Rob, Hello, I'm Rob Hirschfeld, CEO and co founder of RackN and your host for the cloud 2030 podcast. In this episode, we talk about the highly interconnected nature of power, electrical power, and how the upcoming trend of AI data centers is intersecting as a load with generation, storage, transportation, Bitcoin mining and mining, all of these things are using power in really in really important ways, in different ways, but they all intersect in how we use and manage the grid, and that is our conversation for the day. Hope you enjoy it.
What do you what do you think about Nvidia breaking into the top spot is most valuable company, or there is it going to come down? Oh,
what goes up must come down.
That's true.
The question, the question is not if the question is one. So we'll see how sustainable the AI bubble is. But
there's a lot that reminds me of the web 1.0 crash that, oh, happening with Cisco. You know Cisco. Now, what Cisco did that NVIDIA is not doing is that Cisco basically financed an enormous amount of the purchase of their equipment for the and fueled the growth that way, which left a lot of dark fiber and unused, unused networking infrastructure when everything fell, fell apart. The question now is, will the, let's put it this way, will the used will the used GPU market be anywhere near as sustained, huh, when things start to start to back off?
Well, I mean, if we looked at the trend over the past 10 years or so, the answer is yes. Now admittedly, that's because the market kept finding new uses for those GPUs. Right?
What is there used GPU market.
I mean, there is a huge, there used to be a huge market for TPUs, for cryptocurrency mining, right? It's kind of cool down. But just in the heels of that, the AI stuff was growing. So like, Nvidia is a company that was in the right place at the right time. Having said that, though, Nvidia is also known for pretty much only producing power thirsty hardware like there, there's memes about needing, like, a chain rack outside of your house to power the next trend of GPUs. So there is not a current threat, but there is a potential threat, or the next couple of years, like once the this AI stuff stabilizes, and we reach at least good enough performance on the llms, the Asics, or at the release like low power AI processors or CO processors might displace the the demand for GPUs,
well, the demand for high end GPUs, yes, there's, there's reason to think that. Well, you've got a lot of research that seems to be doing, getting done regarding models that are not GPU hungry. And in point of fact, Microsoft itself has had a kind of a blip there and then I. And I keep looking for more read more results. And that's these one, what they called one bit llms, which arguably require only CPUs, although it probably to make it really work for any length of time, requires that a new kind of
CPU chip
get put in place that's a little bit better in memory management on chip memory management that said that that's the kind of thing that would make a change. More distribution is going to be called for, for a number of reasons, including putting more AI at the edge, including spreading out the consumption of electric power so that you don't fire up these enormous clusters and dim all the lights in Seattle. You know, it's
I also don't expect like in via to like crash and burn after this more likely to be a path similar to what's currently happening to Intel, where, like again, past five to 10 years, they started losing their their top Stop, their top spot on the consumer CPU market, and at the same time, now they're also being threatened by arm and regs five platforms, because for regular consumer hardware, like particularly people who only use it for internet and office and not really gaming a power shipping device is just good enough.
It's interesting because, like, they, they, they have the benefit of, you know, adding cards are much faster moving. Can be much faster moving because they're not producing whole systems, right? Intel, Intel, as much as they're making a chip, they'll, they're still system, you know, they end up being part of a system component, where Nvidia is mainly still making things such as slot into it. Other systems, and they can move faster. They did try to acquire arm and and get into the arm market. So, I mean, it's one of the things that we and they have Mellanox, so they're sort of building up. I just don't see it converging into a system. You know, your intel, your intel analogy is apt, right? They could take arm they could take the GPUs, they could take Mellanox. They could actually be moving towards a much more integrated system product. I just don't see enterprises as you know, the buyers being as excited about what they're doing. From that perspective, I see these
acquisitions, or types of acquisitions, being aimed more towards cloud providers, particularly with the mail and ox part, there is a ongoing effort, and Google has showcased some of what they've been doing to be able to just have a form of GPUs that you can connect to arbitrary servers. Are supposed to have them permanently connected, right? And basically, as specialty,
as a specialty, specialty hardware that can be kind of brought in on demand, yeah,
if I had to guess, that's how Nvidia is trying to position themselves, which is which makes sense again, given that they're power hungry devices. Data centers are better suited at powering on our side of those, like being able to provide a platform where a cloud provider can say, like, okay, yeah, I'll just pull up a rack of these devices, plug them into interconnected with my server rack, and off it goes. It makes a much more attractive platform, and then, say, buy a whole set of GPUs, and I'd have to install them on each of my white box devices that may not have. Yes, like the boards to support that.
