Rethinking Grid Connections for the AI Era with Paul Manning (NovoGrid)

Hello and welcome to Transmission. Today's episode is with Paul Manning, who is the co founder and co CEO of Nova Grid. The episode is all about assessing the impact of new infrastructure on the energy grid. So what will data centers and batteries do to that network?

We look at data centers and we look at the the four nines of availability in terms of what they require from their grid connection versus the actual load factor that operational data centers bring. We then progress this to look at what the grid connection costs look like in comparison to the wider CapEx required for data centers. And finally, how those data centers manage non firm connections and what flexibility they and batteries might bring to the grid in future. Let's jump in.

Hello, Paul, and welcome to Transmission.

Thank you very much, Ed. Great to be here.

And as always, let's start off with who are you, who are NOVAGRID, and what's your role in the energy space?

So my name is Paul Manning. I'm the cofounder commercial director of Novigrid.

Novigrid, we are a grid analytics company. That is to say, we're really good at modeling the energy grid and then asking questions of what happens if we do this. We put a new data center in here, or we put a battery there, what happens? So at its core, Novigrid is a power modeling company.

And and for people sort of trying to get into this space, that really means sort of power flows. So you're talking about the the technical ability of the system to flow energy from one space to another and how that changes if the infrastructure changes or if more supply comes in or if, let's say, demand changes from, being base load to being very peaky, that that type of thing.

Yeah. Yeah. Exactly. So, I mean, I often think of it as like a a pool, and you start dropping stones in at different places and they create ripples.

And how do those ripples interact with each other? So if you built a new battery system in a place, is that going to improve the congestion in the area? Is it going to make it worse? If we add in a new solar, generator, what's that gonna have on the impact on the local network and then farther up into the more national networks as well?

So, you know, what happens if we do this in the future? All that planning has to be done before any kind of grid connection work can begin. So we often talk about Novigrid comes in at time equals zero, where the blank page starts and the question, well, what the hell do we do? That's where we talk to Novigrid.

And that's that that is so when, when someone is thinking, let's say, about a new point of demand coming onto the system, they would almost get in touch with you right at the start to say, like, is this grid connection point even remotely feasible for handling the demand that we want to bring onto the system?

Yeah. Yeah. Okay. Like right at the very, very, very, very, very beginning.

Okay. And do you go through into the engineering delivery or is it purely just a a sort of technical and feasibility type role?

So Novigrid was acquired, late last year by Opora, who are a Danish energy services company. So now, I suppose, as a group, we're offering that kinda end to end service from, you know, all the way through into the actual connection works, the the construction works and things like that. But Novigrid as a group with or as a division within the the wider group, we bring you from that blank page up to the point where you get your good connection offer from, say, in the UK, from, say, National Grid or from that one of the DNLs.

Okay. Which is very much a critical part of this process. Yeah. True. You were just describing people dropping pebbles into pools.

Maybe let's scale that up and let's have a Big Yeah. Let's have a giant pebble or let's have someone dive bombing into this pool, which is the emergence of of data centers and some of those enormously large loads coming onto the system. Yeah. How do you think about such a large change in demand for a space that hasn't changed in terms of demand levels for Yeah.

Very long time?

Oh, god. How do I think about it? Well, it's certainly keeping some people up at night. It's not necessarily keeping us up at night. I mean, just to kind of put in context the the change in in scale, Typically, your, kind of, your average operational data center in Britain is about fifteen megawatts or so, maybe twenty. There are, you know, there's the, I think the single largest one's about seventy five in Wales.

But we have applications coming in for, you know, a thousand megawatts, like the gigawatts, the gigacampus, as they refer to it. I mean, I appreciate some people have just added a zero in their spreadsheet and said, well, let's scale that up. But just to put, like, a gigawatt of demand in context, for the entire country of Portugal over the decade, this decade we're in from, say, twenty twenty, twenty thirty, their expectation is the entire country's demand in a decade will increase by a gigawatt.

So effectively, you have these hyperscaler companies coming around saying, I'd like to build Portugal for a decade in Middlesbrough or in Teesside or something like this, and and and national grid going, you what? You know, so there there's just that massive difference in expectation. Now now we can bridge that gap, but it's gonna take attitude changes. It's gonna take a little bit of engineering ingenuity and planning ingenuity, which are all feasible and possible to build all these Portugals all over the world to satisfy the, you know, the AI demand, really, and our kind of growing digitalization, you know Yeah. Trends.

