer to locate it.
Once the Meta Node layer locates the data, it connects to the Resource Layer to either retrieve or upload the data. By using these three distinct layers, we’re able to improve performance and scale the network more efficiently. So far, the system is performing well, but like any complex system, we’ve encountered some challenges. This year, we implemented some major updates, and as a result, both miners and end users should notice significant improvements in performance and availability.
That’s the current structure. I apologize for any confusion earlier. Just to clarify: only miners can participate in contributing storage resources, as it would be challenging for end users to do so.
However, two or three days ago, we made a major announcement—Stratos Drive is live! This means that now, anyone can use Stratos without needing any technical expertise. All you need to do is download the Stratos wallet and, as long as you own Stratos tokens, you can use it just like a decentralized version of Dropbox or Google Drive. You can upload personal data, share it if you wish, or keep it completely private. There’s no KYC (Know Your Customer), no censorship, and complete control over what you share.
That’s the product we’ve introduced to the market. Our goal is to give everyone the opportunity to try it. If you're not comfortable with traditional, public solutions, this offers a decentralized alternative.
Stratos Targetsđź”—︎
Thank you. What is your target aim and goals? Are you expecting how many thousands people to use the solutions? Do you have a market perspective on this?
Of course, we want more people to use it, but we’ve only just launched the project, and we haven’t conducted any analysis yet. For example, we haven’t tracked how many people or how many wallet addresses are using Stratos Drive. We haven’t done this because we’re not sure if we want to track it at all, as we don’t want to compromise privacy.
We could potentially count how many users upload files to their drives, but since we don’t track wallet addresses, we have no way of knowing who they are. We’ve made a commitment to privacy, so we don’t track users. However, in the future, we might explore tracking on the blockchain side—specifically, we could look at how many wallets start using Stratos to upload files. This is something we could research later to better understand how many people are utilizing this feature.
Stratos Growthđź”—︎
I think this is pretty interesting, and that's why I'm very proud and happy of this partnership with Axone, because within Axone, AI builders, data scientists, and contributors can collaborate in various ways to build better AI environment and AI system.
Definitely, they will need some storage.
This is very important for us to be able to to provide the connection with as much storage solution as possible, and specifically easy storage solution as yours.
I think it's very interesting. Providing the capabilities to have a collective contribution and a collaborative contribution to fine-tune open-source models and to help people to provide their data and to help them to store this data is something very important, I think, for both of us, and I'm very happy of these partnerships.
And regarding specifically that collaboration, personally, are there any particular things you Would you be proud or thrilled to see build using Stratos as the storage infrastructure in a wider architecture workflow? Do you have hope, visions about that?
Stratos Drive is designed to protect personal information for end users, but we also have a strong focus on business users. Stratos has spent the past six months developing the layer for IPFS, and our team is now working on the environment for the mainnet. While we’ve already tested this on our test network, we expect to announce that Stratos IPFS will go live in the next two weeks.
Once Stratos IPFS is live, we won’t just be another IPFS provider—we’ll be one of the largest IPFS clusters in the world. Companies and projects using IPFS won’t need to change their code; they’ll simply need to update their gateway to ours. They can then store their data directly on our Stratos network, enjoying better performance and significantly lower storage costs compared to current IPFS providers.
This is another key milestone for us. We’ll be offering services to business users as well, including AI projects that require large-scale, decentralized storage. They can use our API or the user interface to upload their data to Stratos, all powered by the IPFS layer.
Stratos Tokenđź”—︎
Cool. I have also a question because here on the audience, we have developers. We have people who are very interested by the infrastructure side of things, but also, we have people who are interested by the DeFi and also the trading of the tokens.
Regarding, specifically, your utility token, do you have specific mechanisms to provide some value to the token holders? What are your perspective on this side of things?
Also because you have quite a huge experience on the market because you a lot of up and down. How do you see the life of your token by the year passing?
To use Stratos services such as Drive, IPFS, and eventually our decentralized database and computing network, you’ll need Stratos tokens. This is the utility of Stratos tokens: all services on Stratos require the consumption of Stratos tokens.
If you’re a miner and contribute resources to the network, you’ll earn Stratos tokens in return. There is a total supply of 100 million Stratos tokens, which you can use to access our decentralized services. Think of it like using Google’s public cloud services or Gmail—everything is powered by tokens. Similarly, to use Stratos services, you’ll need to hold Stratos tokens.
Currently, Stratos tokens are available on several centralized exchanges, such as MEXC and others. You can also acquire them through Osmosis DEX or via Uniswap V2 and V3 pools.
Stratos and Cosmos Ecosystemđź”—︎
Yeah, because I have seen in your technical documentation that your consensus algorithm is Tendermint. That means that we are a member of the same families of the Cosmos SDK. Yeah. Do you have other contact with the Cosmos environment?
Do you have partnerships with other, I don't know, for example, for some computation with a Akash or guys like that? Do you see a way to participate in the Cosmos family or you are doing your own path?
We have some partnerships in place, but we haven’t fully cooperated yet. This is mainly because we’re using Cosmos SDK and Tendermint consensus, while also supporting EVM. So, in a way, we’re bridging Ethereum and Cosmos simultaneously.
When it comes to computing power, we’re taking a different approach. Our team is currently focused on developing our decentralized database, which is specifically designed for AI applications. We’re still working on it, but we aim to deliver a test network for the decentralized AI database this year. This will support vector data, and we’re also calling it a "vector DB."