Is there an architecture change coming here of a much more dynamic infrastructure choice?
I think it's almost inevitable.
More partially, they already, at least on the like, on the cloud consumer side, like, again, like seeing what, seeing what you need to do to provision a VM with a GPU that's low effort nowadays, like even in Kubernetes, like just religious labels and tags and you say, Okay, I schedule my work here. I would not be surprised again, if there was a lot of work being done on the back end to reduce that effort on the provider side, or reduce the cost, or increase the flexibility in the Dole, go hand in hand.
I guess, to me, the thing that's going to just right now the market's so hungry for these GPUs that we're not really talking about much architectural change, but I could see time with the server interconnect links the SLX access. But yeah, I think that's server si x, I can never remember the it's horrible. But right, you know, there's, there's, there could be what you're just, what we're all. I think talking around is, you know, being able to form machines more dynamically is, would be a really fascinating way to look at how the system systems put together. Interesting. And part of this is the topic for the day is related to this, because I think we're talking sort of around it. And when I want to go back to the used GPU market with because, right, there's a, there's a market for use GPUs in the crypto miners. But I actually heard or vice versa, if you were saying vice versa, the there's actually a market for the power allocation for the crypto miners also. And this was fascinating. I think this is the right the right bridge into the question which is, is power the next limiting factor here in power availability that the crypto miners are actually selling or being being bought for their power allocation? So to data center companies who want the, you know, what these, what these miners have, have managed to acquire from, from power, acquisition rights.
Yeah, it's fascinating.
You're going to lunch.
You absolutely are going to see. Well, the question is, can that ever be made dynamic? I mean, right now, you buy, you buy power rights, almost the way you buy water rights. If that is, you know, it's long standing. There's not a there's not much in the way of of variation. You can't. There's not much of a spot market for this. That kind of power doesn't doesn't make sense. It if you talk about changing the architecture, if you talk about distribution of GPU powers, you kind of then you can start to say, All right, let's, let's start. Let's actually create a different kind of spot market. But yeah, the the issue is also, then you talk about power being the limiting factor. But it's a, it's, in my mind, it seems to be a, there are a lot of knock on effects. So for example, with more distribution, you're putting a lot of pressure on the networks, because you've got distributed you've got distributed compute. You've got data distribution either because you need to spread out the heavy duty, or you're got data distribution because you you're doing a lot more at the edge. So. The kinds of pressures that you're going to see, I don't know which is going to come first. I think power is going to be a big issue, though. I think it's going to come. It's going to it's going to hit, not so much the big players, the hyperscalers. It's that mid layer that comes right after them, but they're going to be scrambling, scrambling for electricity.
I mean, we we see it some with the enterprise data centers, as they're pulling back workloads. You know, because they're, they're, they run, they're running, they're running infrastructure. But they, they are power sensitive, right? They're switching to a, you know, more power efficient AMD systems. They're switching to different cooling mechanisms. It's, yeah, it's a real constraint for for enterprises looking to build, you know, build these machines and buy by systems. Why? I'm surprised that you don't think it's going to be a hyper the hyperscalers are going to hit the constraints. You think they're just already ahead of it?
Well, some of them are ahead of it. Some of them have been planning, you know, their placement, their data centers, and they have been buying up they've been buying up power rights, power access. You know the reviews you take a look at what Google does, what Facebook, what meta does with their their data centers, they are they're relentless in terms of their pursuit of power and placement and and subsidies from states in order to, you know, kind of make the selection of where they put the next data centers. And there they seem to be, you know, running ahead of it. They have a longer, I think they have a longer viewpoint, well, they it's business
critical, right? If their power, if power is threatened, then that that really can foresaw their expansion capabilities, or they're just operational capabilities, does that then drive? I mean, I know what, what you know all of, all of the big providers are fundamentally financing solar farms as offsets. And they're, they're putting solar, a lot of solar, on the grid as basically a captive generation capability. And then Bill Gates was talking about funding nuclear, right? The nuclear, nuclear could be on the comeback as part of, you know, you know, as part of the sustainability for this, are all these, I mean, is this related to, there's a fossil fuel component. But even with fossil even if you turned around and said, We're going to say yes to coal and gas, they're not keeping we're not going to be able to produce enough of that to generate the power required for these systems.
