Your uptime, our priority

0:05[music] So there's been a massive explosion of data centers in UK in recent years and 0:12 all the signs point to that growth continuing. In fact, the UK is now home to more than 520 data centers and it's 0:19 going to continue to expand to especially with AI being such a big driver. And while many of the data 0:25 centers originally built were in the southeast of England, new facilities are coming on in other parts now in 0:30 Scotland, Wales and other parts of England. But while there's a lot of discussion about planning, energy, grid 0:37 capacity, and the challenges around getting new sites built, one area that is talked about less is maintenance, 0:44 repair, and operation support, MRO, um the stuff that keeps these facilities 0:49 running once they are alive. And that really matters because when a data center goes down, the impact can be 0:55 immediate and serious. There's a risk of downtime and if there's downtime, it could follow through with reputational 1:01 damage, operational disruption, and a very real strain on already limited engineering resources. 1:08 So today we're going to look at what happens when the data center moves from construction into live operation and why 1:16 that handover is such a critical moment and also how operators, procurement 1:21 teams and suppliers can be better prepared for the demands that come once society is live. To discuss this, we 1:29 have a fantastic panel. We're joined by Navidid Mayor, executive and board member from the data center alliance aka 1:37 the DCA. Hi Navidid. Hi Ashley. It's great to be here. Thanks for having me. No great pleasure that you're with us. 1:43 Um we've also got Phil Broaden uh business development procurement leader from SIPs. Another acronym but I'm sure 1:50 as probably most of you know it stands for the Chartered Institute of Procurement and Supply. Hi Phil. 1:56 Yeah, thanks uh thanks for getting me on board. Brilliant. Thank you. Thank you for joining us. And we've also got from RS 2:02 Components we've got Richard Graham, industry sector manager. Afternoon everyone. Hi Richard, 2:07 thanks for having me. And each member of RA team brings a different perspective on commissioning, 2:12 procurement, supply resilience and the operational realities of keeping a data center up and running. Navidid brings 2:19 the operator and industry view. Phil brings the procurement and government lens and Richard brings a supply service 2:26 and uptime perspective from RS. So with that in mind, let's get on to the debate. So I want to ask kick off 2:33 Navidbid because I think your core area of expertise is is a lot around the kind of evolution of data centers. But when 2:40 when a new data center moves construction into live operations, what does that transition actually look like 2:46 on the ground and why is it such a critical moment for that facility? I 2:51 mean it sounds like a handover on paper but in practice there's often a very real shift in risk responsibility and 2:59 pace on society's live. Absolutely. On paper it looks very simple but in reality it can be quite a 3:06 scary u situ situation to be in. It can be an absolute chaos to to say the 3:11 least. So um to give you an example, picture a high pressure construction 3:16 site where everyone's rushing to hit their deadline, sign off paperwork and 3:22 ultimately close the budget, which is what this is about. Uh these are very expensive facilities. Um the site is 3:28 handed over to the operations team whose main team uh main goal is maintaining 3:34 absolute stability uh uh and keeping the lights on on the ground. it can create a 3:40 real culture crash. You've got contractors, you know, packing up their tools to leave site while the operators 3:47 are um, you know, getting the engineers together to um, you know, scramble to 3:53 check if the building is actually running safely. So, all that is happening in a chaotic fashion and can 3:59 sometimes be quite um quite scary. U, this is the exact moment where 4:04 construction headaches often turn into operational business risks. Um, if you 4:10 treat handover as just a date in the calendar, that's exactly what you can expect to see on on a site. Absolute 4:18 chaos. And what you end up with is even when the site is handed over, you you 4:23 inherently end up with hidden problems that may not be evident to start with, but in months to come, it's going to 4:30 cost the operator a lot of money. And for the senior leadership team, a messy 4:36 handover eats through their margins and balance sheets in months to come. And often it's quite tricky to recover from 4:44 that without any losses. Sure. David, let's start to talk about MRO then. So looking back at recent 4:51 projects, what is often missing or undercooked in the MRO plan at the point of handover from construction to 4:57 operations? And where do those gaps show up first? And what does that tell us about how the site was prepared? 5:03 Yeah, it's a very good question. What I have seen time and again is um I think 5:09 what's almost always undercooked is the data and operational blueprint uh always 5:16 seems to be out of joint uh um uh things like completing uh maintenance 5:23 schedules, asset [snorts] logs, um you know parts lists um uh obviously you 5:28 know some of the very uh crucial uh blueprints seem to be missing uh right at the point when they should become 5:35 completed. Uh operators um often inherit multi-million pound high-tech facilities 5:41 packed with complex machinery, but they don't have the digital digital playbook to really tell them how to operate the 5:48 facility. Um so those those sort of transitional periods 5:54 always seem slightly disjointed and I think they could be really improved in 6:00 uh new facilities moving forward with some pre-planning uh and getting parties to work closely. 