Not Your Father’s Data Center - From Inertia to Innovation: What Data Centers Need to Know About Grid Health

Episode Date: July 29, 2025

In this episode, Raymond Hawkins, Chief Revenue Officer at Compass Datacenters, sits down with Marc Borrett, co-founder and CEO of Reactive Technologies. Marc brings a fascinating background spanning electronics innovation, telecom engineering, and the energy sector. Having launched technology ventures from battery chargers to Star Wars toys, Marc’s journey led him from semiconductors to tackling the evolving challenges of modern power grids.The conversation explores the global energy transition, focusing on how the shift from traditional, fossil-fuel-based generation to renewables has transformed grid dynamics. Marc explains the increasing complexity of grid management, the need for real-time data, and the importance of measuring and maintaining grid stability in a world of distributed, less predictable energy sources. The dialogue delves into Reactive’s unique approach—leveraging grid physics to both communicate with and monitor assets at the grid’s edge—and how understanding grid strength and stability can shape smarter investments, operational decisions, and data center siting in an energy-constrained world.Timestamped Overview00:00 Entrepreneurial Journey in Electronics03:46 Chip Innovation Sparks NFC Revolution09:55 Energy Management for Carrefour14:27 Grid Communication Challenges Explored18:53 Power Station Grid Stabilization Explained21:43 Grid Stability in Energy Transition25:58 Optimizing Grid Signal Integration28:44 Accurate Real-Time Grid Measurement33:08 Time-of-Use Tariffs & Demand Trends34:30 U.S. Energy Demand: Duck Curve

So as we go through this energy transition, you're no longer in the controller and feeling like you're running a steam engine. What you're starting to feel like is you're in a car. And as you continue down that transition journey, that car turns into a motorbike. And now you've got a grid that is much more, you know, dynamic. And you've got a lot less time to fix things.

All right, welcome to another edition of Not Your Father's Data Center. Today, we are joined by co-founder and CEO of reactive technologies. Mark Borett, Mark, thank you for joining us. Raymond, nice to meet you, and thank you for having me. Yeah, we're excited to talk about energy and the grid. A couple of minutes on football. We're just a hard time not having a call without doing that. If you don't mind, we'd love to hear.

We know you guys are in sort of the grid management, grid, monitoring grid reliability space. But before we get into what you guys do it, reactive, do you mind catching us up a little bit on you? I think you're joining us from the south of England, right? Help orient where you are and where you are from. Yes, sure. Yeah, south of England, we have offices in London,

which is our UK base. And, yeah, we have a sort of a fairly international footprint. And certainly in North America, we have a growing present. with everything that is happening around, you know, the growth and the transition of the power system out there. All right. So before we get into what reactive does, do you mind telling us your backstory? How'd you get in the space?

Where'd you grow up? What sort of makes you who you are and how'd you get here? I would say I'm a country boy. I grew up in the south of the UK, but really enjoyed and had a sort of a effort. family history of innovation. My grandfather was an inventor. Not much happened. He came up with a lot of prototypes, but they didn't unfortunately go too far commercially. But there was a sort of an interest in technology and innovation. When I was 24 years old, I co-founded a company, which was really working with a colleague. We had met at my first company out of university. And we launched a

a battery charger for ordinary batteries. That put 40% on the stock price of the company we were working for overnight. It was a really highly interesting process to go through from getting involved in the development to turning it into a product and going out to Hong Kong and living in China for a year to get the factory processes set up and then to start the marketing coming back. So that was really exciting and I sort of felt that, you know, it would be great to do more of that. And it coincided that me and this colleague had a similar level of appetite for trying different things. So we set up a first company. That started down the route of electronics. We got involved in all sorts of different electronics, but we did actually

a toy for Star Wars. So when the range of Star Wars films came out, we had developed a technology to put a little chip in every Star Wars figure. So it would speak lines from the film. But you wouldn't have a battery or a speaker in the little figure. You'd just have a little chip. I think we made something like 60 odd million of those. And there was an electronics unit that you would hold the figure over and that would play the sound that was stored in the chip in the figure. And we didn't realize but we'd sort of developed quite a big RFID product. And then someone at Nokia bought one of those Star Wars toys and thought, well, this technology could go into a handset in a phone. And they contacted us because they saw the name on the printed circuit board in the in the

handheld device. And that led us to get involved with near field communication, which if If you've got a cell phone and you can use it as a contactless credit card to pay for things, that was the start of that technology. And we evolved into a semiconductor company and we licensed the electronics IP to go into a series of chips. And we licensed that technology to about 80% of the world's semiconductor companies. So, you know, that was sort of the first startup. the first sort of technology business I'd got involved with. But it was a hard business because if you've ever got involved or had any awareness around chips, they're very expensive things.

