Big beautiful:
Hi everyone!
Da Piaoliang bought an Aethir machine before, and it has been nearly a month since then.
So today, with the attitude of being responsible for our own investment, we invited Jaden, the chief architect of Aethir.
Field:
Hello everyone, I am the Web3 Chief Architect of Aethir. I mainly focus on our web3 business and the foundation.
Big beautiful:
You are Singaporean and your Chinese is good, that’s great!
Field：
Fortunately, I grew up in Singapore.
My parents immigrated from China a long time ago.
Big beautiful:
Then let's get straight to the point.
I bought our machine and want to participate in the deployment of our nodes
So what do you think about our entire AI, mining, computing power, DePin and this entire big track?
Why did you choose this track as the direction of your entrepreneurship?
Jaden: Actually, our project started in 2021, before AIGC became so popular.
Our team previously had a very rich cloud foundation and an operating company for real-time rendering.
There have been many large businesses in the past that have operated tens of millions of users.
Our hope at the time was to reshape our reinforcement with a decentralized price, so that different GPU resources around the world could be connected to our network.
Then use this blockchain and TOKEN incentives to make this GPU computing network larger.
In 2023, with the storm of AIGC, we actually found that we had already built this platform that can perfectly support AI training and reasoning.
So we started to support different AI trainings at this time, including edge computing capabilities, so that AI business can have a more stable and better computing interface.
Big beautiful:
So my general understanding is that AI computing power itself is actually a very complete business model.
Your company has simply decentralized the AI computing power using Web 3 and it is built by users.
Big beauty:
But I would like to ask, even if all our users come in and we deploy this node.
Who should we charge for this computing power? Who are the real customers?
Field：
Yes! This is a very good question. What is the use of edge computing here?
Most of them are actually used for a lot of AI inference, low latency, and real time inference.
For example, if you want to do AI live streaming, right? In the live streaming, you want to replace the person you are in with another person.
In fact, this first requires a very stable computing latency.
Then the second thing is that you need to make inferences about this AI very quickly.
We currently have more than 200 customers, including startups and very large companies with valuations of hundreds of millions. They are all cooperating with us and working on this kind of computing power.
In terms of AI training, some of the world's most advanced AI training models have been cooperating with us.
Big beautiful:
Okay, I understand now. The computing power formed by our mining machines actually provides some edge computing power.
We will sell our computing power to anyone who has demand for edge computing power, such as live streaming platforms, gaming platforms, or listed companies, as long as they need it.
Big beauty:
Then I have another question, which is how long can our business last?
That is, how much are they willing to pay for us? How big is the market size of our business?
Field：
I can describe that the entire system of Aethir now actually has four major models.
First is cloud gaming
The second is cloud phone
The third is AI training
Then finally, edge computing
Looking at the overall system, we expect the annual revenue in the near future to be approximately US$250 million.
Currently, the edge computing power part actually has the smallest revenue for us, but it has the largest room for growth.
We can see that AI is just beginning to explode, and many of these early startups are just beginning development.
They will need more and more edge computing power, so the ceiling is actually very, very high.
Da Piaoliang: I see. That means our entire Aethir Group may have an annual revenue of 250 million US dollars!
The small machine we bought is edge computing, which is one of its four major revenue modules.
Big beautiful:
So, for a business as profitable as this, are there other people doing the same thing in our entire WEB3 industry? Have you done any competitor research?
Field:
In fact, we announced our current annual income and expected annual income in the future on the 16th.
Our monthly revenue is probably more than all other DePin projects combined, from their inception to the present.
Our monthly revenue is greater than all other different projects lifetime revenues。
Big beauty:
That's really impressive. Before I knew your data, I heard that our industry has a very old project called Render in the field of DePin AI.
Do you know anything about this project? Is there any difference between the two of you?
Field:
We do have some understanding of this. In fact, we really appreciate Render, because Render was first innovated in 2017 and 2018, saying that we can use blockchain for AI.
What they do is GPU rendering, and most of the rendering tasks they support are relatively simple, including perhaps generating 50 images.
In fact, most of the tasks they serve do not have very high requirements on time expectations, while many of the clients we serve have enterprise-level requirements.
Some of the world's largest gaming and telecommunications companies have very high expectations for rendering and AI generation time.
Now we have more than 4,000 H100 GPUs, which are the highest-end GPUs in the world, so we can serve them.
In fact, what they usually serve at Render is probably relatively small-scale AI production, so the difference in our income can be seen from here.
Big beautiful:
I understand. Then why do we have so many H100s and so much high-end computing power?
Field:
You can understand it this way: if we want to train and generate AI models quickly, we must have a very basic and solid high-end computing power.
When users are using the product, they need the inference to be closer to them.
So if we look at our tokenomics, we have support on both sides.
Our H100 computer miner can mine about 12% of TOKEN, and Edge's machines can mine 23%.