6:06 I think u many of the problems I've seen is due to lack of parties um working in 6:12 collaboration uh due to uh time constraints. Um ultimately these gaps 6:18 create immediate operational headaches uh in in particular in year one. Um you 6:25 get delayed preventive maintenance. You get inflated emergency repair costs that 6:30 were unexpected. Um I've seen how um you know frustrated 6:36 customers um start having arguments with builders on site and so it's not a good 6:41 position to be in. Um but ultimately it creates an invisible drain on your balance sheet which is what really it it 6:49 comes down to and um and before you even started scaling up you're eating through 6:55 your margins. Sure. Sure. Okay. So let's talk about MRO readiness at the handover stage. I think if we if 7:03 I put this question to Phil first and N you can reply to but who actually owns 7:09 MRO readiness at that handover stage? Is it the contractor, the operator, the FM 7:14 partner, engineering or or some procurement division? And why does this 7:20 so often fall between the cracks when the facility first goes live? That's a very good question. Um, and I 7:26 think the I go back to a couple of things that had already been said. Um, I mean, you asked kind of what it looks 7:31 like. Uh, you Navidid was talking about people running around under pressure. I mean, it's an incredibly nervous uh 7:40 moment. I mean it's uh there's a there's an awful lot of panic and all sorts of in your intro you talked about it 7:46 [clears throat] being immediate and serious and that kind of has a a knock-on effect on the way that people 7:51 behave. So it's all about preparedness. Um I think you know again something that 7:57 Navidid said there was about collaboration and team and relationships and relationships is something I think 8:03 we'll come back to a lot in this conversation about the relationship between you know all the parties you've 8:09 mentioned. I mean you mentioned there kind of you know kind of engineering uh 8:14 kind of contractors procurement all sorts. It's about being a team and I 8:19 think it's about the the relationships you have uh and the ability to really 8:25 kind of respond uh respond when things are at a critical point. You know it's 8:30 all that what if scenario. So I think you know one of the things that's kind of you know quite often missing one of 8:36 the things that fall where things fall between the cracks is there isn't that joined upwardness between all those 8:42 different players uh in the team you know kind of when when things go wrong and they probably will you know you 8:48 switch anything on quite often there's there's things that aren't quite right or aren't quite aligned or aren't quite 8:53 working if you've got a good team and conversations and relationships and 8:58 collaboration they get spotted quickly identified solution agreed and implemented really 9:06 quickly and you kind of smooth the bumps. I think we know where that's not it drops down the crack and it's very 9:12 hard to get it back out. So I think it's it is that collaboration that different 9:18 people will own different elements. It all depends on kind of the we talk about big data center. You see it in all sorts 9:23 of different teams and situations you know where there are where there are the disconnects or the poor relationships or 9:30 lack of understanding or lack of alignment then that's where things snap. 9:35 Sure. just to you know ask you know an adjacent question actually then so is there anybody who's specifically kind of 9:43 taking the lead or you know does is does it feel a very very collaborative process is is there somebody with a an 9:51 Excel spreadsheet somewhere is going you know we need to do this and pushing people forward in my view operationally and 9:57 contractually ultimately the operator holds the hold you know must be held accountable for what's going on on on 10:05 site But as Phil quite clearly said in reality there are lots of silos. So 10:10 you're working amongst multiple silos in a disjointed fashion. And often think 10:16 those um uh you know we could get rid of a lot of silos with some pre-planning 10:23 and um uh getting things in a more structured fashion. Um it all comes down 10:29 to um I think the start of any data center construction 10:35 is very positive. I think it's where you're now running out of time towards the end and you got multiple silos and 10:43 not clear about who should be finishing what and also uh at times it feels like 10:49 multiple parties are overlapping um all all with good intentions but it does cause a lot of um delays on site not to 10:58 mention friction between um different subcontracts and and I think there's a 11:03 lot to be said about you know ironing out all those issues uh but you know the the readiness needs 11:09 to be written into the contracts from day zero as as far as I you know I'm so 11:14 concerned if the operator doesn't legally require MRO readiness before 11:20 signing off on the building I think they're buying into a massive risk and and and that can be problem and I think 11:26 that really comes down to the leadership taking lead on that. But I agree completely with that and I think you 11:32 said earlier Nabbit about business impact. Uh yeah and you know it needs to 11:37 be taken seriously at that higher level and that what if what if what if this happens what if that happens what if 11:43 this goes wrong. So if you've got that understanding at a high level top down and you've got the understanding of the 11:50 business impact all the things that Navad listed earlier that could go wrong that could impact your business you know 11:55 short medium and long term then you kind of start to think about taking it a bit more seriously and make sure you've got 12:02 that joined upwardness and you kind of bring through a yeah theoretical dry run and all sorts. 