And I suppose it's not it's not only kind of the capacity load.

Right? So applying for a gigawatt of grid connection where you only use it, say, five percent of the year Yeah.

Is a very different type of pebble to one which is a a gigawatt, and you're gonna use it Yeah.

Ninety nine percent of the time. Yeah. So so so those applications in terms of how data centers draw power from the network, is that something that you kind of you expect them to be really sort of solid base load, or would you expect them to be flexible in some way?

Again, going back to kind of what's operational at the moment, say that fifteen, twenty meg site, typically, they're only using two thirds to three quarters of that load. So let's say it was, you know, if it was fifteen megs, they might be, you know, maxing out a ten actually used. So you could effectively say that additional five, that unused five is wasted, you know, it's a stranded asset, so to speak. To do that at a scale of a gigawatt data center would be an absolute sin, and a huge waste of resources.

So what we're really pushing for is a move away from that kind of old fashioned idea of a firm connection, where basically you're two of everything.

Redundancy, if one failed, the other would kick in, and so to speak. And but that has knock on implications farther upstream all the way up to, you know, national grid and up to super grid transformers. So it it meant that it took an awful long time to basically reinforce the network to give you that guarantee that you would always be able to draw power.

And that was okay when we were talking five, ten megawatts. It's not really going to be okay when we're talking a thousand.

So the idea, really, if they want to get them connect those giga campuses connected in a reasonable timeframe, I mean, anything's possible with money, and I'll happily talk about their money later on, but really, when you look at the hyperscalers, time is their key criteria. They need these data centers connected reasonably quickly.

Now they can build the physical infrastructure, the racks, etcetera, fairly quickly, you know, give them eighteen, twenty four months. But to actually get into connected into the grid and have all those reinforcement works done, You know, they're getting pushed out a decade, fifteen years because that is definitely what it would take to give them these firm connections where you have to overbuild and over engineer everything. So the alternative to that is non firm or flexible connection.

You know, anybody who's worked in renewables or, you know, moto working in in batteries and things like that are very, very familiar and comfortable with the connection, a non firm connection.

It's, you know, it's just the day job for for most of us. So the data centers are gonna have to get it into their heads that the the way to get connected quicker is going to be via the non firm route. And that would mean then matching their computational load with their electrical load, as the grid can take it, because the grid will have to curtail them every now and again to allow for, you know, overloading, something like that. And if there was some sort of grid analytics company that specialized in assessing the risk of when curtailment occurs, you know, that is, you know, a very useful thing. But it is a quantifiable and measurable thing, that grid curtailing.

And you can tell them when, you know, to what extent they will be curtailed. So it can be planned.

And if it can be planned, it can be managed.

Mhmm. Well, let's let's come back to the the money of hyperscalers in a second and and why it does take ten to fifteen years to actually kind of make this change happen. But I think some people will find one of the things you said much earlier, perhaps not squaring with how they think about data centers, which is you were mentioning that maybe you had a fifteen megawatt connection. Your average load might be ten, and so maybe you're only using two thirds of your connection in terms of the usage.

But they'll know for data centers, it's kind of the four nines or the five nines of availability. So this thing has to be so available ninety nine point nine nine nine percent of the time Yeah. And yet they're only using it two thirds of the time. But those two numbers seem so at odds with each other.

So what's what's happening there?

Yeah. This uptime thing, the five nines, and and that means that they can they can only be off for something like half an hour and a year or something like that. The thing is when they get they're giving themselves this additional redundancy by only using two thirds, but they're grouping the data centers into these availability zones. So they might have three or four data centers of, say, fifteen megs within a single availability zone, each providing backup. So it's backups and redundancy on top backups on top of backups and redundancy on top of redundancy over engineered to provide this five nine uptime that they're that they're really looking for.

And they can get that if they're prepared to wait. Mhmm. And that's the key bit. They're not prepared to wait. So they have decided to that they need things earlier. So we're saying fine. That's possible, but this is gonna have to be taken into account.