Why is there such a gap? It is because of the Aethir system. We know that edge computing may be a very large part of the entire AI market.
Many big customers can buy H100 from us, and they will also buy H100 from others.
However, we are the only one who has a solid foundation for this technology and can support relatively cost-effective edge computing power.
Therefore, we reserved a large portion of FDV to support edge computing.
Big beautiful:
So its uses are actually different. H100 is used for large model training, while our Edge computing devices, these smaller machines, are designed to support faster AI reasoning.
Jaden: Yeah
Big beautiful:
I see. That is to say, you have built the entire chain of AI computing power, whether it is the upstream high-end computing power for model building, or the downstream application-side reasoning. You are planning the entire chain of computing power!
Then use your economic model to connect all the interests.
Big beautiful:
I also heard that a new star has emerged in this track this year, called IO. What does it do?
Does it do the same thing as yours?
Field：
Aethir is actually an aggregation layer, which brings together different computing powers around the world.
On the IO side, it is an orchestration layer, which means that we already have many H100 machines, all of which are connected to the IO.
IO may have 1,000 H100s down right now, and about 800 to 900 of them are actually ours.
Big beautiful:
Oh, you guys still have a cooperative relationship, I thought you were competitors!
Field:
Yes, it is a cooperative relationship between the upper and lower levels.
In fact, we are in the industry, and you can imagine that we are engaged in high-end computing power and enterprise.
It is definitely the biggest leader.
Big beauty:
Oh, I see. In other words, we are providing him with such a high-end strategic core asset, and you have a competitive cooperative relationship.
Big beautiful:
Then I have another question. If we use the WEB3 approach and the DePin approach, will it steal AWS's business?
In my impression, AWS should be the leader in this sector.
Field:
We can think of it this way. In fact, decentralized computing power and centralized computing power will jointly support this trillion dollar market, the Trillion dollar GPU market.
We have our own advantages. Currently, when Aethir serves large customers, our rendering capabilities and our prices, because we have TOKEN, still have advantages over traditional centralized computing power supply cloud platforms.
We are now exploring new markets, including cloud gaming and cloud phones, which are not the focus of support for centralized cloud platforms.
Big beautiful:
I see. Actually, the market is large enough at the moment, so the competition is not particularly obvious.
Big beautiful:
Jaden, I want to ask a sharp question. Of course, as a business, I know you may have a confidentiality agreement.
But as a customer, I really want to know which customers our company has won so far?
What is the approximate meaning of this number of customers? Can you tell us a bit about it?
Field:
I can briefly explain our major customers and the subsequent business expansion. As I said before, we already have a very rich customer base for rendering, cloud gaming and cloud mobile phones.
Although we can currently name these customers directly, they include the world's top three largest gaming companies and a telecommunications company.
There are a lot of people, there are hundreds of billions of users using mobile phone services, and there are hundreds of billions of users playing games. In fact, I have been cooperating with them for a long time.
They are all using our relatively high-end computing power, and a lot of our income also comes from here.
The problem now is that customers actually have too many demands, and we want to give them the best service and the best support. We have been expanding our team and expanding the services of our platform.
Big beauty:
I see. That is to say, we now have many customers who contribute a total of US$250 million in annual revenue that you just mentioned.
Then there are many small customers, but their revenue contribution is small.
Field:
The situation is roughly like this: we may have 10 to 20 customers come in every week and ask us if they can use this computing power?
Our bigger question is whether we can support them well on our platform.
Is the support of our software platform solid enough?
Therefore, what actually constrains us here is not customer demand, but the progress of our products.
Big beautiful:
That is, as long as our computing power can keep up, and our software services can keep up with our customers.
This market is still relatively blank, so it will be relatively easy for us to expand our customer base.
I have another question. Now that I have bought this machine, can I use it to see which company is using this small machine to provide it?
After all, you already have tens of thousands of these machines out there, so who is using my machine?
I am very curious about this question, can I see this?
Field：
In the long run, we also want users to see that edge services are actually many chips working together to support a network.
This network is actually used by all our current customers.
We will disclose more of our existing customers later.
As for which customers are using it, we also hope that one day, the background can support this, that is, users can see them one job at a time.
In fact, I think it is also very interesting for users to know which companies to support.
Big beautiful:
This is really interesting because not only am I curious, but I think many people who get this robot are curious too.
Because we can’t actually feel it’s operating, and we can’t feel it’s truly creating value.
We can only see that its indicator light is on, so I am still looking forward to this function.
But it’s getting late. It’s been more than ten minutes since you came to our live studio as a guest.
Can we have a small request at the end of our video?
Even though you've come here, why haven't you brought any benefits to our beautiful community?
Field：
Of course, you can check out the event poster for details of the benefits!
Big beauty:
Well, today’s video is going to end here. Thank you Jaden and bye!
