← 返回

2026.09.28 | TB | APD | Air Products Supply Chain & Capital

原标题：Air Products – Continued Supply Chain Headwinds & Capital Allocation Outlook

#industrials #chemicals

Third Bridge · 2026-09-28

Analyst-led • 22 Sep 2026

Companies: AI FR, APD US, LIN US, YAR NO, 2330 TW, GAZP RU, 000322221 BE

Dan Baciu (DB) Independent Consultant

Liam Dorsey (LD) — Third Bridge Sector Analyst

Agenda

Key Insights

Introduction

LD [00:00:02]: Welcome to Third Bridge's Interview entitled Air Products – Continued Supply Chain Headwinds & Capital Allocation Outlook. I'm Liam Dorsey, and I'm facilitating today's Interview. I'm pleased to welcome Mr Dan Baciu, former Director at Linde.

Before we get started with today's Interview, would you please state I agree or I disagree to the following statement: 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.

DB [00:00:25]: I agree.

LD [00:00:26]: Would you mind beginning with a brief introduction to your background and your time at Linde or in the rest of the industry?

DB [00:00:37]: Basically, you probably searched on my background on LinkedIn. I've been in industrial gases for over 35 years, with Linde for the last 25 years of them and with BOC, British Oxygen, in the United States for the first 10 years of my career. I retired from Linde in beginning of July last year. My last job with Linde was Senior Director responsible for helium in EMEA. Prior to that, I was the Head of Air Gases Tonnage for a number of years, and that's basically the business that's involved in large investments and supplying customers with industrial gases via pipelines. What else? Back in my BOC time, I was Manager of Application Development. Once I graduated from business school at NYU, for about three years, I was Manager of Business Planning.

Questions

[00:01:52] Could you discuss the volume disruption in helium due to the Strait of Hormuz closure? What is the historical importance of that region for helium? How much volume typically flows out of that region?

DB [00:02:22]: QatarEnergy used to be running three major helium facilities, Qatar 1, Qatar 2 and Qatar 3. Qatar 1 started, I think it was about 2010. It's 600 million cubic feet. I'm going to use the… We measure helium in standard cubic feet, which is the American measurement. I'll give you some numbers here. Qatar 1 is 600 million cubic feet nameplate. Then came Qatar 2 in 2015, I believe, and that is 1.2 billion cubic feet. Then came Qatar 3, and that's another 400-450 million cubic feet. Total nameplate in Qatar is about 2.2 billion cubic feet. The worldwide nameplate, it depends how you calculate it, but my view of the worldwide nameplate is about 6.5 billion. Qatar has been running very well, very reliably, all of them, Qatar 1, Qatar 2 and Qatar 3. Now, if you argue that because of the situation in the Middle East, Qatar is producing zero, that means about 33% of the worldwide supply is gone. Now, our understanding is that they are not producing zero. Out of that 2.2 billion cubic feet, we understand that they are producing about 100 million, which they can export through the land route to the Port of Jeddah in Saudi Arabia and then they take it up to Suez and bring it over to Europe.

Remember, from a nameplate of 2.2 billion cubic feet, they are only making now about 100 million of it, and that barely dents the amount of helium that is actually missing from the global supply. Anything specific you have on this? Because I'm moving from 100 million to billion, but the idea is very simple. The worldwide supply is about 6.5 billion. I don't take into account the Russian part of it because that's still not very clear if and when it will be operational. If you think of the supply of 6.5 billion, the amount that's missing from Qatar is anywhere between 2 billion and 2.2 billion, which means that the world is missing about 30-33% of its supply.

[00:05:11] How do you think the helium market has been trending in terms of cyclicality? Does the market expect any inflection in the next 1-2 years?

DB [00:05:38]: Let me explain that. Indeed, helium is very cyclical. We went through feast to famine and then back to feast. Before the situation… We started the war with Iran at the end of February this year. Before February this year, the market was actually long. The reason the market was long was because the Russians started to produce some helium. The information coming from Amur is very sketchy, but our understanding is that in a good year, in a good month annualised, they would produce about 200-250 million cubic feet. Now, because of the restrictions on the Russian helium, the Russian helium has been part of the European sanctions for a very long time and also the US sanctions as well, because of the sanctions, the only place for the Russians to sell that helium was China. China, it's a growing market, growing double digit, etc. They have absorbed almost all that helium coming from Russia. What happened, in exchange, that helium displaced the Qatari helium that the Chinese were importing before. That Qatari helium, let's say, 5-6 months, all of a sudden became available to other markets. Those markets were Europe and United States.