Now the impact is, as you said, it's going to be on solar, it's going to be on to the degree they can find more hydro, it's going to be heavily they're going to be looking quite heavily at battery and storage capabilities, because, you know, solar and wind, you're, you have, you have time of day issues, and then you've got, You know, you're at the mercy of weather. So,
yeah, things that's so surprising with this, if you're doing solar and wind without battery, I mean, maybe batteries, but I think you know, the amount of infrastructure required, it's just enormous,
right?
Yeah, there's been some work being done in essentially alternate, alternate types of batteries, so as opposed to the typical chemical batteries that you would use gravity batteries, for example, is one that's been looked at you essentially just pump water into a area, and then, during the night, open the gate. And other ones are heat, heat batteries like particularly sand. I think it's the current favorite one, largely because, like, you can keep that an open in an open container, you can heat it up, and then you can use that heat again to produce steam and power things at night. And.
The demands to run that right. That That also means that you need to have, you know, functionally, twice as much wind and solar just just to run these base loads, right? That the base load this is so here's let me, let me ask it the way I'm trying to ask it, if you build a big compute farm, you're running that 100% of the time, the gear is so expensive you're going to maximize your utilization, but that means that you now have a base load on the systems, and one that you're not going to easily shut down. At least the crypto miners could turn things off pretty quickly. I guess most, for the most part, these aren't surprise load issue. You know, there's, there are surprise capacity issues, if it's right, I'm assuming, like a ice storm or temperature issues or things like that. Can you, can you can prepare for them to an extent that even so, you build a big data center, you put a lot of servers in it, you have a base you have an expectation to run them, and you have a baseload crypto miner said the same thing.
And crypto miners are effectively on the way out. Like, like, proof of work is a dying technology in the blockchain field. It's all moving towards proof of stake, which, on the other hand, is much more dependent on being online. 24/7, yeah, because if you if you lose a sufficient amount of epochs or ears, or whatever the blockchain. The juror calls it, essentially, I got sufficient amount of check ins. Then through reputationalist, yeah, yeah,
yeah. So they're they, they don't have the flexibility to turn, turn the mine, the the mining on and off at the same extent.
On the flip side, though, like though those instances there are a lot less power 100 proof of work.
The funny thing is, is that Bitcoin is on a tear because of the, you know, because it's been included in the indexes. So, you know, just, is it just not going to be? You know, when you say on the way out, I have trouble seeing Bitcoin as you losing its crown as the central as the central currency, right? Bitcoin asset currency
will might stay. It's just that they might switch from proof of work toward the proof of stake, like Ethereum showed that it's possible to do that.
Interesting. That would be an amazing transition. I think it would. It would make Bitcoin a much more attractive currency from that perspective,
right? Changing the mind, changing
it, changing it to a
currency, from being a a what effectively is not a not a currency is it's still it's it's price is based not on its role as a currency. Price is based on something is based on speculation.
It's a meme stock from that perspective, right? Yeah,
in that, in that way, I know the one. There's another Go ahead. Sorry. So the one
thing that might be a roadblock to Bitcoin switching to proof of stake would be the miners themselves. They they are the ones said to lose a lot of power, both physically, like literally, figuratively and this, I
uh, those, those the conversation we're having about the power being such a premium that make it potentially where they're sitting on, you know, a power, right? Or, you know, a data center asset that they could actually encourage them. They're like, Okay, wait a second. I could keep mining, mining, mining value keeps having or I could sell that asset and those rights to an AI company, an AI infrastructure use. And then, you know, encourage the switch, the proof of stake. I'm looking up how much Bitcoin, yeah, peaked. Announced back. It came back down from a from its peak.
There's also the specialty power um companies that bay that and I know this has been the case in Texas. Um. It, where you've got companies that have put in place power generation capabilities that they can fire up pretty quickly at great cost, but they and they run their their their actual duty cycle over the course of a year can be as little as 10% but what they do is, when you need, you need to, you know, bump up the amount of power in in the system in order to Keep everything in equilibrium, they crank these things up, and they are, they are profitable. They're particularly
it takes us, because they can charge search pricing, yep, yep, exactly.
