← 返回

2026.09.23 | TB | Power | Siemens Energy Grid Tech Backlog

原标题：Siemens Energy – Grid Tech – Could High-margin Backlog Conversion Expand Margins for Years Ahead?

#power #industrials

Third Bridge · 2026-09-23

Analyst-led • 10 Sep 2026

Companies: ENR DE, NG. GB, A004800 KR

Adam Middleton (AM) Director at EnergetIQ Advisory Ltd

Louis Knight (LK) — Third Bridge Sector Analyst

Agenda

Key Insights

Introduction

LK [00:00:01]: Welcome to Third Bridge's Interview entitled Siemens Energy – Grid Tech – Could High-margin Backlog Conversion Expand Margins for Years Ahead? My name is Louis Knight. I will be facilitating today's Interview with Mr Adam Middleton, former Managing Director at Siemens Energy in Belgium, who has also experience in the Netherlands as well as wider experience across Europe. Adam, thank you for joining, and welcome to today's Interview.

Before we begin, please state I agree or I disagree to this following sentence: You understand the definition of material non-public information and agree not to disclose any such information, or any other information which is confidential, during this Interview.

AM [00:00:50]: I agree.

LK [00:00:52]: Could you begin with a brief introduction to your background, relevant experience and various roles, please?

AM [00:01:03]: Good morning. My name is Adam Middleton. I've been working in energy and predominantly grid for the last 35 years in leadership positions. Until summer 2025, I had a Vice President role within Siemens Energy. During the period of five years, I was Vice President of Western Europe, covering executive responsibility for all activities of the company in France, Belgium, Switzerland, Netherlands and Luxembourg. Prior to that, I spent three years working with Siemens in the Energy Management division as Vice President for all of their activities around grid in French-speaking countries, so France, Belgium, Luxembourg, 17 countries in Africa. Prior to that, I spent three years with the EPC company, building substations, main customer was National Grid in the UK. Prior to that, some time with an Indian company, again, in T&D. I had 6.5 years working for General Electric, including three years running their global substation automation centre of excellence in Canada. The start of my career, I had 21 years working for Alstom, Alstom T&D and Alstom Grid, working both in turnkey projects, protection and control and also having responsibility for marketing and strategy for one of their divisions when based in France. Nice to speak with you this morning.

Questions

[00:02:45] Siemens Energy's Grid Tech has a current EUR 51bn backlog, and yearly order intake is currently around EUR 20bn. In what year does order intake peak, and at what level?

AM [00:03:08]: First of all, to say that that backlog will be a mix of compact transformers, switchgear, other equipment, also framework agreements. For example, you will buy transformers or the major supplier customers in Europe will buy their transformers on a framework. You'll have a multi-year product agreement. There will also be in there a package of grid services, turnkey projects and systems and solutions. Between that portfolio of different orders, you're going to have a schedule of somewhere between orders of several weeks duration and some of that are 25 years. Therefore, I think it is quite simplistic to actually say how long does it peak and at what level? At the moment, I don't think there's any sign of any peaking. Demand continues to outstrip supply. For the foreseeable future, there is no sign that this is reducing.

[00:04:14] Order intake must peak at some point. Even if it is in 2035, this is a cyclical industry. We have a huge amount of CAPEX approved by regulators in the RIIO [revenue = incentives + innovation + outputs]-T3 period. How far out do you think it might be?

AM [00:04:29]: Again, you look at the projections. The work that's going on at the moment is electrification. That's the fundamental change. You're looking to double the amount of infrastructure. It's taken the industry 70 years to put into place the electrical infrastructure that's there already. In the next 14 or 15 years, for example, the plan is to double it. It will take considerably longer than 14 or 15 years because of the rate of change in availability of equipment and availability of human resources. There are projections out to 2040 or 2050. An assumption that it will peak, I think you've got to understand what are those subparts in there. We always say there would be 10 different factors affecting growth today, any one of which is material. Clearly, the fastest-growing one today is data centres. You've got a mix of factors. You've got a fundamental electrification and doubling of the electrical infrastructure. There is certainly no sign of it peaking, certainly for at least the next 15 years, if not up to 2040, 2050.

[00:05:51] Sell-side analysts at banks with access to Siemens Energy Management have modelled the peak of order intake and sales around 2030. After that, it declines for 2-3 years through 2033. Do you think this is a conservative estimate?