Actually, we understand that some of the Qatari helium made available by the influx of Russian helium to China actually ended up on the US shores. Europe, in particular, was very much affected because Qatari helium would go to Europe relatively easy through the Red Sea and almost Red Sea and then the Suez Canal, but also the United States. Qatari started making so much helium, so much helium was available there that they had to find a home for it. The far eastern markets outside China were doing very well. The Korean market was doing well. The Japanese market relatively well, but they could not absorb all that helium that was made available in Qatar. Therefore, the market was long, the prices went down substantially. Of course, that all changed when the Hormuz Strait was blocked by the war.

The other issue that you have to be aware of, during that period, middle to end of 2025 last year, Linde commissioned one of their plants in Freeport, Texas, another very large plant, about 270-300 million cubic feet per year nameplate, but that plant being commissioned, my understanding is that they were planning to store most of the product from the plant simply because they didn't need it and the Linde storage was right there next door, about 150 miles from Freeport. Still, the plant was commissioned and started making helium, but the storage was not yet commissioned. Some of that product actually percolated into the market and was an additional contributor to the oversupply and pressure on pricing.

LD [00:09:23]: On the last point regarding the plant itself, the product was commissioned from that plant, but the product that was in storage was not. Is that right?

DB [00:09:35]: Not the product in storage. The storage itself was not finalised. Right now, I understand… As you might expect, I'm in touch with my former colleagues at Linde. The storage is operational, but right now, they have nothing to store because everything they make in Freeport is good to go. They will sell it in the market. Again, you have a specific question on storage, and I'm going to go through that when we get there, but suffice to say that if you have the product available and the choice is between selling it outright to the retail market or storing it, then obviously, you want to sell it, right?

[00:10:19] If the Middle East conflict were to end in the next few months, how long would you expect the Qatar plants to take to return to nameplate capacity? How close do you think those plants were to nameplate capacity before the Middle East conflict?

DB [00:11:00]: I understand your question. Again, we need to triangulate here because I hear all sorts of opinions from various people, but I listened to Mr Al-Thani the other day. This is the Minister of Energy in Qatar. What he's saying is that it will take 2-3 years to repair the LNG plants associated with making helium. Mr Al-Thani doesn't say helium specifically, but we know that without the LNG production, they won't be able to make helium because the feed gas that comes to the helium plants, the Qatar 1 and Qatar 2 is actually produced by the process of making LNG. I would suspect that the Qatar 1 and Qatar 2, and as a reminder, the Qatar 1 and Qatar 2 combined are 1.2 plus 0.6, 1.8 billion cubic feet, so quite a substantial amount of helium, even if the war ends tomorrow and the Hormuz is open and all the rest of it, because of the repair required to the LNG plants, I don't expect that product to be available for the next 24-36 months.

Qatar 3, on the other hand, is different. It's smaller. Again, remember, it's 400-450 million cubic feet per year. Qatar 3 is not associated with the LNG production. Qatar 3 is actually extracting helium from the natural gas that goes into Qatar as a product, to make power and to run air conditioning and the rest of it. Now, Qatar 3 has a different problem. Theoretically, that natural gas that was required before the war in Qatar to make power is going to have to be required now as well. The issue they have there seems like they have a problem with the liquefier itself. To the extent that this war ends somehow and the Hormuz is open and the traffic of engineering and information and so on goes back to a reasonable normal, my expectation is that Qatar 3 is actually going to start before Qatar 1 and Qatar 2.

The other thing, this is not zero to 100, Qatar 1 and Qatar 2. Qatar 1 and Qatar 2 are getting their feed gas from 20-30 LNG trains. Now, not all those LNG trains are damaged. When I say 24-36 months, I mean the time required for Qatar to go back to where they were before, which was close to the nameplate. Some of those trains are not damaged, so therefore, there is some production. As I mentioned to you before, Qatar 3 is not running. Between Qatar 1 and Qatar 2, we understand that they are making about 100 million cubic feet. That 100 million cubic feet is going to be available, and as long as they fix a couple of those trains, which have been damaged, they will be coupled immediately to the helium production and the 100 million is going to keep on growing. By the time of 24 months or 36 months, they are going to gradually go back to where they were before.

[Expert misspoke. Sheikh Bandar bin Mohammed bin Saoud Al-Thani is the Qatar Investment Authority Chairman.]

LD [00:14:43]: How quickly do you think Qatar could bring it back from 100 million to nameplate capacity? I assume it would be a one-year process.

DB [00:14:55]: I would say 24-36 months, but it's going to be gradual. Right now, they are about 100 million. Imagine a straight line from 100 million to 2.4-2.2 million. That straight line is going to be reflective of their ability to fix their LNG trains.

[00:15:25] Do you think nearly two billion cubic feet of volume won't come out of Qatar over the next 2-3 years? Given that the industry was already long, how much do you think is still held in excess cavern inventory either in the Middle East or in the US and Europe? How long do you think the market could withstand that displacement over the next 2-3 years?