12:08 Absolutely right. And it I think it really comes down to um you know the the 12:14 leadership the leadership forcing these internal teams to work in a closer 12:19 collaboration and and I think it really all comes down to the fact that the 12:24 building is powered up now. You know if there was no power to the building perhaps people could take a more relaxed 12:30 approach but now you have power to the facilities and the clock is ticking. Mhm. It's an expensive exercise in a 12:37 high-tech facility. And so if what you should really I mean what 12:43 I've seen is people plan in certain facilities people plan for the absolute best and they get average and that's 12:51 good. That's brilliant because you know you're bound to hit problems. You're not talking about 12:56 facilities where things are smooth sailing and and you know when you hit problems less about fingerpointing and 13:04 more about working together to get things sorted. Uh and you know because 13:10 everyone wants to finish everyone wants to submit their invoices. So you're all on the same team but under pressure 13:17 that can be a big ask. So, so now it's almost like you're saying in some instances it's not necessarily a bad 13:23 thing if things do go wrong because you know things will then maybe get sorted out at that time rather than if it just 13:30 kind of is the perfect handover. Quite often problems will emerge later down the line when they might actually be 13:37 more of a drain on the bottom line. Yeah, you could and and I've often been down this road not just in the UK but 13:44 Europe international markets. You can put together a a the perfect framework and everyone will buy into that and and 13:52 quite rightly. But as you progress through the project, you start facing 13:57 time constraints. That's where things start going sideways. That's where one 14:02 has to decide we we only have a year to go, but there's two years of work on paper. What do we do? 14:10 It's not a good place to be. You don't want to take any shortcuts in a mission critical facility. And so it needs 14:18 revisiting in a in a structured way. You can't just say we're going to cut that 14:23 corner and we're going to cut that corner cuz ultimately that's not what the client signed up to. Um but I do 14:31 feel that at the start of the contract if people were going at a faster pace um 14:37 they wouldn't necessarily be facing problems at the back end of it because you're always going to get 14:42 problems. So you want to buy yourself as much time as possible. You really want to hit the ground running. 14:47 Yeah. With these projects and you can only do that if everyone's on the same page. 14:52 Okay. So let's move on to talk about stocking supplies and readiness. So again, Navid, it' be interesting in your 14:58 views and from uh from RS's perspective, what Richard has to say on this, but how 15:04 should an operator decide which critical spares and consumables need to be on site from day one and which can sit 15:11 safely in the wider supply chain without increasing risk? So I think um you know this is ultimately the the RS business model if 15:18 you like. This is how we support our customers whether it's looking after a data center or any other piece of 15:23 critical um national infrastructure or manufacturing asset. Ultimately 15:29 understanding the the impact of downtime or you know a product going offline if 15:36 you like or breaking that ultimately causes a you know cascade effect or something else not to work or the 15:41 service not to be provided. uh understanding the um the criticality is ultimately the point and you know what 15:47 is criticality? Criticality I suppose for me is driven by a number of different number of different measures. 15:53 Um it might be looking at the uh the level of redundancy that's ultimately baked into the thing that breaks. So if 15:59 that first thing breaks is there something else to take over from it to to carry on running that service? um if 16:06 not what is the ultimate impact of that asset going offline or breaking and 16:12 what's the lead time like to get the replacement product online or you know back into back into place to to to take 16:19 over from the thing that's broken ultimately looking at that criticality measuring that criticality. Some 16:24 industries do this better than others. is um you know coming up with a scoring matrix or mechanism that allows um you 16:32 know stocking policy for spare parts to be geared by that criticality and also 16:37 then um you know engineer and you know human resource to be applied to that um 16:43 to that problem to go and fix that problem. Um you gearing it around those things is is really important. Um I 16:50 think when you look at you know the wider supply chain and the stuff that our customers are asking for from RS in 16:56 this space uh it's all about partnerships and testing people's people's capabilities. So it's not just 17:02 looking at price for a product is looking at their availability to supply that product. The value they can create around that whether that's be stocking 17:09 it locally stocking it on site through a managed inventory service. um just having a broad range if you like for 17:15 people to be able to to pick from. Uh because ultimately if you're looking after a complex facility like a data 17:20 center uh thinking about um you many people picture the white space when they think of a data center but really the 17:26 the the life support system is that sits in that gray space. It's the HVAC. It's the power. It's the water provider water 17:33 provision if you like to keep that thing cooled powered up and working efficiently. there's there's many thousands tens of 17:39 thousands of complex assets in that in that space and you can't predict which one will go wrong at any given point. So 17:46 having a supplier that um you know understands firstly scoring the criticality of those 17:52 assets in that space is important. Having a supplier that's got a broad range to supply things quickly uh to 17:59 predict when things might go wrong is also important. uh but then measuring that and managing 18:04 that supply with with agreed SLAs's and um you know through a proper partnership you tested response times I suppose is 18:11 is vitally important asking the question you what's the impact of this thing if it goes wrong within 24 48 uh 72 hours 18:18 having a supply chain and supplier suited um to be able to support you in that instance and n I mean do do you think you know um 18:26 yeah