So, you know, you you may find that there are, there's, a lot of variation in the business model of power generation for exactly the purposes. Well could be just with regard to AI and data centers, but it goes well beyond that. Go ahead
tangentially. Also gotta keep in mind that the production of power doesn't necessarily it's not necessarily driven by ideas and the demand or high power demand either. Like look at, for example, what's happening right now in Europe and France and Spain that the rollout of renewables has been so successful that they have a glut of power that they don't know what to do with it,
not when you say they're when is is The glut on a daily basis or a long term basis? Is it time of day
issues? That's fascinating.
I think it's been sustained for a couple of weeks now. I'll have to dig up the article again, but I mean, it's a trend on it's it's not the first time that it's happened. Like there's there's been other situations where power essentially came started going in, into the negative numbers, right?
Like, what happened with the way you do to
consume the power, because otherwise they can't distribute it, right? Right?
That's like, what happened with oil? When, when, when, during covid, when people stopped consuming and they couldn't, they didn't have any place to store it. But, yeah, I mean, it seems, it seems like some of this to to our earlier point, there's a, you know, when you have a power surplus, those are times when, yes, we'll start seeing stories showing up because it's, it's very cost effective, just the power arbitrage at that point.
Interesting thing. It's fascinating to think that we could be the US moves sort of slowly, but in getting into a place where, where our where electrical power, at least, is, is considered, have we have surplus? If that feels weird, feels weird to me, because, the way, because, yeah, the consumption models don't line up from that perspective.
It doesn't because where you're getting, where you're generating it, and where you where it's being consumed, are disparate and distributed. It honestly, in a highly, highly concentrated, you know, in a geographically dense kind of dense consumption of power, then I can, you know, which is more the case in Europe. I can see that happening earlier there than happening here.
The other part also is like, to a certain degree, in the US, there's a lot of regulatory capture, where the current power generators are holding all of the cards, and the they're essentially pushing law Market lawmakers to make rules to discourage individual power generation.
Like, for example, meaning private like solar is yeah,
like and in Texas, for example, if you run solar or wind for you at home and you. Produce more than you're consuming. You have to pay the utility company for that, right,
right? The distance they've been disincenting over generate over capacity on solar. Yeah, yeah, which is bizarre to me, but that's I guess in California had the same problem with people running the meters backwards and causing problems too, but on a consistent basis.
And and when you say causing problems, let's talk about and I think, because I think it's it's valuable, what are the what's the nature of the problems? Their economic problems.
They were economic. They were economics. I didn't realize it. I didn't know if there was tech, tech issues, of them like overpowering the grid from it's
not so much overpowering the grid is that they don't have the power switching infrastructure to reverse the flow. It's it's built very unidirectionally, and the it's essentially coming on by them in the rear because they they're not able to take advantage of of a much more fluid to call it market.
Well, I mean, part of the I've made the comment before, and it always knocks me out. Power semiconductors are built and put and and put in place in systems very one way. I mean they're the assumption when they're planned and when they're put in place is, you know, power flows one way, and it's only fairly recently. So what we're really talking about is switching, power switching to to Klaus's point. And we're also looking at, you know, how rapidly must that switching take place in order for power to remain in equilibrium? Then that's,
that's we all remember what happened when, when, when that fails. Look at the what was it a three or four, when we had the big, massive cascading failure in in the northeast,
yeah, yeah,
and the brown on Canada,
yeah. I mean, does that get solved? Because, like, household battery systems that I've seen one are very expensive. But also, you know, like, like, the one my son got has solar system with batteries, batteries in it, but the batteries are strictly a UPS. They don't, they don't really charge, like, they're not used to load shift for him,
and so that they're, they're meant for, for very short outages, right? Short cycles.
Like he can, he can charge it if, if the system's down, he can charge his batteries to maintain his refrigerator, basically, but it's not designed to take excess power he produces and store it for later. Yeah,
we might see a change with that. Sorry. I should do that. What
I was going to say is it may be depend less on the individual consumer
and more on micro grids so that you have what effectively are hyper local centers of storage and resale and those hyper those local markets, in fact, engage in the economic markets of trading, buying and selling power that to meet demands in various parts of the parts of The Grid, I
see a potential or increased household level power storage, particularly now that we've been what now, yeah, seven, eight years since EVs have been popularized, we're going to start seeing a lot of EOL batteries that are not good enough for electric vehicles anymore, but they're perfectly fine for a household. Yeah, on one of those batteries, you can power your your entire home for days, days.