You know, the worst thing you can get out of going with a new chip design is a coaster mat for your boardroom that cost you, you know, one or two million dollars because a few resistors were rooted wrongly and you cannot use. the chips and you just get the wafers that you can use as coaster mats in your boardroom. And everything was getting smaller and smaller geometries. All the costs were getting bigger and bigger. So I didn't want to go back into semiconductors, but I really have a strong passion for innovation. So in looking around, the energy sector seemed like it was at a very interesting moment. This was in 2010. And there seemed to be a

a desire to, you know, change the way the power system was operating. There was a need for, you know, more generation, but of a much smaller scale. There was an emergence of electric vehicles where they might take power, but also push power to the grid. There was more devices that were connected and could, you know, change their time of use and how they actually operate to support the grid. So it sort of felt very much like a telecoms problem that, you know, lots of things connected, a need for control and diagnostics and sort of measurement. So that sort of led us or led me into the energy sector. And having had the experience of semiconductor world and telecoms, you know, we set up the company with a lot of engineering out of,

of the telecom sector. Engineers who had been involved with lots of these sort of high volume type of implementations, but there was a high degree of communication and control and understanding how a signal can actually be turned into whatever you need to measure. So that's really what got me into the energy sector. Gotcha. So as you start down that path, tell me, is that reactive or is there something before reactive? No, that was reactive. Okay. Along the way, so when I did that sort of first semiconductor company, I mentioned someone in Nokia bought the Star Wars figure and said we could use this. That person became a very dear friend and became my co-founder in reactive. So you mentioned looking around and saying, hey, I thought this energy space could be interesting. So this is kind of

of 2009, 2010 time frame. What problems were you trying to solve right at the beginning and what problems are you solving today? How has the business changed over almost 15 years? Yeah, I think it's quite an interesting story. As I said, you know, we started our engineering function from the mobile communications sector. So pretty much all of our engineers were based in Finland.

Pretty much all of them came from Nokia. And we wanted to bring a very different approach to the energy sector. And the challenges of really helping the grid transition from a very linear setup, which by that I mean you have a big power station that produces the energy at very high voltage. And then you have a network of subgrids that step that voltage down. so finally it gets into homes and businesses at safer voltages. But the power only goes one way. It goes from the big power station, eventually the home or business.

And the grid was organized like that. And now there was a need for or a desire for homes to have their own solar production, potentially batteries, potentially EVs. So now the grid was needing to operate in a bi-directional way. we would call that the edge. So the edge of the grid is really where the power ultimately gets consumed. And there was very little, you know, intelligence at that level. From a network perspective, it was really just a meter that recorded how much power was being consumed.

And there's clearly a need to have more, more intelligence to understand the state of something, what it could and couldn't do safely or in line with its operational. rating parameters. So we first started out working with, actually, our first customer was Carfour in France, and we basically set up a control system that would allow us our system to look at every fridge and every air conditioning unit in Car fours. We had like, I think, 200 supermarkets, we had about 30 big sort of hypermarkets, and then we had some big distribution depots. And our software platform was monitoring every single fridge, every single air conditioning unit,

and basically allowing us to play with the settings of the aircon or the fridge to create a reduction in energy on a temporary basis and through some sort of clever techniques allow us to create a shift of energy from one time period to another, but keeping everything operating within the limits and the operating parameters that Carfour wanted. So we're able to shift large amounts of power by two hours without affecting any kind of operating parameter. And that created very value for car four because that shift in energy could then be sold into the grid operator and they would be paid for it. Then there was a change of policy and two hours wasn't enough. They needed four or six, which sort of meant that that was beyond our reach really and they

played more to diesel backup generators. So we sort of felt we needed to have a change of direction and we'd had a side project going with the grid operator in the UK. And that side project was really about how can we use the grid as a communication channel. How can we talk to a very simple device like a water heater that is just like an electrical coil that heats water? And that typically is not connected to anything, hasn't got a Wi-Fi connection. So we thought we could send us a code through the grid, and we would create that code by creating a pattern in the frequency of the grid. So in the US, you have a 60 hertz grid. In the UK, we have a 50 hertz grid.