AM [00:06:22]: If I may refer back to something we were touching upon a few moments ago, you've got to understand in a little bit more detail the mix of orders that are coming into certainly the three major players, of which we're talking about one today and where those are coming from. For example, Siemens Energy are very selective today in the products they're selling. They have a limited capacity, which they're growing, but you will pick and choose the customers you want to deal with. Therefore, the demand far outstrips supply. For example, in the UK, they are not on the transformer framework agreements with National Grid. These are public domain information. There is far greater demand than there is supply at the moment. That's why the lead times are going out. Therefore, even if there is a softening in the coming years at some point, there is still significant wider demand that's not being satisfied or not being accommodated by the three major players and the three major players are doing pretty much the same things.

[00:07:47] Analysts expect Grid Tech to see sales of around EUR 24bn in 2030 and then that tails off to around EUR 16bn in 2033. This seems to be a considerable decline post 2030. Do you think they're not appreciating the long-term demand that could actually come to fruition?

AM [00:08:19]: If you go back and have a look at the way the market has developed, there's always been a concern that the bow wave of increasing workload associated with electrification would not happen. This has always been a concern, especially further down the supply chain is, can you believe the numbers? The projections and the numbers are still there and still significant is, how much do you believe? Then, as I say, you need to peel into those numbers a little bit. For example, if you're looking at something major like the high-voltage DC connections to offshore wind farms in the North Sea, there are plans for 300GW of offshore wind. Again, each of the 2GW connections and certain of the utilities have ordered forward significant volumes up to 2031, if I remember rightly.

You can see, at the moment, there is adequate visibility on those major projects, which are multiple billions. After 2031, you can see those really huge projects calming down unless there is further growth in the connection of the offshore wind. It's really important that you understand the sub-elements to that. I think the second one, of course, is how quickly data centre development progresses. The major constraint at the moment is being able to plug them in. You may be able to build a data centre and you may be able to build the high-voltage connection, but you can't… The grids are inadequately sized to be able to take the high-voltage connections. It's really important to get down a level below and understand where those major factors are.

[00:10:31] We have the upcoming RIIO-T3 in the UK, and RIIO-T4 will be in the 2030s. Beyond that, there could still be more expansion CAPEX. When you think about the equivalent investments and regulatory periods in Europe and the US, what will fill that huge gap? US President Trump has completely wiped out offshore wind pipelines in the US. What might support the huge level of growth and, therefore, not see a bit of a tail-off in the 2030s for Grid Tech?

AM [00:11:15]: (Audio distorts 11.17) about Siemens Energy, if I may. There's two or three things here. First of all, the American President, unless there's some major change in governance, has another 2.5 years to complete his second tenure. After that, I think most commentators are assuming there's going to be somebody else in the White House. You're going to need every source of energy possible to electrify the world. You're going to need nuclear, you're going to need wind, you're going to need solar, you're going to need local generation. You're going to need small modular reactors. You're going to need everything. A policy that pushes aside wind in North America, I certainly agree with the commentators who say it's a time out rather than a complete cessation. I'm absolutely assuming the offshore wind industry, certainly on the East Coast, will come back, maybe it takes a few years to come back, but I'm assuming it will come back once there's different incumbents in the White House. The demand for electrons is only going to go up. I mean, that is crystal clear. The supply of electrons at the moment is in a (audio distorts 12.51). We have to find ways as an industry to make more electrons available at a lower accessible cost, and that's a huge challenge. The only visibility is upwards at the moment.

[00:13:09] During the 2030s, what might growth from LATAM, the Middle East, Africa, Asia, India and SMRs [small modular reactors], as that revolution comes through, do to support Grid Tech's growth? Do you think it could potentially sustain EUR 24bn revenue per year through the 2030s? Is that too aggressive?

AM [00:13:37]: No, no, forgive me. Let's go back to fundamentals here. What line of business is Siemens Energy Grid Technologies in? It plugs in generation. Fundamentally, that's what they do. Irrespective of the generation, you're still going to need technologies and network technologies to plug in and provide stable, safe power flows. If you build a wind farm, you're going to need to plug it in, you're going to need to ship bulk amounts of power. If you're having far greater implantation of renewable energy, whether it's solar locally, you're still going to need to reinforce networks everywhere because the networks over the last 70 years are just not designed for the energy demand and the topology of the future. There's got to be ongoing significant increase in grid technology and electrification.