Okay. So then let's come on to the the money question then. So around you hear billions being thrown around, and Yeah. I think billions is probably playing it too likely in lightly in terms of the size of of the depth of the pockets of these groups. Yeah. So why did they have to wait ten, fifteen years? Surely money can move everything much faster.

It can to a point. And, I mean, if you take, like I mean, it's it's hard to keep up almost with the the CapEx per megawatt. I mean, at one stage, we were, you know, we were going to conferences and people were telling us it's about fifteen million pounds per megawatt.

I'd say that number has changed dramatically even in maybe the twelve months since I've been at that conference, which was last September. You know, that number is only going up. So, I mean, let's just stick with fifteen, million per megawatt.

And for listeners, let's just describe quickly what fifteen million per megawatt actually buys you?

It buys you the racks. It buys you all the racks, the chips, and the cooling.

It doesn't buy you the land. It doesn't buy you the building. It doesn't buy you the electrical substation or the grid connection. That's all on top.

So this is just for the processing part of the day system?

Yeah. Yeah. Yeah. Yeah. Just the bit that's in the hole that you can see. I I often think of, like, I I'm gonna digress a little bit, and I'll come back to this one, but I started working for my father who ran a painting company, and he specialized in commercial and industrial painting. This is relevant to data centers because this is how I found out how much money data centers make.

And my father's company would have painted things like shopping malls and pharmaceutical factories and things like this. And in two thousand and six, I think it was, I was working on a site in West Dublin called Grange Castle, which is now where lots of the data centers are, but there's a lot of pharmaceutical factories around it. It's a large industrial area to the West of Dublin. And, I just completed the works, which was basically we were laying epoxy resin floors in the halls before the racks were put in on top of all the plinths, okay, to keep the static away.

And we'd finished the job. Builder was happy. Went into the quantity surveyor's office on-site to settle up the final bill, and I knew John, the the the QS for many, many years, and, settled the account. And I said, fine, and, you know, shook his hand, and he said, I have another job for you if you're interested.

And I said, yeah. Sure. And this was kinda back before they'd email you the the thing. So he literally would hand you a large envelope with all the site drawings and the bits about the job.

And, he gave me the job, and I said, where is it? And he just pointed out the window at the building that we just finished. And he said, it's there. And I said, but we just finished that building.

He said, yeah. They're gonna knock it in eighteen months' time and build a bigger one.

This was a data center for Microsoft.

I, from there, gained my curiosity about the the the sheer levels of money involved. So to go back fifteen million just for the racks times a thousand, I mean, it's fifteen billion.

Yeah. It's it's a chunky number. Let's compare that. So I think, interesting for the the rack cost and well done, by the way, tying that story back together. That was, it really was a diversion. We were we were off of the detail.

Yeah. I I I'm I'm the best bit about it. It's true.

Yeah. Yeah. Of course.

I never doubted it. The let's let's talk about the connection cost. So the the the connection cost versus the fifty million. So, like, what's the difference between those two numbers?

Oh, it's a fraction. And this is this is the point where, again, you know, kind of to go back to, you know, a story I know from a it was a guy who used to work for Novigrid who was very experienced network planning engineer, and he can remember when they received their one of the very first applications for grid connection for a data center. And this was not by modern standards a large connection, but at the time, it was huge. And they took the price and they tripled it thinking the data center operation would go away, PFO price, so to speak.

I'm not gonna allow it to swear on your your Definitely not allowed to swear, but, you know, I think I think we can get away with that one.

Oh, yeah. Yeah. But, basically, they went, that's fine. Where do we send the check to?

And the guys were like, what the hell was this? So the point on it is is that they are completely price insensitive because if you're looking to build a fifteen billion pound, that's just the racks and everything else. Like, you could throw another five on top of that. You know, let's say twenty billion.

Just, you know, why not?

And someone turns around you and says, oh, that will be twenty five million for your grid connection. You're like, oh, they're spending that on stamps, you know, or, you know, the the stationary cost is is less than that. So it's nothing to them, on it. So they're not interested. The the the the price is nothing to them. Now you'd asked me originally, but if they have more money than creases, then how come they can't just buy their way to a connection tomorrow?