At the same time, we’re continuing to explore what a computing network could look like. For example, ICP is a major player in the computing network space. They’ve recently announced their own computing network, but we believe there are challenges with how ICP and similar projects are designed, particularly when it comes to user participation.
These systems are quite complex and difficult for most people to use. We’re focused on creating a more user-friendly solution and figuring out what we should deliver in this space.
Currently, we’re working on integrating DeepSeek, which is a highly popular model, especially with its recent implementation from China. We’re in the process of testing it and considering the possibility of integrating it directly into the Stratos computing network.
The idea is that, for each of our miners or users running a node, they could access DeepSeek at the edge of the network. While we will operate our own nodes and models, we plan to provide edge access to this larger model.
The goal is to enable decentralized AI agents or users who are concerned about their data being shared with public models like ChatGPT, to interact with a decentralized DeepSeek model. Users could contribute their data to train the decentralized model while ensuring that their data stays private. This way, they would receive the benefits of the model’s knowledge without their data being shared with others. If users own Stratos tokens, they would be able to access this service.
In the future, if you don’t own Stratos tokens, you may need to rely on other models, which could come with additional risks. Our goal is to leverage our computing power to support large models, but with an important distinction: the knowledge from these models will not be shared with other communities or models. Instead, the knowledge will remain localized within our network. This means that the model’s data will stay secure within the decentralized storage, ensuring that it isn’t shared or used by others.
This approach is designed to offer a safer environment for users, as their data and the insights derived from these models remain private. With our decentralized storage and computing network, we’re focused on providing this enhanced privacy and security.
Stratos in the marketđź”—︎
I cannot wait to see a connector between Axone and Stratos because definitely I'm sure that a lot of people will be able to use the data storage, the computation and also the model that you may provide with other data sources and everything.
I think it would be very interesting to have a connector as quick as possible between our two projects.
This is very interesting. Yeah, this is pretty cool.
The case is that you are not afraid to offer a complete offer because you have people who are offering computation, people are offering storage. Do you think that there will be an aggregation of that services by the same actors? Or this is the way you are going with Stratos?
Currently, we’re offering storage services with Stratos Drive, but we also have plans to provide computing power and a decentralized database, which will be available as services for users. These services will be easy to access, offering a plug-and-play experience, so users can start using them directly without any complexity.
For your project, there are plenty of opportunities for collaboration. We can provide storage for your project and any developer community working on it. To support you, we can offer a trial or free pilot program for storage usage, potentially lasting six months or even a year. This is part of our commitment to helping new projects scale and integrate decentralized storage.
In addition, for AI service providers, if you need access to large models like ChatGPT or even DeepSeek, we will provide open access to those models via our interface. This will enable your team to tap into advanced AI capabilities, while maintaining the security and privacy benefits that Stratos offers.
We encourage you to communicate with us because we can likely offer you more cost-effective and faster solutions. If we close the project with you, any off-structure manner node that is located closer to your project will be able to serve you. This proximity means you can access the node from the edge, which will speed up data access and processing.
We have an application named KatAI, which is a collaborative text to image training model. And definitely, I think soon this application will be open source and we propose some basic connection for this application. And definitely in the field of this grant, we could be able to propose this partnership in the months to come. Definitely, it's a very good idea.
Great. Yeah
Endingđź”—︎
Great. That's super cool. We are getting closer to the end of this AMA. Do you have some elements to add? Would you like that there would be maybe a specific items that we didn't discuss about? I let you the word of conclusion.
There are two parts to our journey. One is the Stratos infrastructure. We've spent a lot of time developing it, and implementing infrastructure is like building a great building. To make the building higher and stronger, you need a very deep foundation.
For the past couple of years, we've been working on this foundation, which is mostly underground and out of sight for end users. Now that the foundation is solid, we’re ready to start building the DApps and services on top.
This year, we will deliver many exciting updates and new offerings.
This is very good to hear that speech because we are in the world where everything has to be done so quickly and so fast. And that a lot of people are forgetting that, as you mentioned, from Foundations are so important to be able to build something solid. And I cannot... I mean, I agree more than that with you.
I know it's tough timing for builders because right now, many are focused on pump-and-dump schemes—issuing tokens, creating hype, and seeing people jump on the bandwagon. But most forget the long-term value. Sure, everyone wants to make money, but we want to do things differently. We truly want to deliver valuable products and services to the community, not just issue a token. In the current world, that kind of token might already be dead.
That's not the scale we're aiming for. On the other hand, we’re committed to continuing our work on decentralized storage, the database, and AI models. We want to combine these elements to create unique services that everyone can use.
With the new products and DApps we’re developing, our focus is to increase the utility and demand for Stratos tokens—building real value and creating lasting demand in the market.
It was a great moment with you. Thank you to all the audience. And thank you to you, Bin, for your time. It was a pleasure discussing with you.
I just have to give a To give a code for the Task & Quest, for the guys who are following this Task & Quest.
The code is KattDrop, K-A-T-D-R-O-P. Don't stop following us on the network.
I mean, Maybe you don't know Bin, but you already passed through this, but we will launch our main net in a few weeks now. So everyone is on the edge to provide the best network as possible and to launch the network and the token in the best condition as possible. So I know that you already passed through this. Now it's our turn and things are going pretty well.
Keep following us on the networks. And, Bin, it was really a pleasure to have you. Thank you, everyone, for your audience. It's a pleasure to have so many people here for this AMA and talk to you next week for a new Axon AMA. Thank you very much. It was a pleasure having you. Thank you. Bye-bye.
Thank you, Have a good day.