And we came up with a technique of embedding a tiny, tiny pattern in the frequency, which goes through the whole grid. And we set this project up with National Grid in the UK, who thought we were talking, you know, really about black magic and, you know, how could this possibly work? And we created these receiver devices and we put them all around the UK. And we sent us a command, which was called UK OK. And every single one picked it up. So we could, and this was in, I think, 2012 or so, 2013. we sent a command through the whole grid and it was received at every single corner of the grid,

which was a big deal because now we had a communication way of actually sending commands through the whole grid. When you say send commands, did you use the example of a water heater? Did you get something back from the water heater or do you just know, hey, it got through? When you say you got to every device,

talk me through how you get to know. Yeah, it was one way to start with. So yes, we confirmed that the water heater would have received the command. And the command would have been something like if you're in this region and you're not on, then we want you to switch on for 30 minutes, for an hour, two hours. And then that would be a command that would go to the water heater that would then execute the command. and we realize that we could also effectively ping the water heaters if that's what we're interested in, and they could send a command back that we could then work out how many were on at any given time.

So not a typical IP network, but rather still electrons going through a wire, but not IP, you're just doing it through electrons on the power grid. Correct, yeah. Interesting. Okay. Because also grids do want to know that what they get back is absolutely sort of untamperable with. So, you know, one of the things that grid operators were struggling with is, well, if we buy these services from, you know, water heaters or from EVs or whatever, how do we really know we get that back? You know, how can we measure it? And because we were using the power grid and no IP network at all, it was very hard for that to be spoofed, for something else to try and mimic or emulate or try and fool the system.

So there was a lot of interest and potential to use that approach. But, you know, being inquisitive engineers, such as we had, something bothered them. And the thing that bothered them about the result, because we were all delighted that it worked, you know, this is, you know, world first, brilliant. But there was a number of them who were bothered by the results. And the thing that bothered them was, well, you know, the UK isn't very big and we know, you know, that an electron travels at the speed of light. And why did this, you know, part of the country get the command quicker than this part? Why did we have to repeat the signal a few times for it to get through? And that really bothered them.

And then we looked and looked into it more and more and more. And they realized that there was something happening in the grid. The way the grid was behaving was different in certain parts to others. And in the parts where there was less of that sort of, you know, blocking phenomena, the signal got through quite easily. And the parts where there was that blocking phenomenon, took more times to get through. And we realized that actually what was stopping our signal was the stability of the grid. That where the grid had effectively a stronger, more stable

part of the power system, it was harder for our command to get through. We still get it through, but it would take a few goes. In areas of the grid where the grid was weaker, our signal got through first time. Mark, can I ask you to quantify what strong and weak means? And I'm going to give you dumb data center guy understanding as context our headquarters in Texas. We don't get a lot of cold weather, right? No. But when we got this horrible winter storm and everybody's using all of this energy to warm their homes and the generation side is having weak performance because of freezing weather, right? I think you have to seven or eight degrees. There was talk about, hey, the grid is not performing because there's equipment asking for energy and the energy is not on the grid

and that that's a problem. And then the other side of the problem is that there's more energy on the grid than equipment's asking for. I know that's a super simple data center guide description, but when you say where there's blocking phenomenon where the grid was healthy or it's unhealthy, strong and weak, what is strong and weak to me? Is it the infrastructure or is it the load? Okay, well, it's the physics.

So what you should, I guess, know is that not every electron is created equally. So an electron created by a solar farm or a wind farm is slightly different in one particular way compared to an electron that's produced by a gas-fired power station or a coal-fired power station. And the reason they're different is because the way the electron is produced. So in a solar farm, you have effectively a way through photovoltaics of converting sort of infrared into DC power, and that DC power is then transformed into AC power to go onto the grid and then get transported to homes and businesses wherever they are. But there is no, we would say there's no synchronous connection to the grid, because that power is produced as DC first, and then it is transformed into AC.