It all comes back to a few things, but the supply chain ultimately and transformers.

Now transformers are the complete heart of any electrical network, stepping up the voltage and down the voltage. You cannot have a grid really without them. You know, either on-site ones all the way up to the super grid transformers or the small, old pole mounted ones you see every night again. You need these transformers.

And at the heart of every transformer is what they call grain oriented electrical steel. So it's, it's at the very core of the transformer and it's critical to what a transformer is. And this is highly specialized steel and it's not made in very few, in very many places. So, I mean, in Europe, I think there's only about three or four factories that make it.

In America, they went down to one. And the reason for that was, back in kind of the twenty tens, late noughties, the Americans were big into recycling and repurposing transformers, which was a great idea until they ran out of old transformers to recycle. So now they needed new ones to build new ones. The problem was they'd run out of the skills to to basically make the grain orange electrical steel, except for a plant, I think it's in Indiana, had to be rescued, by the American government, nationalization in America.

Crazy, but it happened.

And, the key bit now, they have since increased their productivity of that, but it's a highly, highly specialized thing to make this core of it. And there's a queue, so you've got to get in a queue, and the quality of the people in the queue, so when I say quality, if everybody has all this money, you know, Microsoft can't buy the way ahead of the queue versus Google, who can't buy it ahead of Amazon, etcetera, etcetera. So everybody else in the queue has the same amount of money. And it's like trying to buy your way out of some posh members club when everybody has money.

Then the other problem you have is when you do have the core, you have to make the transformer. And to make the transformer is a highly, highly labor intensive process. That is all done by hand. And the reason for that is all to do with the charge in the transformer. They can't use metal tools.

So what happens is around the core, they build a wooden frame of balasso wood, and he would have made models in school and things like this out of balasso wood. And then they have to wind copper by hand, a highly exact way around each one, and each transformer is made by hand.

And it takes thousands and thousands of hours to make each one.

So you can shout at these people. You can roar at them to build more. You can throw money at it, but it takes years to train a winder in, you know, the likes of the Siemens plant in Nuremberg and in in Germany.

But there's only one of Fritz or Hans who can actually do this highly skilled work.

He's not gonna work, you know, rush himself very, you know, just because your name is Jeff Bezos or Bill Gates or whatever, you know? You just have to get in line. And there's a certain lovely equality to that.

It's also highly frustrating that it is this choke point in the supply chain, but it's it's, it's just that is the reality. We often talk about, like, your money is fantastic until it hits the physical reality of the grid, and the grid is a wonderfully physical thing.

It reflects real life in how patterns are used and the energies used on it.

But there's something lovely about that, that there's still this point in it that there's just some guy highly skilled, and he's whining his copper, and you'll all just have to bloody wait.

It's it's it's almost unbelievable Yeah. The process that it goes through to be made. Yeah. It feels like something for another time.

I mean, maybe one thing that kind of jumps out to me is that, let's say, we talk about inverters. So inverters, another essential part of the electrical system, and a lot of the electrical system wouldn't work without them. Yeah. But they have, not exclusively, but to a a large portion of that market is done in other economies.

So for example, China can be involved in making lots of inverter equipment, which has obviously hit the news in, various parts of this year, but but that to me feels like like it's in a really similar space. Right? That's power electronics. Mhmm.

Yes. It's complicated.

But why why are we so why have certain regions retained such an advantage around making of transformers when we have things like, the sort of the the beast that is sort of Chinese manufacturing Yeah. That could pick up and run with this and potentially do it in half the time or half the cost. Yeah. Do do do we think or is is this something that you think that China will will sort of expand out of the inverter space into transformers, or do you think that that is unlikely to happen?

I I I expand out of the inverter space into into transformers.

Absolutely. I mean, China is, you know, they talk about petro states now, China being an electro state in the way, you know, look at what BYD have done with cars. It's amazing, you know, and incredibly impressive what they've done. And there's there's a couple of elements to it.

The one is, like, transformers, they're so critical that they are a national strategic challenge. So this isn't something you're just going to let go to somebody else. Like, the Americans made that mistake, and now they're rectifying it. You know, certainly, you know, here in in in Britain and Europe, we didn't lose the capacity, but we'd certainly need to to up that capacity.

