00:00 Speaker A
So, um, Arch, how much does this cost to install and maintain?
00:04 arch
Yes, that’s a good question. There is no cost. The idea here is very similar to how we can leverage existing capacity in the form of electricity distribution networks in our homes and install devices that provide greater economic goods. So we
00:23 Speaker A
That’s why it comes with the new house I buy.
00:25 arch
That’s correct. It comes with the new home you purchase. And the way we’re thinking about building this is that a third of the new homes built by our homebuilding partners like Puti will have an additional node, but 100% of the homes in the building will have span panels. This is the material that is unlocked to allow you to add this large electrical load and battery. Therefore, the benefit for you as a consumer is that you pay less for your home and continue to have free energy, free internet, and free backup.
00:54 Speaker A
This will also reduce your electricity bill. How much do you think it is?
00:58 arch
It’s just that the electricity bill doesn’t go down. We discount your electricity bill because we can monetize your computing in the rapidly growing AI market and use that revenue to help compensate you for a certain amount of space on your premises and access to power.
01:21 Speaker A
Other basic mining questions that come to mind for consumers include privacy, security, and how loud is it?
01:31 arch
Yes, that’s a question. Again, the design of the system is actually such that it can be installed on the side of the house without causing any disruption or visible impact to the user. That is, in units of 60 decibels. So it’s quieter than your average air conditioning system on the side of your house. And it is designed to provide economic benefits. Therefore, if you consider the average household energy consumption and Internet consumption, it is consuming much more energy than the energy consumption of the computing nodes. So we essentially pay for all of our customers’ energy consumption and the energy consumption of their computing systems with the revenue that we earn from providing GPU capacity as a service to cloud service providers.
02:18 Speaker A
And Mr. Arch, how do you all make money? What is the business model here?
02:22 arch
Well, it’s a lot like owning a large network of cell phone towers, for example, right? These are pieces of technology equipment that are typically quoted in commercial buildings, and you end up paying a lease fee to the commercial property owner. This is similar to that. So we’re occupying a small space in our backyard that would normally be unusable, in a setback space between our house and the property line. And it’s quiet, harmless, doing its job. And in exchange for having access to that space and power, you can offer some pretty healthy rewards in the form of discounted or free energy and internet.
03:10 Speaker A
This is so cool, Arch. Or rather, has it actually been proven? Are these installed or are these boxes currently installed?
03:18 arch
Yeah. Well, we’ve been working on this for almost two years. And over the last few quarters of this year, we’ve actually been installing units in manufactured homes. One of our first production units is located in a community we are building in Sacramento in partnership with Pull Homes. And we were able to demonstrate that it provides extremely long uptime and high reliability for computing services, while at the same time not disrupting the power being supplied to homes. right.
03:52 Speaker A
Now, what do you think of Arch? Are there some hurdles that might come up when you’re trying to scale up, i.e. cost of capital? Is it a regulation? What is it?
04:00 arch
I think we’ve addressed a lot of the fundamentals of computing business models. I think what we’re spending more time on is helping consumers understand where the net benefit lies for them. Well, scale is important in delivery. Fortunately, thanks to our home builders, approximately 1 million single-family homes are built in the United States each year. Even if you take 5% or 10% of those homes, you can meet your entire inference computing demand on an annual basis. But it’s possible to do that in a way that doesn’t increase the cost of building electrical networks for large data centers, right? So we’re doing it in a way that we think allows us to distribute the economic benefits to all ratepayers or all consumers, rather than just requiring huge capital investments in large data centers.
04:54 Speaker A
Finally, Arch, please explain your partnership with NVIDIA. How did that happen?
04:59 arch
Well, Nvidia is clearly one of the leading providers of computing today, especially artificial intelligence and inference. As we were beginning to design and build X for Node, we saw an opportunity to partner with Nvia and also use their technology to build a liquid cooling system that could be installed on the side of a home or small commercial building. So we’ve been working with them on the design of the system. We look forward to continuing to work with them to deliver a very high quality service inference computing network.