As I said, there are probably 10 different factors that are driving growth within grid. We've talked a lot about new generation and new wind farms, and we'll come back to SMRs in a minute. Also, remember, there's a lot of this infrastructure that's already there is 50 years old. There are transformers and power cables, for example, that are life expired. There is an absolute requirement to replace that basic infrastructure that's already there and service it and keep it in a safe and operational condition. It's not only building new and adding new, you have to manage and replace and upgrade what's already there, which is another huge piece of work and probably even more complicated than adding new because you've got to understand how it knits together and keep the lights on in the meantime.

Bear in mind also the political sensitivity. I'm sure you'll be acutely aware of what happened in 2003 with blackouts. We are a lifetime away from the situation and the impact of reducing reliability and interruptions would have major, major implications as would something like cybersecurity, which is… You can imagine, again, the impact of significant attacks and there are tens of attacks per day on the electrical power industry, tens of attacks. If I may, just going back to the thing about SMRs, SMR is, again, it's just another generation technology. We'll have a different connection requirement. We're probably still 10 years away from getting commercialised SMRs. You may see the first ones around 2030, you may seriously get into commercialisation probably 2035. We're still a little bit of time away, but the demand is clearly there.

[00:17:12] Given everything we've discussed, to what extent do you think the EUR 25bn annual revenue can be sustained through the 2030s for Grid Tech as opposed to it reaching EUR 24bn in 2030 and then dropping to EUR 16bn through 2033?

AM [00:17:28]: As I suggested, what you need to do is to look at these subsets and then run some scenarios for each of the subsets. We talked a little bit earlier about the…

LK [00:17:42]: What is your gut feeling based on your experience, do you think that is conservative or aggressive to expect it to drop off significantly post 2030?

AM [00:17:51]: I would be very surprised to see it significantly drop off.

[00:17:56] A significant portion of growth expectations hinge on load growth from AI data centres and electrification, especially in the US. How much of this pipeline consists of binding equipment reservations vs speculative project proposals and might not materialise, given grid infrastructure expansion can frequently hit regulatory bottlenecks, environmental delays and local opposition?

AM [00:18:22]: You don't book orders unless they're firm. It's not a situation. The equipments we're talking about are manufactured to order. You are not going to commit your manufacturing capacity to something that's not firm. You may have a framework agreement for five or 10 years with a projected capacity. The capacities are only going up at the moment.

LK [00:18:56]: Could you define what a firm commitment is?

AM [00:18:59]: Order with money. Let's use a precise example. When you buy a transformer today, what you do is you pay for a production slot. You're paying, I don't know what the number is, EUR 0.25m to firm your commitment to a production slot. Then on top of that, you will have a calculation mechanism to work out how much it's going to cost at that point of time in the future. You're putting money upfront, a serious amount of cash upfront to take that production slot. You don't do that if it's not a firm commitment.

LK [00:19:40]: If a hyperscaler pays 10-25% reservation fee for a slot, it could potentially not pay for the rest of it if it really has to cancel the CAPEX for the rest of the project.

AM [00:20:03]: Yes. Then there's still a queue of people wanting a transformer. At the moment, demand is so much greater than supply that the suppliers can pick and choose who they work with. Yes, the hyperscalers and the constructors of data centres are the fastest-growing segment, clearly.

[00:20:28] To what extent could we potentially see delays or cancellations on a significant chunk of the EUR 51bn order backlog?

AM [00:20:40]: To say there's not a risk is incorrect because what you need to do is have a look at the impact of DeepSeek and an alternative technology or an alternative algorithm were to show that you needed a lower volume of data centres. Again, the growth is such, it's massive or even more massive. There is still going to be significant growth even if there will be some adjustment going forward, that's for sure, because of the technology of the AI that will change. Now, whether we need 1,000 data centres or 10,000 data centres, I don't know, but it's more than the 10 we've got at the moment. You've got significant growth. That growth will be adjusted and changed over time as the technology changes. I would suggest it's a technology-related issue. Whatever happens, you again need to plug them in.

[00:21:45] Would you expect the Grid Tech order backlog of EUR 51bn to be one-third TSO [transmission system operator], one-third offshore wind and one-third data centre connection?

AM [00:21:58]: I wouldn't like to estimate that.

LK [00:22:08]: Broadly as a rule of thumb in the right direction or…

AM [00:22:12]: Siemens Energy Grid historically has had the majority of its business from transmission system operators. That's where the bulk of their business has been. They've done some industrial business, but that backed off to some extent and preference coming for data centres. I wouldn't like to say what the mix is, but I would say that the significant chunk of this is still with transmission system operators.

[00:22:47] Could you estimate the order backlog split across TSO, offshore wind and data centre connections for the wider HVDC [high-voltage direct current] industry, thinking about the Big Three – Siemens, Hitachi Energy and GE Vernova?