Klaus, do you see that as as more realistic than neighborhood microgrids, or, you know, small regional microgrids that have, you know, kind of a local a local reservoir, if you want to think of it that way,
absolutely yes, because I. Seen enthusiasts do it themselves, like they literally buy used battery packs and build them into their own home. And these are people who, again, who are on the bleeding edge, who've installed their own solar, who want to decouple things from from the grid, but it is already being done. The only thing that's missing here is scaling it on a larger market, and with battery technologies evolving, particularly EV battery technologies evolving and more used batteries coming back into the market as again, EOL, for vehicle purposes, still perfectly good for a household where you have, like a
slower duty cycle. It's a slower duty cycle,
yeah, so what you're saying is, there's a there's there's kind of a longer, there's a longer economic value to EV batteries, which can be, you know, in other words, you remove them from the EVS, but you're basically dedicating them to a an environment that doesn't place the same kind of demands. Yes,
the only barrier that see right now is the cost like, there are, there are commercial, like, off the shelf options to do this. And I'm not talking just like the Tesla Powerwall, which is probably like the one that popularized this idea, but, but there's other vendors that that offer the same kind of power bank with like, 60 to 200 kilowatt hour capacity. The cost doesn't make sense right now, but if again, the cost of battery, of batteries is going down like, like, high voltage batteries. And at some point, we're gonna get to a point where it will just make sense to buy one for your home. And I would imagine no suit, no not much later than that, we will start seeing legislation to say every new home should have the capability of having one of those installed, or just have one installed of the bag would be
the this is, this would be, this is where it's interesting, because there's regulatory components, if, if EV batteries, when EV batteries start Flooding, the secondary used EV batteries come into market, you know, there, I suspect there will be regulatory requirements, because municipalities aren't ready to deal with them. There's no recycling for them, so that there needs to be a way to, you know, to absorb them in market. And it's interesting to think that the main barrier to putting in a battery from us, most people, is that the the panels in their houses are not structured to add in the type of intelligence and storage. But it's really a panel problem, more than it is a utility problem, right? So from if I wanted to add batteries into my system, right? You know, I have a place I could put them. It'd be nice if they could be, actually be outside, and you could, you could not worry about internal storage on them. But, you know, the main issue is, I'm talking about a full panel upgrade, but if there was regulatory support, like there is for switching to electric, electrifying my gas, my gas appliances, putting in heat pumps, if there was regulatory support for putting in a new panel, some of which I'd have to do anyway, if I was going to switch over to heat pumps,
you know what? It will make sense, especially with multi unit dwellings, which you know, more highly regulated than the than the individual. You know that's true, private, private home. As part of that, you make sure that it's pre wired, if you want to think of it that way, for exactly that kind of a thing and and it would make a lot of sense for a multi unit dwelling to have exactly this kind of a setup.
Well, it's, it'd be interesting, because it's a, it's a regulatory it's, it's a commercial benefit from for the, for the for the dwelling, in that the landlord, one can, you know, charge for power, and two has less, you know, should have less, fewer outages. So it makes the unit more the units more attractive. Also, it also means that,
it also means that they're less susceptible to search pricing on the on the grid and. And
their risk reduction. There's risk reduction in the event of, you know, weather events, climate events, outages like that.
In turn, it also means stuff the utility themselves doesn't have to have maximum capacity at any time they can. They're allowed or not so loud, they the impact of a brownout is going to be mitigated effectively if, if this is a rolled out at scale.