yeah do you want to just add to that but also you know kind of bear in mind how well do you think data centers is in 18:32 your experience how how good are they at managing? I think it really depends on who you asking the question. Um so you've got 18:40 the engineering team on site who needs spares to you know react to failures but 18:46 also you've got the operator who is running a business and um I think this 18:52 is a classic challenge of um balancing cash flow against your risk. Um, if 18:58 you're talking about um um I would say non-critical items, if that's the way to 19:03 put it, you know, relays, valves, filters, stuff that you can get next day and are well stocked. Uh you don't need 19:10 to spend a lot of capital on those items. However, if you're looking at any item, probably I used a transformer UPS 19:18 as as as a good example. Anything that has a long lead time and you are likely 19:25 to encounter problems in in getting that um piece of equipment, you need to plan ahead. And if you don't think you're 19:31 going to have the parts on site in time, then you need to keep it on site. You 19:37 need to keep spares on site. Um which obviously comes at a cost. So um but 19:43 that's part of the risk management and the framework that you need to put together to determine you know if 19:48 tomorrow we have a transformer failure what does that look like? Is it going to 19:53 drag down the entire business and um uh and you know uh you know what do we do 20:00 to to address that? And so for those types of um um components I would say 20:08 really the only way is to have to have it on site from day one. There's there's no other you know uh we work in 20:15 facilities where there's no um downtime. So but uh once upon a time um operators 20:23 used to you know stock a lot of filters and valves and all that stuff. Now uh with nationwide suppliers that's no 20:30 longer required. you know um most suppliers can get it to you uh even on 20:35 the same day. So that's no longer a problem. It just needs establishing a good relationship with that supplier and 20:42 giving the supplier to the opportunity to understand what your requirements are throughout the year so they can plan 20:49 ahead that does help suppliers dramatically. Um knowing you know what 20:55 equipment you're likely to be using, how often you're likely to procure and so on. So, it's all about minimizing. I 21:02 mean, things go wrong. It's it's impossible for a data center facility not to encounter problems. It's a, you 21:09 know, it's a it's a highly tuned engine. So, and like all highly tuned engines, 21:14 it's bound to go sideways from time to time. But the idea is you put it back on track in, you know, in the in the 21:21 shortest possible time. I would suggest that as these sensors are new that, you know, the understanding of what may go wrong may 21:26 be maybe a challenge. uh it may be only after you know certain maintenance cycles have gone through that a customer 21:32 gets clarity over what they might need to uh buy or stock on a any given basis. So um you know that we do appreciate 21:39 that that that may be a difficult challenge as that sensor comes from you know construction through to um through 21:44 to being on online and is operationally handed over. I think one of the the ways 21:50 in which RS looks to support that is through high quality data. So helping people understand what they're using and 21:56 how they're how often they're buying things to help inform that procurement policy and understanding of you which 22:02 spare parts I suppose they're buying and buying frequently. Um you know one of the challenges that we see within you 22:08 know across industry is customers having a lack of data uh and if um if people are buying on pecards or sporadically 22:14 you know using lots of different vendors um in lots of different procurement methods and it's not necessarily 22:20 controlled you can think about that data as a um you know it's lots of information but it's not particularly 22:26 insightful. You know the more that you can consolidate with suppliers that offer high quality data and management information the more informative that 22:33 data can be. Um also um transitioning from um preventive maintenance to 22:39 predictive maintenance has allowed u many of the operators out there to um 22:45 collect the right type of data I would say because you can collect data if you can't turn data into something 22:51 meaningful it becomes pointless and unless you can translate it into into 22:56 something that everybody can use it has no value. was with predictive maintenance now becoming more and more 23:02 popular. Now the operators have a chance to look for anomalies within their operations before you hit a major 23:09 failure. And so communicating that with the suppliers and understanding what could potentially fail in a month from 23:14 now on also goes towards um you know uh planning you know uh planning to reduce 23:22 your your your your failures. Great. Should we uh you know we've already kind of touched on this next 23:27 question actually. So um you know Phil and Richard and and Nav'i too if you want to come in here but um let's start 23:34 with you Phil. Um but yeah fragmented supply lists are a real common headache 23:39 in MRO. So how do operators consolidate spend and create a manageable supplier 23:44 base without losing resilience, technical confidence or flexibility? 