AM [00:23:13]: Okay. Let's go back and understand what you use HVDC for. HVDC is for shipping bulk quantities of power over long distances. Historically, before offshore wind grew, you would plug together different parts of power network where you wanted to ship volumes of energy. For example, in India, where you have an unstable grid, you can plug together the eastern region and the northern region through a DC link. One, you can ship a lot of energy through a relatively restrained transmission channel. Secondly, those two pieces of network don't have to be synchronised, which normally is quite a difficult job, especially in places like India. Once you added in offshore wind, so the whole idea with offshore wind, so typically, you would use high-voltage DC when you're more than 75km offshore, if there's an economic model you can use. Rule of thumb is anything over 75km, you would be looking to use DC, it becomes economic because the investment cost at the beginning is that much more significant.

It depends on that mix between, let's say, the core HVDC business and markets and then the growth in offshore wind connections. At the moment, the constraints are really around capacity. For example, you may have gone for a factor of 10 change between doing one project per year for each of the major players maybe six or seven years ago or 10 years ago and where each of them is now probably doing 10x that. The major constraint here is capacity and manpower and knowledge. Then that is pushing out deliveries and also then there's the partners building the mechanical infrastructure, the jackets and then the module that goes offshore.

[00:25:45] It seems the data centre pullback is the biggest potential risk to order intake in the long term. Could you estimate the share of global order intake across the Big Three that is purely DC [direct current]? Is it 5% or 50%?

AM [00:26:03]: It's not 50%. I would say it is… You need to look at the major projects, and these things are lumpy if they're EUR 2bn a piece. That's a significant part. I mean, you're quoting your numbers if they're looking at orders of EUR 20bn, you may have somewhere between EUR 2bn and EUR 5bn on DC. As I say, these things are inherently lumpy. There's been a recent, last two or three years, very significant set of orders put out by TenneT. I mean, they spent, I think, EUR 23bn or EUR 24bn predominantly between the three major players. That obviously will skew the numbers. You have to look in more detail at where those big orders are coming from in order…

LK [00:27:06]: If we average that out over 3-5 years, it seems to be 15-20%, then, therefore, is about a reasonable estimate for data centres and hyperscalers…

AM [00:27:16]: As I say, you need to look at where those major projects are, but it's that sort of order of magnitude. It's not 50%.

[00:27:25] Grid Tech announced multibillion-dollar investments to ramp up global transformer and switchgear manufacturing capacity, aiming to add 50% by 2030 from a 2026 base. To what extent could this aggressive CAPEX plan be executed without cost overruns or margin dilution?

AM [00:27:48]: I think you've got to have a look where historically, where you have problems if you're running this sort of business, it's not in the products where you have the problems. It's in the projects. If you're ramping up transformer production, you've got good processes that are well-honed and well-executed, you're building a new factory, you're recruiting new people. You're going to have a ramp-up period to get to a stable situation. That is not going to cause you major cost overruns beyond normal operational issues. Once you've got your supply chain quality sorted out, the major problems tend to be in the big projects. Things like the HVDC or the major substation projects where you have far greater complexity and also far more propensity for change.

[00:28:52] To what extent is a 50% capacity increase to 2030 a good indicator of scaling of the rest of the Grid Tech business?

AM [00:29:03]: I think it goes back to what business is Siemens Energy Grid Technologies in. It connects generation to power networks. As a starting point, you need transformers and you need switchgear. Those are the fundamental technologies. As a starting point, as a first objectionable proxy, that's not an unreasonable place to start. Then you've got to dial out other factors like what percentage of your business is actually service-related. As a starting point, it's not a bad proxy.

[00:29:43] Can Grid Tech manage supply chain bottlenecks, secure raw materials such as copper and grain-oriented electrical steel and ramp up production quickly enough to meet the major contractual delivery dates?