I know, there's a a new, large transformer manufacturing facility being opened in Finland, for example, and and there's Wilsons here, for example, in in Britain. But, you know, I often talk talk about, like, you know, we talk about Great British Energy, but we really need to have great British transformer manufacturing company.

It's blue collar skills that aren't gonna go anywhere soon. You know, certainly Ireland would buy ten, fifteen percent of of of the capacity easily off of of you guys. But it's it's that nationally critical strategic element that you have to and you should keep it here. It's not a price thing, but making it cheaper.

Certainly, the Chinese will, I think, yeah, highly, if they're doing anything with regards to the batteries and and and coming from that under the mineral element of it, the material element, because obviously copper is is a huge element to it and the the steel that's in the core of it. Will they figure something out? Probably. I mean, I always, always have great faith in human ingenuity.

Will they do it? And we should certainly buy them because, you know, a transformer's not something you can hack. It's a physical thing, so to speak.

So let's then talk about the there's the ambition, right, of the data center space, which is enormous. It's it's kinda potentially, as you say, adding a Portugal in a single grid connection, but it's also potentially doubling demand for some grids in, or depending on the the time frame that you're considering. How much of that do you think is, like hyperbole versus actual this will come through and this will deliver? Because you must you must see hundreds of applications coming in for lots of potential. Yeah. How many of those do you think really are gonna come through and be successful?

Hard to put an exact number on, you know, the number of applications that will come through. I I can I can certainly say it'll be less than what's being announced by the hyperscalers, etcetera, but certainly it'll be several multiples of what National Grid publish in their future energy scenarios? So everybody's going to be wrong at some point. But I think that it's that key piece of if the flexibility, the non firm question, if they can get their heads around that, you'll see a big step jump in the number of connections, rather than just that kind of bragawatts, you know, we've announced we're doing five hundred megawatts.

Actually, we're doing twenty five now, and we're hoping for the other four hundred and seventy five later based on this, that, or the other thing, in there. So so certainly, everybody's going to be wrong, and I'm not going to add to that.

Okay. Very good. Very, diplomatic answer.

I think from from our from our point of view, I think you were describing things like great British energy as needing to go into the transformer space.

We certainly see in our modeling that when we get into the future, there's a lot of movement in terms of wind and solar that gets added, but there's not very much demand that has shifted, in the last few years. In fact, demand is kind of classically quite stagnant.

Yeah.

There's no I think there's some, like, there's something there around the more that we can encourage demand to actually connect, the better the system operates.

This whole thing doesn't work very well if we have thousands of hours of negative pricing. That is kind of just calling out that we've really got the sort of the the scale slightly wrong in how we plan and build this all out. So, yeah, maybe there's maybe there's space there for groups like GBNG to think about how do we encourage flexible, consumption from data centers.

Yeah. Yeah. And and another way of looking at that is, I mean, the fact that negative pricing exists and there's a glut effectively of generation that's going and begging, and there's all this curtailment. I mean, pick up the Daily Telegraph on, you know, of a weekend, and they'll happily tell you how many pounds have been wasted, from Scottish wind. It's one of their main bugbears.

Another way of looking at that is there's a massive opportunity to connect a load of load to gather this all up. And we see this, you know, particularly this summer, actually, we've had a large influx of queries in Scotland from that kind of central belt down.

Because between the b six boundary between England and and, Northern England and Scotland, so there's a a choke point there. It's very hard to get the energy that's produced in the lowlands of Scotland down to England. And even when you get to the north of England, there's not an awful lot of people live there until you hit kind of Manchester, really, which would be the first kind of most normally sizable area of demand. So you have this long stretch of area that needs demand, and while it's a grand place to be put in a big data center, particularly those kind of data centers that aren't required for, you know, what we call the latency or the distance to the eyeballs.

So, we often talk about, like, if you kind of imagined a two by two classic business school matrix with air distance to the eyeballs on the x axis and size of the data center, make a number of megawatts, so we say, on the y. You know, and at the in the top right quadrant where it's very far away, you need more megawatts. So they're the ones where you're doing this kind of, you know, the learning and models and things like that from AI.