23:50 And what works best when you're trying to reduce complex complexity without creating new risk? Yeah, sure. Uh I mean Richard kind of 23:57 touched on this in his um his previous answer where he was talking about the what makes a good supplier. Uh you know 24:03 he was talking about responsiveness. He was talking about support and commitment. He was talking about cap you know capability and knowledge and kind 24:10 of you know how in this instance kind of RS look to support their um you know their customers. And I think you know 24:17 from a procurement point of view it's you also want to be looking at the you know supplier preferencing model how a 24:23 supplier sees you um off the back of that if you are looking at being in the 24:29 core box where a supplier sees you as a good customer yeah good business long-term commitment then that 24:34 relationship and relationship is a word that's come up repeatedly through this discussion you know you can build a 24:41 really strong relationship where you've got you know a really good, smooth, uh, 24:47 productive, sustainable, reliable, um, supplier, you know, relationship with 24:52 your supplier. And that's a great place to be. You know, it gives you, uh, it's ease of management. much much better 24:58 that you have that kind of relationship rather than managing a frag fragmentated fragmented supplier base of multiple 25:05 suppliers doing bits and pieces you know and again you know Navi was talking about being forward looking and having 25:11 good data and planning for the future and making sure you know of supply you 25:16 know that's where you want to be it's absolutely where you want to be do not want to be kind of in any of the other 25:22 three boxes really where you're either being exploited you're seen as a nuisance or you're just there as 25:27 potential potential for the future. You want to be in that core box working really closely. And that goes back to the earlier question about, you know, 25:34 kind of avoidance of downtime. Full stop. Yeah, I agree with that, Phil. I think the key the key word there is like 25:40 leverage for me. Ultimately, if you can uh pull your spend with um more capable 25:46 suppliers, larger suppliers that can create value for you, ultimately you can leverage that that spend with that 25:52 supplier for for better commercials clearly, but also and I think important in the data center sector more so than 25:58 anything. Um you can leverage that for for for greater sort of service. Um and 26:04 um you know, you can ask for of more you can ask more of them if you like. um hold that supply to account more more 26:10 than uh than you can do if you're dealing with hundreds of suppliers that you might have spent you know transactionally with through the course 26:16 of the year. Yeah. Um you know that that supplier ultimately is not not going to be um bought into um you 26:24 holding you know I think that Phil used a sort of uh you know that box analogy 26:29 around you being a core supplier to them not a nuisance or a pain. Um well then you're looking for you're 26:35 looking for the innovation, you're looking for the creativity, you're looking for market knowledge, you're looking for suggestions and and constant 26:41 improvements. Yeah, it's that it's that equation that you suppliers ultimately um expect value 26:47 for money and distributors are looking for for money for the value they create and ultimately the more value that 26:54 you're putting their their way the um in terms of money monetary spend the more value they can create for you in terms 26:59 of that innovation that that attention that account management support etc. Um and ultimately in my mind you can do 27:05 that through uh you can consolidate without creating additional risk by looking at suppliers with broad-based capability or ones that have like an 27:12 omni omni channel approach. Ultimately it can trade with you in a in a variety of ways. Um and think about the way that 27:18 you consume product and look at offering a a solution to to match that that 27:23 process that procurement process. If you've got stuff that you're buying on an ad hoc basis uh we talked about 27:28 inventory and stock before you can't stock everything. That's why you have suppliers because your otherwise your 27:35 data center would be 98% full of stock rather than the server racks that ultimately create the value. So you do 27:41 need to have suppliers that can supply you product quickly. Um and for the stuff that you do use frequently and 27:48 regularly, you ultimately pouring time and people's people's energy into 27:53 managing repeated spend. So looking at things like vendor manage inventory where where it sort of pushes that uh 28:00 management and expectation of stock through to through to suppliers where they're better placed to do it is a more efficient use of engineers time 28:06 especially especially in this space where there's um um sort of skills gaps and and and sort of challenges with uh 28:13 the levels of engineers we've got in the industry. The the final thing I'd say is yeah Phil's touched on it there about 28:18 innovation. uh if you're dealing with broad-based suppliers are ultimately dealing with thousands of different suppliers themselves, the OEMs that 28:24 manufacture the kit that they distribute and I know for for us at RS we have really really strong relationships with 28:30 with key OEMs in the sector where we can be the conduit if you like to the innovation that they can bring uh to end 28:37 users in the data center space. Okay. So let's talk uh move on to talk a little bit about the uh downtime and its 28:44 relationship with procurement. So a question again uh for Phil and maybe 28:49 Richie you could come in too on this one. When organizations model the impact of unplanned outage which costs matter 28:56 most in the boardroom and how often do those costs actually show up properly in 29:01 procure procurement conversations? It's one thing to talk about unit price but it's quite another to capture the wider 29:08 commercial damage of downtime. Um I think the costs that um matter most for 29:14 uh an unplanned um outage are the unplanned unplanned costs. Um you can 29:20 kind of plan for things you know if things are likely to go wrong you kind of there should be some planning in there you know kind of uh to allow you 29:28 to mitigate for for things going wrong. But when things go seriously wrong and the unplanned unplanned things go go 29:34 wrong they're the things that really really hit uh kind of hit the bottom line. So I think you know a lot of this 29:41 is around storytelling and a lot of the storytelling is about 29:46 the in the importance of MRO because I think MRO you know kind of tends to