Now, compare that to a power station. So in a power station, you have got a fuel, gas or coal or nuclear, whatever, and you are creating a heat source from it, and you are heating water. up to create steam and then with that steam you're forcing it through a turbine or a turbine and that is that turbine is spinning at 3,000 RPM typically and it's spinning in the middle of an electromagnetic core and that is the thing that is then producing the electron at high voltage but it's it's that spinning turbine that is actually providing a huge amount of of stability or strength on the grid because in every power station each of those turbines that are producing the power are synchronously connected to each other so what that means is if

one power station fails those turbines can speed up or slow down their their sort of operating speed to compensate for that sudden loss of generation somewhere else. So they think of them like shock absorbers on the grid. They can take a sudden stress and they can help mitigate that stress on the grid. Is that synchronous nature at the power station level or is it at the grid level? In other words, can there be one in pool and one in Fortinman and one in Guilford and they can all talk to each other, or is it just the one in pool? No, no, no. They're all synchronistically connected. And the thing that is connecting them is that frequency that I mentioned. So they're all maintaining their speed based on that frequency

being maintained. So in the US, again, you've got a 60 hertz power system. So that means when your supply of power or your generation is exactly matching your demand, the grid is operating exactly at 60 hertz. When suddenly you lose a power station or you suddenly have a huge increase in consumption requirement, that changes the frequency. So the frequency can go up or it can go down. And the grid has to somehow bring it back to that same central frequency. And a big part of how difficult it is for the grid to bring the system back to that balanced state is a function of how many big spinning assets they've got on there. So if you had a power system that was completely run by fossil generation, which we had 20, 30 years ago, in the control room of the grid, that would feel like you're running a steam engine. you've got a big heavy grid it can keep going you just keep filling it with coal and it will run

through anything it will stay on those cast iron tracks and nothing is going to knock it off balance but you know that at the expense of that is CO2 so as we go through this energy transition you're no longer in the controlling feeling like you're running a power a steam engine what you're starting to feel like is you're in a car and as you continue down that transition journey that car turns into a motorbike and now you've got a grid that is that is much more you know dynamic and you've got a lot less time to fix things and you've got the potential to actually you know everything starts to happen much more quickly and you've got to have much more insight much more real-time data to give you the time, additional time that you perhaps wouldn't have if you didn't have

that visibility. So that's what I mentioned with the strong and weak part. That's super, super helpful. That helps me understand a bunch. I'm going to try to say back to you the other side of the equation and see if I'm tracking. So I've got my network at 50 or 60 hertz. A couple generators go down. That makes the network now running something less than 60 hertz. And all the equipment recognizes it and higher than 60 hertz, okay? And it makes everybody recognize it and we add some more generation and that comes us back down to 60. Yeah, yeah, exactly, exactly.

The problem is, I think, where I hear you going and where I think we might be going in talking about the strength or weakness of the grid is I can't tell a solar cell, hey, soak up some more sun. No. And I can't tell the wind, hey, could you blow a little harder, right? Because I need some more, to quote Star Trek, I need some more power, Scotty, when I call down to the lithium crystals in the power room for my Star Trek fans who know what I'm referencing there. So I can't do that. And this is back to your comment about I've had a car that's four wheels, it's super stable.

And now I've lost it and I've gone to a motorbike where I'm not nearly a stable because I don't have the same kind of stability of all these big engines. I'm now got engines that I can't kind of turn the wheels on as irks. Dial up the productivity or dialed down. Is that where when you say it's weak, is that what you're referring to? Exactly. Yes. So you cannot suddenly ask, as you said, Ray, the sun to shine more, or the wind to blow more.

And that's one of the challenges that you then need other assets that can provide that compensation. So batteries become a potential route to be, that fast response because when you are in the energy transition and depends, you know, we work with many countries around the world, but they all are at different stages. So some still have a few minutes to respond to an event. Others, you're talking maybe sub-second. So you start as a grid, you start to need to build in different kind of services that can immediately support you in that kind of, you know, very fast reacting situation. So when I was saying about strong and weaker, then

you start to see that when you look at a grid, you have areas where you have maybe a concentration of big generation and areas of concentration of high renewables. And the grid behaves quite differently in those two scenarios. So back to how we were trying to get our signal through the grid, it was linking back to that sort of synchronized part of the generation. That was harder for us to get through our little coded pattern in the frequency because we weren't able to move the frequency as much in those stronger areas, but where there was less of that big generation, the frequency moved a lot more so we could easily embed our pattern in there and get it through the grid. So, yeah, it just starts to come back to, you know, grid physics in terms of

how you can manage this energy transition. And you need to do that, but you also need to do it with keeping the grid secure. And that's why you start to look at, you know, how you invest more smartly in the grid, where you should upgrade and strengthen the grid in terms of, so in the UK, for example, National Grid has spent about 1.7 billion on buying stability. So that is buying almost like a generator, but it's basically a battery, or it's a generation asset, and it has a spinning mass that is spinning at that 3,000 RPM to provide that stability. So the more sort of stored stability on the grid, the less your frequency is going to move for a given stress on the grid.