But it's the premise, though, is that that these units aren't competing for new batteries. To me, this is what's interesting, right? Is if, as long as those units are competing for with with vehicles, for batteries, you're you're going to have trouble funding that. If, Klaus, if your idea of them getting aftermarket, you know, after, after EV use batteries, then that create, that reduces the No, rich you don't think,
Well, I think that you're, you're right short term, but I think long term, not even long term, mid term, there's if enough, if it's clear that there is enough demand for either because of regulation, zoning, construction, things like that, there's enough time to put in place the production of the of appropriate battery technology and and again, EVs are going to, we're at a kind of a plateau right now in terms of EV adoption. I think that's going to, it's going to bump up again. All you need is a is a change in, you know, a massive, you know, oil crisis and
spiking oil price. I actually think the battery, the battery issue to me, because I went from driving an EV to driving a gas car. And a big part of the problem I had was that, was that I uh, the battery, the cost of replacement batteries was so onerous for EV so I would we, you know, I couldn't convince my wife that it was useful,
that There is, there is the potential, and it's something that Tesla had toyed with at one point and then abandoned. Sad to hear that. But particularly Chinese manufacturers are reviving the idea of swappable batteries or EVs um es or EVs, and we've seen this model work in places like Taiwan and Central America where or even like Colombia, Venezuela, where they have electric motorcycles, like mopeds And yeah, you go up to essentially a vending machine slot in your your your discharged battery, pay a nominal fee, uh, out pops a fully charged battery of an L slot that you can put in your motorcycle, and you go off, and it takes less than a minute.
Yeah, that's incredible, if
we're for larger EVs, and we see that this already being worked on for particularly for fleet vehicles, like trucks and so on. But if it if it picks up for commercial vehicles, then the demand it will happen or for a visa will go up, and at the same time, the availability of certified used batteries will also go up, which means that the capability of placing those in a home or rotating those out again is going to go up at the same time. Also, we've seen bio manufacturers like cattle work on technologies that reduce or try to eliminate the amount of rare earth materials in them, which is another thing that, again, EV dissidents have been pointing out as an issue. But if that. Is addressed, and it seems like ceramic is currently the main material that is being eyed for batteries. Then, yeah, high density, cheap, high density batteries that don't require law digging to produce
if you can make them swappable, or you can address the whole issue of the kind of the physical standards for making them swappable, it's
actually stunning. It's a difference, stunning to me that we have not arrived at standard form factors for these batteries. It's, it's so,
yeah, you can barely been able to get standards for for the adapters to charge to go home, much less, you know, we're
getting there, yeah, or at least in Europe.
In Europe, it is here. It's still a craze, yeah,
no, but it's, I mean, we're, you look at, you look at having what's, you know, standard, standardized batteries, is, you know, did for the industry in, you know, electronics. It's, it's, you know, it creates, it creates an opportunity where you could actually buy cars and then swap the battery chemistry. You could recondition batteries and other uses. It would create a whole nother. It would create secondary markets that aren't even, aren't even, you know, obvious at the moment, and then whole products like what we're talking about, yeah,
well, actually, let me ask you, in Europe, is The,
is the predominant supplier of battery, of batteries? Um, is China,
currently, yes, although with the new terrace, the manufacturers are looking to establish plants in Europe themselves to sidestep those tires,
who are the major manufacturers, uh,
BYD, uh, cattle, that's like, as in, C, A, T, L, yeah, yeah.
Interesting, yeah. I mean the fact that we have kind of, if not a monopoly, something about as close to a monopolistic control of EVs in the US because of Tesla,
which is potentially faltering,
yeah, Well, yes, it's
an interesting. It is interesting.
It it runs. I mean,
the other thing that we haven't really talked about is the power production companies themselves, you know, the Duke, everybody else who you know, I think I mentioned it a couple weeks ago. Were, you know, on the stock market, are just going great guns, yep,
that's what led us to put this on the topic. Yeah, exactly. It's fascinating to me, because I think one of the things that you would you would put on as municipal or small micro grids, which I think are regulatory and municipally very hard to do. Whereas Home Home, home users have a lot more choice, right? And property owners have more choice, the the it, we are definitely in, in a mode where there, there is a there is a in home, homes are actually decreasing their base load as they electrify and move to more efficient appliances, while we're seeing a huge upsurge in EV transportation uses of electricity and computational uses for electricity. So the computational base load is actually very consistent and predictable. The transportation base load, I think, is also predictable, but but buffered because of batteries. It's fascinating to me that the thing that we keep coming back to is storage capacity. Is the IS THE.
Storage and distribution? Yeah, I mean power distribution is still, I mean, people are basically saying, you know, invest either in distribution infrastructure, right, or invest in storage and make it almost hyper local.
And the it's interesting that our AI conversation translates into, you know, increased, increased demand that's going to be saturated through on, you know, solar and wind, which is, which is non, non consistent. We're going to have a consistent demand with inconsistent supply. Unless nuclear, unless we start ramping back into nuclear.