sort 29:53 of sit at a slightly lower level than it should. Uh so it is about um you know 30:00 kind of the storytelling that goes into the boardroom to get the commitment and the focus uh to make sure that you've 30:06 got the resources. Uh we're talking about data, you know, everybody's got the greatest data. Nav was talking about 30:12 having good data. If it's not good, it's pointless. You know, you need to have that make sure you've got that good data. You need to have those good 30:18 relationships. You need to have a good supplier base that you can rely on. So that kind of planning I think is really 30:25 really important. I mean in an earlier conversation we talked about uh kind of you know what does [clears throat] a 30:31 mature approach to critical spares and mro look like you know and at that point 30:36 we're talking about having a spare strategy um that focuses on you know you 30:41 know rather than just a spare strategy it's about business impact it is about that business impact all the time so uh 30:49 in mature uh and you know strategic uh 30:54 situations Yeah, it does get talked about in procurement you where you're at that kind of higher level but quite 31:00 frequently um it doesn't you know and again you know Navidid was talking gave plenty of examples uh beforehand about 31:07 where things drop between the gaps and all sorts that's where things get board get dropped that's where there's an 31:14 unplanned unplanned cost and they're the nasty ones that really bite sure Phil I you know I'm I'm curious 31:21 actually just you know so from your perspective you know what level of understanding handy do you think there is of MRO in you know in in data centers 31:29 boardrooms but more generally across various industries um you know is it getting better you know are executive 31:36 members starting to understand it more or is there still very much an education job to do I always an education job um I think 31:44 there's you know kind of uh skills the kind of skills that are being focused on 31:50 this you know could improve I mean they're always improving but you know it's It's I'll kind of bow to to Nav's 31:57 sort of um uh view in a minute in terms of him being probably closer than I am, 32:02 but I think uh with the data centers um they're new, they're big, they're 32:08 expensive. So I think the the the focus is better uh and uh you know than it is 32:15 in some other industries uh just because of the the high risk uh you know kind of 32:20 the impact is absolutely immediate. you know, if it goes down, it's a problem and everybody knows about it really 32:25 quickly. But so I think because of that, it's better. There's more focus. But um Nabid 32:32 will probably either agree hopefully or disagree and tell me I'm wrong. No, absolutely. Phil, um spot on. Yeah. 32:38 I mean, in the boardroom, the the the costs of an outage go far beyond repair 32:44 costs. that's not what they're interested in hearing or even paying penalty um you know payouts which could 32:52 be quite costly. It's about reputational damage and it's about customers who have 32:58 zero tolerance quite rightly because of the nature of the infrastructure and losing those customers to a competitor 33:05 can be very expensive. Those are the subject matters that are um primarily 33:10 discussed at boardroom. um you know how how do we um ensure that you know we 33:16 don't hit these problems uh what plans have be in place because you know as as I um touched on earlier things do go 33:23 wrong and so when it goes wrong have you got a proper protocol in place if you 33:29 have a pro proper protocol in place you can reduce your damage if you haven't 33:35 every day that you're losing is quite costly yeah I think it's a really interesting thing because I think in this industry 33:41 more than any the the link between something that happens in a remote facility somewhere um you know a data 33:48 center ultimately the link between that and an end user a consumer is immediate. You know, we know that, you know, we're 33:54 operating on a on a call at the minute over the internet that's ultimately running through a through a data center. 33:59 Your banking transactions, social media, um email, all these things are so 34:04 interconnected and so dependent on these on these these facilities. Um the 34:10 reputational damage and and consumer end user consumer awareness of that is is probably more so than any other 34:16 industry. So it's you know it's unique in that in that fact in my in my view and that will get get the attention as 34:22 the other other guys have said you know in the boardroom that reputational damage is is probably more um more 34:28 prevalent in their minds than anything else but caused by something that could be quite inconsequential in terms of cost um so you know if a in the world of 34:36 MRO if your standard average order value is between2 to500 for a you know basket 34:42 of goods and something in that in that basket is is designed trying to go and fix an asset and repair something, get 34:48 something back online. And you're looking at downtime that isn't measured in the hours like it is in normal industry. It's measured in minutes and 34:55 could be costing, you know, recent white paper that RS produced around data centers. Um, it measures the cost 35:01 between 7 and£10,000 a minute. Um, most of the customers in industry talked to us about downtime costs in hours. So for 35:08 me that that that distinction is um is really really um paramount when we're 35:14 talking to customers in this space about how RS can create value for them in mitigating that and and and preventing 35:19 that from happening in the first place. Um but yeah ultimately that that risk could be you know what if a 2020 20 part 35:28 was causing £7,000 of downtime per minute. you know that that the the scales and the the comparisons are um 35:34 are crazy and the consumer expectation is um you know is increasing all the 35:40 time. You know, it's we're not expecting our service to drop from these providers and we're all using it every day. Um so 35:46 