Scotland is a perfect place for them, you know. I mean, there's the population of Scotland's about five million people. It's not enormous.

It's not London, for example. It has a small internet demand, relatively speaking, but it's a bloody great place to put those kind of large data centers for the learning models. But then, as you get down to kind of, okay, let's get out of Slough, a key bit really where everybody went initially because I think it was something to do with, it was an ICR Dulux paint factory or something like that, going back to my paint story from earlier, which is where the data centers went into in Slough.

If you if you drove up that way from, you know, from from Slough up to Birmingham, up to Manchester, that kind of access to everything, you're, you've got the population there. So you've got the eyeballs there, people watching Netflix and doing whatever. So the data centers can be much smaller, and we do an awful lot of work quite successfully on much smaller data centers in and around those kind of, urban metropolitan areas, because they're all about the inference, they're about delivering stuff really fast to you, as as it's being used.

So there's a there's a an interesting kind of mix and blend that that can be done, and it's all that distance to the eyeballs, kind of is the kind of the main criteria for it.

Mhmm. So So what's holding up those those data centers in South Scotland? Because as you say, it sounds like too good to be true. Right?

There's Yeah. Plenty of power coming to Scotland. It's there is room for data centers there. Yeah.

Surely, it seems like an open goal, so to speak.

It is. And and until you're you're you well, it's like anything else. It's new. You know?

It's the first time they're they're getting them. So, you know, you're you're making your application to Scottish Power, and they're going, oh, the hell is this? You know? And and they have to go back and they have to look at it, and you're getting that kind of initial hesitation, from them.

But it does make sense. It has to be kind of detailed out and planned out and analyzed, and we do an awful lot of work in that area kind of showing this is a good thing actually for the wider network, because what it can often do is it can often offset, in those kind of congested generation areas, it can offset a lot of reinforcement works that that probably shouldn't really be done. If you could have put a load in beside these generators, you would reduce the strain on the network, for it. So I think we'll see it. Well, we will get there. You know, there's obviously questions around fiber and and planning and and and things like that as well that have to be overcome.

But from a just from thinking from a systems perspective, the the kind of the the main logic here is if you can take your demand off system in Scotland Yeah. Then you don't need to build a very expensive transmission line between Scotland and England, which by the way, from a planning perspective, is significantly more challenging. Right? Because you're laying hundreds of kilometers of cable Yeah. Rather than just putting demand onto one particular site.

Yeah. And running one, I don't know, five kilometer cable or ten kilometer cable, which is is fairly easy in a straight line. Yeah. You know, it makes much more sense, to do. And and we're seeing, you know, clients of ours kinda coming in and and looking in those areas as well.

Okay. Super super interesting. We are, we are, on the face of it, a battery storage, podcast. And so making sure we kind of loop back in the battery storage side.

Yeah. You must be seeing some large connections coming through from the battery storage side and also around for both non firm and firm connections. So we'd really like to ask, how are you seeing the battery storage applications changing? But but also to ask a question around how are they reacting to the potential non firmness of some of those connections?

So the interesting one, everything we're seeing, we're seeing, certainly with the grid reforms, there's been a there's been a large slowdown in standalone batteries.

But everything we've seen coming through has been, I'd say, eighty, twenty co located solar PV and battery, with maybe just twenty percent PV solo, in particularly this year. I think everybody's waiting for the gate two to whole queue, outcome, which, you know, coming next month.

Yes. Today is the sixth of August, right? So Yeah.

Just in case this comes out in a slightly different time and nothing September What's going on?

September twenty five, everybody will get their, notifications as to whether they're getting a gate one, gate two, phase one, gate two, phase two. And the expectation, I think, is that only eighty percent of the standalone battery applications will get them.

Saying that, for those that do get them, they're gonna be fantastic projects. And the other key bit as well is for those reapplying, I think, in gate gate one, so they can go back and have a second bite maybe next year, it will provide clarity.

We'll be able to sit back, look at the flushed out queues, shall we say, so, you know, there's, you apply in a particular area, you're put into a queue system, some of those people ahead of you in the queue have been zombies, projects that are never actually going to be real, NESO's going to come in, flush those out of the queue, we actually get a realistic queue of projects that are actually going to connect, but then we get a very clear view of what's what and where the areas that are actually good to put batteries in, and will work in Britain. It's a really good idea, it's a really good process.