AM [00:29:57]: Everybody has got the same problem. It's the same fundamental issues. If you go back maybe eight years pre-COVID, if you were signalling there's going to be a major increase in demand, certainly, the tier 2 and tier 3 suppliers who are critical here, just didn't believe it. They've been burned too many times and they would not have invested. You hear the end user demands and you take it with a pinch of salt, and they just didn't invest. Now, the demand is there and there's a lot of work actually going on in the industry trying to make sure that tier 2 and tier 3 suppliers are actually scaling up and have the confidence and have the necessary resources to be able to invest. That's the critical bottleneck. I would add…

LK [00:30:54]: The suppliers, in particular, are you referring mainly to the cabling players and that being the number one constraining factor, ie, Nexans, Prysmian…

AM [00:31:03]: No, no. Forgive me. We're talking about transformers and switchgear. For transformers, you will buy some of the tanks that the transformers sits in or the fundamental coils are inside the device. You won't make them all yourself or even if you make up any of them yourself these days. Also the skills around and training the people to wind coils, it is a very, very manual activity. These things are not done on an automated production line with robots. These are handcrafted individual products, literally handcrafted. One of the big problems in transformers at the moment is there are relatively few people with these skills and they're not being trained. Let's take an example in the UK alone, there is only one factory with the skills. In order to ramp up, it isn't necessarily the issue about the plus 50% is you've got to have the tier 2 and tier 3 suppliers on board and with the investments.

Secondly, and most importantly, is the expert human resources. Have you got those skills in the volumes necessary? The final point is you need to understand what is the percentage of the people who work in this industry who are going to retire in the next 7-10 years. It's very significant. I would really recommend researching that in some detail. For me, one of the absolutely critical questions is, are you going to have enough key know-how and expertise and competencies in order to be able to undertake this work if half of your employees are going to retire? What are you doing about it? For me, that's the biggest issue of all.

[00:33:10] Is it close to impossible to automate the coiling for transformers to the level of automation achieved in the automotive OEM industry factories with robotics, automated tooling and soon, humanoid robots?

AM [00:33:27]: The straight answer is each power transformer is unique. It is a project in itself. The coils are made by hand with insulation, which is paper and pressboard and copper wire and people with hammers manipulating copper wire. This is artisan. There are certain parts of it that absolutely you can automate production of the laminations, the laminated steel that goes into the cores, absolutely. The coil winding is a critical skill. You can have it manipulated such that you're minimising the amount of actual handwork being done, but there is still a very significant amount of handwork being done. The other thing is… Sorry, I mentioned those things like transformer tanks. Today, there's only one manufacturer I'm aware of that automates transformer tank manufacturing.

LK [00:34:43]: Why is it not able to be achieved through automation or machine handling as opposed to craftsmanship?

AM [00:34:52]: Because they're all different. When you design a transformer, you optimise the performance through specific design of the windings. Therefore…

[00:35:05] What does the inability to automate large power transformer manufacturing reflect in terms of the bespoke nature of the project? Is it the equipment size and how it fits into the substation, voltage requirements or something more digital?

AM [00:35:22]: Inside a transformer, you have a number of copper windings placed one over the top of another on a set of normally three laminated cores. These things are massive. It can be three or four metres tall. The windings themselves can be one metre in diameter and three or four metres long and have multiple sets to them. These things are predominantly made by hand. The smaller the transformer, the more automation you can do and you can wind some of these, the more standard products. Then if I may, probably the other word is standardisation. There is greater pressure on standardisation today quite rightly and less preferential engineering. The bigger the transformer, the more complicated the transformer you have, the more bespoke it needs to be. Therefore, the more handcrafted, more artisan approach that you need to actually make these things.

[00:36:35] Ramping up manufacturing and system commissioning globally requires thousands of specialist engineers and technicians. To what extent could labour shortages limit revenue conversion and delivery capability?

AM [00:36:50]: As I alluded to a few minutes ago, access to competencies and competency shortage is critical. In an industry, on average, every 5-10 people who are leaving, there's one joining. You have a growth in demand and a reduction in supply. Now, this is going to be a growing situation. Train a commissioning engineer, you touched on a commissioning engineer, you're talking 10-plus years. This is not a five-minute job. If you have five years of fundamental education and then five years of experience to be able to commission these things. I would say the key bottleneck above everything is competencies and competency shortage. That's why you have a war on talent at the moment between the major players, fishing in a very small pool of people. That's why the main players go out to India, for example, to get expertise. It is a global challenge.

[00:38:10] To what extent would standardisation change the dynamics of the bottlenecks? Will customers always need an extremely high level of bespoke customisation?

AM [00:38:25]: The higher the voltage, the greater the bespoke nature. Remember, you've got 70 years of installed base. It's not as if you're starting with a vanilla canvas and you start and design everything to be standard. That's not the case.

LK [00:38:41]: So you've got to match the voltage ratios and oil configurations to the rest of what's around the grid and substation?