yeah, it's a it's a critical industry that um we know we're involved in supporting. Sure. So Richard, you know, just to kind 35:53 of follow on from that, you know, do you think we're moving away from, you know, 35:59 um, uh, data centers and, you know, you know, companies in general, actually 36:04 across the industry having a a lowest unit cost mindset towards valuing other 36:10 things. Uh, when it comes to MRO, I think it it probably should do. I mean, if it depends on who you're asking, I suppose. If you're asking um 36:17 someone in procurement or finance, they might have a different view to someone in operations and and engineering. Um that's something we often we often see. 36:24 Uh but in in the in the report that I mentioned before, the um speed of delivery, responsiveness, and 36:29 availability were all paramount in a in the end users, the customer's mind uh when looking to source goods over and 36:36 above price. Um, but I imagine if you're the financial director of of that facility, you know, you'd want that 36:42 price added back into that conversation quite quite quickly. Uh, but clearly if your KPIs are are geared by um and 36:49 measured in minutes and downtime um then you're going to want that part pretty damn quickly. So I think um yeah, there 36:57 probably is a an evolution and a maturity moving away from just putting cost in there and it is ultimately about 37:02 total cost. So, it's everything from looking to source the item, then the procurement process to actually go and 37:07 buy it, the lead time to get hold of it, and then the um yeah, the lag, if you like, of getting it back on the on the 37:13 asset on the on the piece of equipment that was that needed it. So, um we always encourage customers to think 37:19 about total cost rather than unit cost because there is a um you huge amount of time and effort that goes into in terms 37:25 of people's time. So engineers, procurement, finance, all going into sourcing goods irrespective of the the 37:31 cost in that industry of downtime. You add that back in in this instance in data centers and the the impact is 37:37 exponential. So um yeah, speed of delivery, responsiveness and availability are the key things that our 37:43 customers in this space are talking to us about. What makes those conversations land with boards, finance teams and operators in 37:50 your experiences? you know what what how do you persuade that tricky finance person who's going to be like we've got 37:57 to be completely cost focused here. We would take the view that you've got to have um all of those interested 38:02 parties, all those stakeholders in the room to to create a collective decision and make um some of what we do often is 38:08 facilitating those internal conversations within our customer base to make sure that the finance person understands the pressure the the the 38:14 engineer or the operations team are under and procurement understand what um what finance are looking for too. So 38:21 it's you know it's facilitating a cross party um multi- multifunctional 38:26 conversation if you like with those people to raise awareness of um you know KPIs that are set forth with all best 38:32 intentions and how they may impact and create some some behaviors that aren't that value adding to other departments and other colleagues. Um and ultimately 38:39 how we we look to to create value is through that through that type of partnership. You can only do that if you're um you know if you're working 38:46 with fewer vendors if you like and you've consolidated working in a more mature space to a point then you're 38:52 trying to extrapolate the value which they can create to alleviate some of those concerns across function. I agree totally. Um it's it's about the 39:00 overall value and the value isn't in cost savings. The value is in uh all the preventative stuff and the the added uh 39:07 bits and pieces that Richard's just eloquently listed. And you can't have, you know, it's about being strategically 39:13 aligned as a business across all functions and all departments. You can't have uh procurement uh targeted with 39:20 cost savings when you're trying to mitigate risk and everything elsewhere. that total cost and total value. 39:26 And there there was a um so backtracking a few years. Um 39:32 um you know the the costing exercise was was the main driver uh with some of 39:39 these projects and the finance team in most all cases were always keen to um 39:46 return as much cost savings as possible. But cost savings at what cost? Um and 39:52 they soon realized that saving 5% 10% even as much as 15% on a cheaper 39:58 component was um a counterproductive exercise. 40:03 And so the mindsets were changed in particular as we're moving towards 40:09 facilities that really cannot be switched off. You know they really have to be on and you know any downtime comes 40:18 uh at a very high cost. So, you know, getting the best price is no longer I 40:23 would say um certainly the clients I speak to are more leaning towards less 40:29 about transactional partners with a supplier, more about a strategic partnership with 40:36 the supplier and getting to know the supplier well uh and creating that 40:42 ecosystem that they can rely on. Well, therefore, so uh the cost is secondary 40:48 now up to obviously everything has to be budgeted, but it's no longer the main driver. And also, I mean, Rich, you know, we 40:55 Richard and I both mentioned innovation earlier. If you're hammering a supplier on cost, you're hardly likely to get uh 41:00 innovation and creativity and commitment and forward thinking. It's um you get 41:05 what you pay for. Okay. So, continuing to uh you know talk about cost issues, I I guess it's a 41:12 question for all of you. Yeah. How how can operators and suppliers structure agreements around service levels, 41:18 stocking, procurement, and inventory so that procurement's actively protecting 41:23 uptime rather than simply just managing spend? I mean, what what does a good model look like when you're trying to 41:29 balance resilience, cost, and responsiveness? I