Yeah, I don't have a I don't know that. I have a particularly good argument to back this one up. But I keep thinking about, and I hear you talking about the individual, you know, residents, putting in place and spending pretty investing a pretty good chunk of money into local Residential storage, power storage. And I can't help thinking about the initial flurry of Web. Everybody was going to run their own web servers, and you were, you know, you'd run it, you'd have small office, and you'd run your own web you'd run your own email server and then hook it up to the net to in a very tiered approach. It works early on for those who are in an economic position to invest long term in that they have housing that you know can be converted again with the expenditure. I guess my question is, does that also hit a ceiling pretty quickly when you compare that cohort to the next, the next range of people, where their houses are not easily converted, their incomes, their revenue, their household income doesn't support it. I guess my question there is, who comes into to save that day?
I don't see it and they like looking at other examples, like, for example, kitchen layouts were, or where the lowest level on the economic scale of households tend to have gas Rangers, which are now known to be carcinogen, or carcinogens problematic. They're installed because they're cheap. The next lineup, like the lowest level of electric ranges, they're not very reliable, like they tend to work often going up to one step is quite a price change. So unless stuff and then, similarly, again, with, with the power generation, like, even like, like working rights and minimum wage, there is, there is, there's a strong potential for this to cause inequality or worsen inequality between those who have and those who don't have and those who don't have end up having to pay search pricing and be even less capable To dig themselves out, but,
but this, this is, to me, where, you know, electrical systems and home systems are highly standardized and and to the point about kitchen, kitchen layouts is a good one, right? We have standard every you know, everything at this point has gotten standard. You could panels are standard. You could replace panels, and you could include battery capacity into a panel layout, and with modular batteries, then you could actually start selling, you know, putting in, you know, you could, you could actually make that up, you know, some, some type of variable. Pricing system,
but it would be a long time before the grandfathered households end up replacing those.
Unless there's unless there's government subsidies or somebody something else, there would have to be incentives for for consumption, like we did for solar, solar effective incentives from that perspective, that's exactly
the point I was making earlier, that like this change to prevent inequality would have to come from the government. It the market is not going to do it.
It's a combination of standards, as you just pointed out, and incentives I completely agree with you on that are interesting, gentlemen.
I appreciate the time. This is, this is exactly where I was hoping this conversation would go, which is pointing out how intertwined these industries are. Thank you. Next week, we are still on and our topic, let me see, I have the topic defined vectorization and digital twins. So Oh, and then we have two weeks off, and then July 18, we have the two butt rule. And Joanne was able to get the author, ah, to join us. So July 18, for that one,
great when you talk about vectorization.
Yeah. I think the idea with vectorization is converting things into
inputs for representation. We're talking about representation, digital 10, basically simulation, representations like discrete events, emulation and things like that. Yeah, okay, that's a good topic.
It'll be fun. I'll see you.
I hope I know Joanne will will show up for that.
It's her wheelhouse. It won't be much of a discussion without or she tells how we how we did a draft of the conversation, and so we got it. We have to update for update, yeah, otherwise, I'll change the graphic, probably see All
right. Bye.
One of my favorite things about the cloud 2030 discussions is that we dive into the intersection of these topics. We're not afraid to jump from Ai, Blockchain, proof of proof of work, into generation, storage, nuclear, right, all of these things, home, regulatory, building codes. These are all interconnected in how technology works together, and I love that our group takes that on, is willing to discuss and explore and think through what the ramifications are the limitations of adopting technologies that might be two orders removed from the actual technology, or maybe even three. In this case, if this is fun and you're enjoying these conversations, if you're listening to me now, you clearly are, please consider joining us at the 2030 cloud. You can see our whole schedule. Read the books, be part of the book club. We want to have you included. I'll see you there. Thank you for listening to the cloud 2030 podcast. It is sponsored by RackN, where we are really working to build a community of people who are using and thinking about infrastructure differently, because that's what RackN does. We write software that helps put operators back in control of distributed infrastructure, really thinking about how things should be run, and building software that makes that possible. If this is interesting to you, please try out the software. We would love to get your opinion and hear how you think this could transform infrastructure more broadly, or just keep enjoying the podcast and coming to the discussions and, you know, laying out your thoughts and how you see the future unfolding. It's all part of building a better infrastructure operations community. Thank you.