And that's really where we realized the problem that we were seeing, that was the thing that was stopping our message getting through. We approached, went back to the grid operator and said, well, we could turn this whole thing on its head. and we could actually measure the stability of your grid and give you that on a real-time basis. And again, that had never been done before. I'm not sure they really believed we could do it. So they said, well, go and do it. But we're not going to tell you anything about what we think the stability is.

You go and do what you say you think you can do. Give us the data and we'll assess it. And we did that. We turned the technology. on its head. So instead of a communication technology, we turned it into a measurement technology. And we took some measurements. We really didn't know whether they were going to be accurate or not. We had no idea because we were given no guidance by the grid operator. We gave them our homework and they took it away for about a month or so. And then they came back and they said,

this is very interesting. We've assessed your data and you're in the right ballpark, but actually when we were doing our measurement campaign, a power station did fail. And when a power station does fail, that gives the grid operator one data point that they can see, they know how big that power station was, and that power they can check came off the grid, and they can see how much it affected the frequency of the grid and that gives them one measurement point and they said we were able to do a measurement of the stability using that power station and your data was exactly correct with the amount of stability that the power station showed was on the grid at that time so that turned us into a measurement company because we realized we could give a

real-time measurement of this very specific phenomena, but we didn't need to switch a power station off each time. So you could do it without a power station going down, which is much better. Yeah, exactly. Everyone's happier with that one. But we could do with a very tiny asset, and the way we do it is imperceptible to the grid. So it just, you know, they can't know what we're doing. They can't, sorry, they can't see that there is any kind of negative effect on their power grid.

but it gives them a measurement point and that measurement point can be delivered on a continuous basis almost every 30 seconds if they need it and that starts to help them understand what is happening on their grid and equally where on their grid are the stronger areas where are the weaker areas

so they can target their investments and I guess to be true to your audience the grid can also start to help understand where best are site energy intensive assets like a data center, but they can do it empirically and they can really help narrow in, yeah, this would be a really good location because, you know, the grid is changing and, you know, a strong area of a grid might be become a weaker area if that particular set of power stations is decommissioned and if, you know, solar and wind are put in their place,

suddenly you've gone from a very strong grid to something that is actually, you know, quite weak. And so you wouldn't really want to build an energy-intensive critical assets in that area if that was going to be the case. So this kind of real-time measurement data builds up a very deep understanding of the whole grid. And that is really where you can then make better policy decisions, better planning decisions, as well as better operational decisions to maintain the security of the system. Well, understanding the strength and weakness and what's the generation telling you

where to put big assets, right? Data Center businesses are big assets, right? Hundreds of megawatts, if not gigawatt plus nowadays. But before we get into that side, which I think is going to end up being a whole other episode of the podcast, I'm going to ask you two, again, dumb sales guy questions.

So I just can't get my arms around. I've got this massive power grid that's generating all these electronics, and that it's perfectly balanced to the demand of the electrons that all of my edge end users are requesting. I can't believe that we're electron-to-electron-match. That just seems too exact to me. What happens when there's too many electrons,

and then conversely, what happens when there's not enough electrons? Or am I just dead wrong, and it's exactly matched? We're never going to get to the electron-match. But second-by-second is really where the grid is operating, too. So second by second, the grid needs to make sure that there is that match. And if there isn't, then you start to see frequency deviate from that central point, the 60 hertz or the 50 hertz. And what then happens is some automatic controls start to happen.