for me, if you if you were to start on 41:35 that, I think it's about understanding the consum consumption of the products that you're looking to buy and not treating everything with with the same 41:41 approach. So something that you buy frequently or has a high value of spend against it in the year or forms part of 41:48 your planned spend for that period should be treated very differently to something that you buy in an emergency 41:54 basis and just need to get hold of quickly. Um so focusing on that on the way that products are consumed for a 42:00 commercial lens is one thing and then also how you create value around that through the services that suppliers can 42:06 offer. So you know we've talked about the fact you can't stock everything. So if you understand that you buy something frequently and regularly, having 42:13 commercial agreements that we put in place, KPIs to measure that supplier for those core basket of goods, um you know, 42:20 talking to that supplier about services they might be able to offer to support the procurement of those goods, things 42:26 like VMI, vendor managed inventory should be, um different to the things 42:31 that you need to buy quickly in in an emergency, often underpinned through, you know, quick availability or easy 42:38 ordering processes. is in controlled ways. Um you know but in the same sort 42:43 of ecosystem if you like of of that supplier so you get great data and understand the nature of consumption. Um 42:50 but you know having proactive conversations with that supplier around you know master service agreements, KPIs, SLAs's etc. Um that feel 42:57 appropriate based on the nature of the product that you're buying. I think for me is a key key starting point. 43:02 I agree with that completely. I think data, you know, again, we've talked about data uh um and I think if you've 43:09 got that data and you understand the consumption um flow and you know the 43:15 consumption rates and what's you know what's needed when that kind of having that information and being able to sit 43:20 down and plan forward uh is is vital. So I think I agree with Richard. Sure. Okay. So um let's look into the 43:29 future now. Um there's a question for you Navidid really. Um obviously there's a big conversation going on about AI at 43:36 the moment and you know AIdriven demand is definitely changing the pressure on 43:41 data centers. So I wonder what does that mean for infrastructure power cooling 43:47 maintenance and day-to-day operational planning as as demand grows does that 43:53 make mro even more important or does it mean that it's kind of being eclipsed by 43:59 other things and is left behind by the bigger build convers I hear so many people talking about AI 44:05 AI is becoming it's become a regular regular conversation um and and quite 44:12 rightly Um yeah, absolutely. AI is completely rewriting the rules of our industry. There's there's no doubt about 44:18 that. Uh tradition uh traditional data centers handle standard cloud computing. 44:24 We all know about that. But um in the case of AI, they require massive GPU 44:31 compute clusters and you know they pack a lot of power, they generate a lot of 44:36 heat and so you're up against um uh you know a different ball game here and um 44:46 the facilities have to be designed in a different way from the ground up. Um you 44:51 can't just uh retrofit an existing facility that was built for cloud computing and convert it to AI. Um uh 45:00 you can't the structure is just not man enough to you know to take the weight to start with. Um we're moving rapidly from 45:07 standard air cooled. We all know that to liquid cooled for the obvious reasons because air isn't sufficient enough um 45:15 with you know to take care of the higher uh operating temperatures. Um and also 45:21 um if I'll go back a few years in the past um data centers had um scheduled 45:27 quite periods where engineers could go in and carry out service work. Mhm. Um you you can no longer just shut down 45:35 um uh parts of the data center facility to carry out lengthy lengthy work. And 45:40 so uh work has to be done on the fly. Um what does that mean? It means that um 45:47 you know um the maintenance windows are shrinking dramatically. We have to 45:53 service high stress equipment on the fly and all that goes towards raising operational stakes. So it's not just uh 46:01 so while it's moving and it's creating a lot of challenges uh we have to also put 46:07 in place um um revised framework to to take care of to take care of those those 46:13 um issues are going to be coming coming our way. Um, when legacy servers go down, it's an inconvenience. When an AI 46:20 training cluster goes down, you're talking about a multi-million pound 46:26 training run. Basically, it's it can cost a lot of money. Uh, and so, uh, 46:32 from a balance, I guess from the investor's perspective and looking at their balance sheet as they do, a 46:38 failure in an AI data center is a major financial event. That's sort of how they're seeing it. And I think 46:46 those are some of the reasons why a developer who is looking at building an AI facility for the purpose of AI 46:55 training for example. Um so high density racks lots to think about uh liquid 47:02 cooled facilities. They've got to really think about who they're going to be partnering with who's going to be 47:08 delivering that facility that's going to be operating the most efficient way. And that's where the due diligence processes 47:14 are getting longer in terms of qualifying a partner to you know to go 47:19 in and uh cater for um for the for these challenges. Um the other area that a lot 47:26 of people don't seem to touch on but it's absolutely key to efficient operation of a data center is BMS and 47:33 EPMS. So on the digital infra infrastructure side while you've got 47:39 many of the facets that feel uh uh um drop under the digital infrastructure um 47:45 framework uh one of those facets is BMS and APMS um you need specialist contractors that 47:52 really understand how to control the environment within the data center how