I appreciate there's some, you know, people are, there's always little bumps along the road in any process, particularly something like this, which is so complex and being done at such a speed for a regulator. You have to give Ofgem a hat tip on that one. Not often you'll hear that from me, but on this particular occasion, I think they're doing a really good job on a so I think we'll see an uptick next year in battery applications once the field has been cleared of the zombie projects.

Saying that, then the other area that we're seeing an awful lot of battery coming in is not an area that you might expect. It's go back to the data centers.

When we look at the non firm connection offers for the data centers, they still need to be able to provide the backup, the the the service, sorry. So they're going to need on-site storage. So what we're seeing and what we're doing an awful lot of is the design and scoping of on-site storage for exclusive use by that data center to basically help it ride through the curtailment periods effectively to provide a firm, kind of a, let me say, a de facto firm, if not de jure firm, a connection to it. So, there's a huge element to that as well. Will we ever see those come on the market? I doubt it because I think the market revenues from battery users like that is not going to cut us.

You know, a data center's earning, I think it's about a hundred and fifty pounds per kilowatt per month in rent. And if you're building a gigawatt, you do the maths, the number the zeros get ridiculous. So the the revenues from market revenues for the use of those batteries is is not gonna be, this is not gonna be existent. But But I certainly think from, like, the manufacturers, that's gonna be a huge market for them is is supplying to that on-site kind of effectively backup storage.

And that could be a really, really large, drain on kind of the supply chain of of battery storage. Right?

Idea.

In the does your non firmness, does it look like half an hour? In which case, no problem. It's a very straightforward battery. Yep. Or does it look like sixteen hours or twenty hours? In which case, a heck of a lot of battery to put to compensate for a hundred megawatts for sixteen hours. It's such just an enormous cost.

Well, relatively speaking, I mean, let's go back to what we spoke about earlier.

And if you're building a fifteen billion pound project that's just for the racks, you know, what's a few hundred million on, on battery storage on-site? I think I think the land and the planning permission is probably gonna be more of a challenge.

One interesting one, actually, complete aside, on it. In Denmark, because we were acquired by a Danish company, we found out recently there has to be a ten meter gap between each unit of a battery in a in a battery park. So it's just this massive thing, and they have to build individual walls between each container in Denmark, so it's really holding it back.

So I think that's a bit of a old school design, actually.

It is.

Definitely seen that change in other regions. Yeah. So maybe not to Denmark here to take a look at what sort of some of the best practices Absolutely. In other in other areas. Okay. We could we could definitely talk about this, for much longer, but I think I'd like to move us on to our final two questions.

Oh, we're there yet. Okay. Yeah.

Yeah. And so, is there anything that you'd like to plug?

For Britain, the gate two to whole queue, process is coming to a head this year. Everybody should get in September their notification as to whether they're going to be gate one, gate two, or gate two point two, and then they're gonna receive their revised offer.

Every single one of those projects is gonna require a curtailment analysis because the single biggest impact on your curtailment number is where you are in the queue. And the queue is gonna change. Everybody's queue will change. So that means that every single non firm connection offer is going to change, and the curtailment is gonna be impacted, usually for the better, by and large, not always.

So everybody's gonna need those turned around within thirty days from their offer.

And as far as I'm concerned, there's only one company that does the best, curtailment analysis in ten working days, which is no for it. So get your curtailment reports here is the way I look.

Very, very good. And and and maybe just to add a little bit more context to the non firmness, we've talked about it a lot. I I think that there are does non firmness, does it, to you, mean sort of fifty percent of the time running, fifty percent of the time not running, or does it mean sort of ninety nine percent of the time running, one percent not? I mean, like, how should people think about this?

Think of it like a dimmer switch.

And National Grid can walk into your living room at any moment and turn your lights down a bit for half an hour and then turn them back on again, or they might turn them off completely for two hours and bring you up for fifty percent for half an hour and, you know, just be really kind of a bit of annoying, playing with your lights. And and and that's kind of the the way to think of it. It's it happens at different points, and it can happen to different degrees. It It can be off completely.