AM [00:38:50]: You have a grid and a topology that already exists and a dynamic and changing demand. Therefore, you're having to mix and match to make it fit. It's not as if you've got a whole greenfield or a vanilla country and you're building it from scratch, you're modifying, extending, enhancing, replacing a lot of what already exists. Then you're bolting on the side… An offshore wind farm is bolted on the side and you have to provide the infrastructure. There was never the infrastructure down the East Coast of the UK because there was never power generation on the East Coast of the UK, and now it's having to be created. Historically, you go back 30 years or 40 years, power generation was the coal fields in the Midlands and Yorkshire. That's all gone. You actually need the infrastructure around the nuclear power stations and offshore wind because that's where the major generation centres are and the growth of distributed generation with solar, for example. The topology of the grid has changed completely, and therefore, you need to be investing in different ways to connect this both distributed generation and the new major assets.

[00:40:18] What is the bigger bottleneck for the industry's growth? Is it transformer coil craftsmen and commissioners, or is it the copper cable supply?

AM [00:40:32]: I think you've got two parts of a bigger jigsaw. You're going to need power cable, in a parallel industry, things like oil and gas. One of the big problems at the moment is the auxiliary connections to offshore rigs are the things that fail first. The time to actually get cable at the moment is very significant. Yes, I mean, the capacity for power cable, especially the very high-capacity power cables, demand is very, very significant from the major players.

LK [00:41:13]: If you had to say one, which one would be the greater supply constraint?

AM [00:41:20]: I think there isn't one or the other. It's both. It has a different timeline and a different requirement. You can't exist without either. Forgive me, it's a little bit simplistic to say, well, do you want power cable or do you want windings for transformers? You need both.

[00:41:46] How are key players dealing with the potential loss of know-how from the retiring workforce?

AM [00:41:54]: There are ways to deal with this, and there are organisations that are active in, first of all, extending the careers of experienced people. People like ScottishPower are quite renowned for extending people's tenure with inside the company with flexible employment contracts. Perhaps when colleagues get to 60, they don't want to work 60 hours a week anymore, maybe they want to work a different work pattern. They're quite innovative in looking for solutions. National Grid are well-publicised about looking for mid-career people to retrain. It's not just we need more apprentices. ScottishPower, for example, again, are very good at attracting young talent and training them in an excellent way. Grid also focus on the same mid-career retraining. It's a continuous process, and this has to be ramped up in order to get the basic manpower requirements satisfied. I think also for the suppliers as well for particular technical skills, there are activities ongoing about capturing that knowledge using AI. Schneider Electric are quite well-published on the subject. Siemens Energy are also working on these sorts of tools and techniques. You don't lose the knowledge.

I think what organisations are going to have to do is come to terms with all of the pieces of this jigsaw. It is a complicated jigsaw. First of all, is understanding the profile of the employees you've got. Also, the different countries have different employment laws. For example, in the UK, you can't ask me when I'm retiring. It's illegal. Then there are different ways to actually understand what are my perspectives going forward in terms of what I want to do after the age of 60. There are ways and means to work around this. I think the final thing, there's some basics, you keep your employees healthy, and it sounds a little bit trite. There's a lot of basic support mechanisms have been reduced over the last few years as competition has become more and more difficult. There has to be a pretty strong reset and understanding just how critical the key knowledge workers are and then how you best manage.

[00:44:54] Grid Tech is targeting high-teens revenue growth through 2028. To what extent is this a conservative vs aggressive estimate?

AM [00:45:03]: I think you need to look on the subsegments between transmission, distribution, industrial, data centres. I think you'll have a look at the mix and what the projection will be for each of those subsegments over that period of time. Then you need to knit in their aspiration to grow the service business as well, which is well-publicised. You'd need to look at that mix.

[00:45:32] Isn't the sales growth of data centres actually constrained by HVDC? Therefore, would the growth rates be limited by the things we've discussed in terms of capacity of craftsmen and commissioners in cable copper?

AM [00:45:52]: If I may, you just said linking HVDC and data centres. Data centres are not connected with HVDC. That's different parts of the business. There may be some skills overlap, absolutely, and you're using substations in both cases to connect. I wouldn't necessarily connect those two factors. I don't think it is a direct…

[00:46:18] Would you more likely correlate HVDC growth to TSO CAPEX across Europe, the US, etc?

AM [00:46:30]: I think you've got to go back to fundamental demand for each of the subsegments. If you listen to the major TSOs at the conference, you listen to people like Dominion, Virginia, who've got the biggest installed base and the biggest growth of data centres on the planet. Their projections are just more and upwards and they're looking at standardisation, they're looking at teaming with other utilities in the region as a way to increase the capacity they've got available. Again, looking for different ways to do these things. I think the industry has to work together even more profoundly to be able to address some of these growth challenges.