So if there is too much generation on the system, some of that generation will be tripped off, you know, in a controlled way to... Will they slow down the turbines or fully shut them down? It can be, and or it can be, you know, renewable energy sources can play a role to sort of be downrated or temporarily disconnected. Equally, you know, with sort of from a consumer and a business perspective, there are time of use tariffs, which sort of guide... or nudge, you know, consumption behaviors in the right way. So, you know, if you have an EV,

then, you know, you're incentivized to charge it at night, you know, at 2 in the morning when, you know, there's obviously not the same amount of demand on the system. So those kind of things are the first protocol to try and, I think, get, you know, everyone to operate and consume in a general way. But, you know, in the U.S., I don't know, Ray, if you've come across this thing called the duck curve. I'm not familiar. Okay, I'll send you, I'll send you a picture of it, but it's almost like an outline of demand, the demand profile through a day. So basically, you have a morning bit of demand. I guess, you know, people are, it was going to work. It was going to work, and there's an aircon on. And then, but then when people are at work, you start to see

demand tail off. So you start to see demand drop down. And then, Through the working day, as you near towards the end of the day, demand picks up again, and you then start to see actually, you know, more aircon on at home. And so that demand really starts to recover. And the shape of it looks a bit like the outline of a duck. And what has been happening in the US for the last five or so years is really that as more renewable generation has gone in, especially solar, there has been more generation happening during the day, which is.

has been pushing, it's sort of a net demand curve, I should say. So it's what is happening in terms of the amount of power consumed and the amount of power produced. And during the day, especially around lunchtime, there is a lot of solar generation, but there's very little demand. So the belly of the duck has been getting lower and lower and lower. And then as you come back and there's more people now with EVs where it used to be just aircon, that there's a spike in demand towards the end of the day as people plug back in to

recharge and they put the aircon on and maybe lighting and those sort of things. So the duck has started to be sort of elongated. And what is very interesting for utilities with data center loads is that actually they are a source of energy consumption during those parts of the day when typically there has been little demand. So suddenly, you know, utilities can see that data centers can help bring back that demand and start to use more of that excess generation that has been produced during the course of the day from solar and wind. So there's a real positive for utilities to absorb and bring more data centers onto their grids.

Yeah, because those computers want to be cooled at 12 in the afternoon, just as much they want to be cool at 2 in the morning. That's right. They want their aircon on all the time. All right. So I'm still, because I don't think I quite get it, you said, you know, this excess that's been produced during the day. Where do the, if I don't have an air conditioner using that electron at one in the afternoon when the sun's at its peak, where do those electrons go? Again, there's a variety of ways. The worst is they're curtailed.

That's the worst. And you can see, if you look, you know, across grid operators around the world, curtailment is part of the playbook. It's not a desirable part of the playbook, but it's a necessary reality. So define curtailment for those of us who don't know. I mean, I can guess what it means, but in the energy space, what does it mean? In the energy space, it means that you have to stop producing on the grid. You are paid to not produce. Gotcha.

So that is clearly not an ideal scenario. Now there are more batteries or grid scale batteries going on to grids. That's a better alternative now because you can charge your battery at a low price and you can store that energy and potentially release it at the higher price. So that's why you start to see batteries. being built alongside solar farm. So, you know, even if it's a, you know, a really sunny day, you can store that excess energy, even if you're being curtailed,

and then you could potentially resell it back on when power is demanded and when prices are at a higher price. When you can get a better price, yeah. When you get a better price. So that energy arbitrage that batteries provide gives a better way of managing that sudden surplus that you might get. during different times of the day. You know, there are other things.

So hydro plants will also use, you know, when energy prices are low, that's when they will sort of pump the water back up the hill, back into the reservoir. Right. So it's just like a, you know, water base. The same thing as a hydro. Right, right. We're going to let that water down.

Pump it up at night, let it run down in the day. That's the model. So I get generation at the highest demand period. That's right. All right. Well, so, so Mark, if there's any. Anything I've learned in our 40-some-odd minutes of talking is that you have forgotten more about the grid than I understand and that we can't get it all done in one episode because I'm fascinated by what you're talking about and trying to understand. And we haven't even scratched the data center surface.

So here's our ask of you. Will you do another episode with us so we can record again? And we'll just pick up and start talking about data centers in the next episode. if we can get you scheduled. Yes, absolutely, right. I mean, you're right. We haven't even got to the exciting bit yet. I'm sorry, we haven't even started talking about data centers.

I'm still just trying to understand the grid, but this has been good. I feel like I understand it a little bit better. But, I mean, now is the time to start talking data centers. And we've sort of hit the limit of what we try to record in a single session. So, Mark, we'd love to have you back if you're willing. It would be my pleasure, Ray. Absolutely.