You can be off fifty percent. You could be off five percent, or anywhere in between for varying different lengths, and it's all to do with how your project is interacting with everything else on the network to which the grid to which it's connected. And that is quantifiable in what we call a grid availability envelope. So it's literally time of the year, time of the day, and, you know, the percentage, let's say, charge in your case, for battery.

And we'll say the grid is available to you in this time, for this amount, length of time, and for this amount across the year. It's quantifiable, we can measure it on it. So it really is about being flexible as the grid needs, because think about what the grid is ultimately.

The grid is there, the grid exists to serve demand, and what is demand? Demand is just real life. Demand is just your house, my house, this office building, factory down the road, the tube here, you know, they're all just normal rhythms of life consuming energy, and they go up and they go down. And we have generators on the other side then going up and going down, and in the middle is the grid trying to match and balance all of this.

And it needs that little bit of flexibility to say, could you turn down a little here? Just having a bit of difficulty on this line. The congestion has passed. You can come back up now, please.

Thank you very much.

And and on that, if it's based around human rhythms, you'd think kind of, well, then maybe a two or four hour type non firmness seems like it might be a, a thing that would occur, or maybe it'd be driven by solar. And so it'd be, you know, maybe over this course for middle of the day, but unlikely to be, you know, six months back to back of this thing.

Unlikely. Yeah. That would be more of an outage scenario. Yeah. And, I mean, we're seeing, like, we had one, one data center, and and they're not their firmness was driven by two half hour periods in the year, one in December and one in, I think the end of January.

Then non firmness.

Yeah. They're and that was it. So basically, for half two half hours at two different points of the year, they couldn't be guaranteed a hundred percent of their access to the grid. That is non firm. Mhmm. And that is an extreme case of just these tiny little fractions of the year.

But if it's not a hundred percent, it's ninety nine point nine is not a hundred percent, you know? So that's the, the difference in it. Or alternatively, it could be, look, guys, pretty much every morning and every tee time peak of the day throughout the entire year, you're not gonna get access to the grid here. But the rest of the day, you know, fill your boots and do what you like.

Mhmm. And depending on who you are, so if you're a data center, those two half hours might be too much for you to bear. But if you're a battery asset Yeah. You might love it. Yes. It's fine. I will I wasn't gonna I wasn't gonna do that at that time anyway, so no issue.

Exactly.

Super interesting. Okay. And then, final question. So is there a contrarian view you hold something that you believe that the majority of the market doesn't?

So my view is between the hyperscalers, the data center operators on one hand having their shouting match with the utilities on the other, trying to get their group connection, speaking very different languages with a bemused regulator and government in the middle wondering why we can't all just fix this. I think that the data centers and hyperscalers would pay extra above and beyond for their grid connections to basically facilitate all of the grid connections that they need plus all of the grid connection work required to deliver the clean power, in in Britain's case, twenty thirty, at zero cost to the actual consumer. I think we could take the entire build cost for everything to facilitate all the renewables onto the grid, plus all the data center requirements, and have the data center operators pay everything at no cost to consumer.

And and just to be clear, is that from a financial perspective as in they would just be willing to pay for or are you saying that they would be happy to run a process alongside the network owners to almost build out independent parts of the network off their own back and then pass that over?

Yeah. Pass that over plus all of the works required for everything else in terms of bringing all the renewables onto the grid and basically paying for the lot in some sort of additional surcharge. And I don't think the utilities realize just how much money the data center operators have and are more importantly, willing to pay. There is a willingness to pay on the part of the the data center operators, the hyperscatter that the utilities and the government haven't grasped yet. And they could charge an additional surcharge per megawatt, whatever way you wanna charge it, effectively to do all of the upgrades without any cost to in your bill or my bill or anybody else's bill.

I think it's a really good place to leave it. People can take that away and think about it, chew on it, and think whether, they are willing to see Facebook sponsoring part of the, northern transmission line going in. Who knows?

Substation has been brought to you by Oracle.

We we we shall see. Paul, thank you very much for coming on transmission. You've been a fantastic guest, and we hope to see, lots of projects coming through very soon as a result of the, new grid connection offers.

Not at all. It's been an absolute pleasure, Ed. Go. Go. Go. Go. Go. Go. Go.