LK [00:47:24]: It would be better and more accurate to go by subsector and look at data centres, TSO, various HVDC projects. On the face of it, do you think the high-teens revenue growth for Grid Tech is conservative?

AM [00:47:41]: You need to burn through the EUR 51bn. At some point, you're going to have to deliver that. You're going to have significant revenue growth associated with the length and duration of those contracts. It's quite easy to see there will be a significant growth and backlog is there and also new orders.

[00:48:07] Do you think revenue of EUR 24bn by 2030 is conservative or aggressive?

AM [00:48:14]: I think you make an estimate of that EUR 51bn and do some linear interpolation and come up with a number, if they're going to keep on schedule with the delivery commitments.

LK [00:48:26]: What is your personal gut feeling on that?

AM [00:48:29]: I wouldn't be surprised at all to see them at EUR 25bn.

[00:48:34] What are the rough industry average margins as percentage ranges for grid products vs grid solutions and grid services?

AM [00:48:47]: You're in a very, very special market situation at the moment where demand far outstrips supply. I'm sure colleagues have studied the numbers that are public domain published every quarter by Siemens Energy. There is an overview of that. Historically, you would make your money on the products and on the…

[00:49:18] What gross margins would grid products typically have? Would it be 30-50%?

AM [00:49:24]: Even more in some cases. I don't know, and that's obviously not public domain. I wouldn't want to project. Again, it depends on the different technologies. What you do know in terms of net margins, overall, they're very rapidly approaching 20% EBITDA. In that mix, you're going to have service business, which is a higher margin historically. You're going to have projects which are historically single-digit EBITDA. Again, you wouldn't always make your money on the products. That's how it's always been done.

LK [00:50:12]: Could you rank grid products vs solutions vs services?

AM [00:50:17]: Services would be highest and then products, and then projects.

[00:50:26] How would margins generally rank or compare across greenfield, brownfield, offshore, TSO and data centres? Could you potentially rank them?

AM [00:50:42]: I wouldn't like to say because I think it's skewed significantly recently because of the demand outstripping supply. It depends… When you're looking at… Let's take products would be, I think, fairly consistent across all of these domains. It also will depend on how much historic data the customers have got in terms of benchmarking, although certainly margins on products for TSOs has been clearly driven upwards by, again, listening to the TSOs complaining about the way prices have rocketed upwards. That's not just down to material costs increasing and labour costs increasing. On projects, certainly, risk is being more adequately covered with margin these days. I think it's maybe very generic to split greenfield, brownfield, offshore and onshore. I think you need to look at risk. Certainly, offshore risk is very significant. Data centres is completely different because your buyers are very, very different. I wouldn't like to say. Certainly, it's not public domain information as far as I know.

[00:52:13] High demand has granted Siemens Energy significant pricing power, driving high-margin contracts. How sustainable and expandable are these elevated margins once global manufacturing capacity scales up?

AM [00:52:28]: I think there's two or three things. First of all, it's a complicated question. I sat with a chief procurement officer yesterday of a utility, and there is acute sensitivity and understanding that the costs have gone up and also the reported margins have gone up very significantly. I think the second factor is the fear of utilities using products and systems from China. I think there's only one example, for example, of HVDC in Europe where Chinese technology is used because in very simple terms, you don't want critical national infrastructure and HVDC is critical national infrastructure for shipping bulk energy. You would have deep concern about putting control of that transmission channel in the hands potentially of an unfriendly country administration going forward. There is deep concern over buying any technology, which is going into a critical area from China, especially things like control systems. There's absolute sensitivity around that.

[00:54:02] Do you think that 20% EBITDA margins for the Grid Tech business and other peers in the industry are sustainable?

AM [00:54:16]: I think it's extremely sensitive. How long you can sustain it at that level, I would be surprised to see it go higher because of the absolute sensitivity with the customers because they certainly feel that they are being badly treated. Again, that's very open information that they feel that the suppliers are making very, very substantial returns. Yes, at the same time, the suppliers need to be investing, and then that would be part of the argument. Would I be surprised then to see a little bit of softening of that? No, I wouldn't be surprised because certainly, the pressure from the utilities will be such as to, at what point is this excessive?

[00:55:10] How successful have the Chinese HVDC projects been in Europe?

AM [00:55:19]: I don't know the technical details of that. I'll be quite honest. I know that I have some understanding of the deployment, but a first-off project with first-off technology, there's always going to be issues.

LK [00:55:40]: Is it in Portugal?

AM [00:55:43]: No, it's in north of Germany.

LK [00:55:47]: With TenneT?

AM [00:55:49]: Yes.

[00:55:51] I assume the HVDC contract to a China-based supplier is almost a leveraging tactic by TenneT to demonstrate that there are other options as opposed to relying on Siemens, GE and Hitachi. Would you agree?

AM [00:56:11]: At the moment, there are three suppliers. There may well be a fourth and fifth supplier. A fourth supplier, there was one of the Korean manufacturers, Hyosung, was demonstrating HVDC package, and they've got a first project they're deploying in Korea, a 2GW project. In the next 3-4 years, there will be another legitimate supplier, which is not Chinese, also on the radar. Now, I'm sure there's going to be others as well. Again, it's also political. Where do you trust?

LK [00:56:58]: The political pushback from China being a less friendly administration is more understandable than Korea. Do you think there will be any political pushback from the Koreans who have been decades back as good partners to Europe?

AM [00:57:15]: I'm not aware of any.

[00:57:20] What is the limitation or ceiling for Hyosung's expansion in Europe?

AM [00:57:27]: They've just built a new factory for HVDC at Changwon in South Korea, a whole new factory.

[00:57:36] What do you expect Hyosung's market share to look like in 3-5 years?

AM [00:57:44]: Very small because it will take some little time for things like certification and homologation. These technologies bearing in mind, they ship so much energy that it will take a period of time.

[00:58:02] How long will it be until Hyosung could have an impact on pricing and margins for incumbent HVDC project players?

AM [00:58:10]: I wouldn't be surprised if it's 7-10 years.

LK [00:58:15]: Not today? If orders are being made today and there's a factory involved, if the lead time is 3-5 years, could the utilities use…

AM [00:58:26]: No. Forgive me, you're not going to plug in new technology unless you've seen, first of all, some initial projects. The first project they have, I believe, is going to be with KEPCO in South Korea. That needs to be built, commissioned and put into place. You're looking at 2028 at the beginning, then you've got a process of certification may well take you to 2030 or 2031. You're at least seven years away.

[00:58:57] Is the duopolistic structure of Hitachi, GE and Siemens going to be undisturbed until 2030-31?

AM [00:59:08]: I don't see any really major changes. No.

LK [00:59:12]: Therefore, will it have any impact on the sustainable EBITDA margin of almost 20% until the Koreans scale up? Do you think the Chinese presence is enough to add significant price pressure?

AM [00:59:26]: The Chinese presence and the threat of going to alternative suppliers will maybe bring some stabilisation. I think the continued growth of margin, there is acute sensitivity at the moment. I would be surprised for it to go significantly higher.

[00:59:50] The Chinese have now broken into Europe with the north Germany project. Why would there not be further projects?

AM [00:59:59]: Political pressure.

LK [01:00:01]: There was political pressure a number of years ago when it was signed. Is that correct?

AM [01:00:07]: Absolutely.

LK [01:00:09]: If they've done it once, why not…

AM [01:00:11]: Forgive me, I don't have the details, but I know there is significant sensitivity and growing sensitivity as sovereign security becomes a bigger and bigger question.

[01:00:28] What do you expect a concrete sustainable EBITDA margin to be, given what we're discussing around competition?

AM [01:00:39]: I think in a bandwidth of 15-20%.

[01:00:44] Is there anything else that you would like to mention that is important, and perhaps investors don't understand or appreciate in the industry?

AM [01:00:57]: I think we've talked a lot about people and the potential wave of people leaving the industry and the amount of time it takes to educate and train. I think they also need to have a look at things like the administrative delays around regulation. The whole view that the regulatory models are based on historic situations, not on future requirements. The political decision-making to change that is a major constraint.

LK [01:01:33]: How would that appear in results for companies such as Siemens? Is this an order intake problem, or a milestone and delivery problem?

AM [01:01:42]: Slow decisions on projects.

LK [01:01:46]: It's future order intake as opposed to current order backlog?

AM [01:01:50]: Yes, correct.

LK [01:01:55]: I'll close the Interview there. Thank you so much for your really invaluable insights. It's been really great speaking to you, and I hope to stay in touch. For now, I'll say thank you and close the Interview.

AM [01:02:07]: My pleasure. Thank you.

Transcription ends at 01:02:10 of the recorded material