DB [00:16:17]: Again, let's make some assumptions here. Assuming that Hormuz is going to be deblocked in the next couple of months, I would say the market will most likely get back to some sort of balance, not surplus, but some sort of balance in 1.5-2 years. Again, the reason behind that is Qatar actually is only gradually going to go back to where they were before, not immediately. It will take some time for them to build up. The other thing that you have to be aware of is that the industrial gas component or helium suppliers are being affected by the crisis in the Middle East in very different ways. Linde, for instance, is the least affected because… I think it was in the middle of 2025, they chose not to extend one of their contracts with QatarEnergy. That contract went to Messer. Their exposure to Qatar is actually relatively small compared to the total supply.

On the other hand, Air Products has a hole of about 400 million because… Qatar 3 is 100% Air Products, by the way. Air Products has a hole of about 400 million, which is about 20% of their supply. This is Linde and Air Products. Now, Air Liquide, which is arguably the number three in the market, they have a huge problem because 60% of their supply used to be coming from Qatar. Imagine that right now, that 60% is close to nil. Air Liquide, we see them in Europe being very, very short, declaring force majeure all over the place, etc. On the other hand, Linde doesn't seem to have as big of a problem as Air Products and Air Liquide.

[00:18:36] You mentioned Linde's decision not to extend the contract with QatarEnergy. Do you think the Strait of Hormuz closure and overall supply disruption meaningfully change any of these players' aggression and stance towards regional diversification, or do you see many of the dynamics returning to pre-Middle East conflict levels?

DB [00:19:10]: Again, the diversification in helium is always a subject to your ability to sign a new source. The new sources are not available. Helium is only being made in a handful of places around the world. Algeria is sold out. Qatar has a problem with the war with Iran. Russia is under sanctions. Poland is very small. I understand that Linde signed the entire amount or entire quantity of the Polish supply there. It's about 70 million cubic feet per year. Again, one of the reasons why Linde is doing reasonably well is because they're actually plugging the hole with product coming out of Poland as well. What I'm trying to say is when Qatar is going to come back to normal operation, etc, the landscape of source diversification and source ownership is not going to change that much. I think the wildcard here is Russia.

Allow me to make one comment or two on Russia because this may be familiar with you. Amur was supposed to be the largest helium supplier worldwide overtaking Qatar. Gazprom, there, were planning to invest in three helium plants. Each helium plant was 700 million cubic feet. If you multiply it by three, it's 2.1 billion, very, very close to Qatar. Now, all these plants, the three of them were Linde vintage. They were made by Linde, which is one reason I'm familiar with them. The first plant was installed. They had some issues there back in 2020. There were, 2021, two explosions. They call them fire, but call them what you may, the bottom line is that they screwed up the production there. In February 2022, the war with Ukraine started and all that production there ceased. They built one plant, and they had the second plant in boxes, so not yet installed. We pulled out all our engineers from Russia within one month from the war started in February and by the end of March, there was no western engineer left in Amur.

Russians are very resourceful, engineering-wise, pretty smart people. What they did, they cannibalised plant number two, which was in boxes, to make plant number one work. Our understanding is that that didn't work very well, but worked well enough to make about 200 million per year, which, as I said before, in turn went for the most part to China. Again, the wildcard here is Russia. Now, imagine that you have two events happen simultaneously, one is the end of this war with Iran and the deblocking of the Hormuz and the Bab El-Mandab, etc, and the other one is a peaceful resolution between Russia and Ukraine. At that stage, I would say that it will take Russia, it will take Gazprom, again, between two and three years to make all these three plants operational. Gazprom doesn't care that much about profit. They care more about world dominance. They think of helium as a lever. They have a lot of cash. The point I'm trying to make to the extent that the western technology is allowed back into Russia and assuming those two scenarios will happen more or less simultaneously, we are going to end up in a huge surplus. No storage can take care of that surplus.

LD [00:23:30]: If Russian supply were to enter the market after some resolution, what would be the magnitude of that in terms of making the industry long? As I understand it, and based on what you're mentioning, the market is slightly short from the last couple of years, but if Russian gas were to enter the market, what would be the magnitude in terms of making the industry long again?

DB [00:24:08]: If we assume the Russian entered the market at the nameplate and Qatar produces the nameplate, I think the market would be about 20% long, by the way, which is not possible because with the exception of the three storages, Air Products, Air Liquide and Linde, you cannot store that helium anywhere. Under that scenario, what I think is going to happen, either the Russians or the Qataris or maybe both will have to mothball some of their production until the demand is high enough to absorb that product.

[00:24:53] How much helium is typically stored in caverns globally? What is typically held in inventories on a YoY basis? Is it months' worth of supply?

DB [00:25:15]: A couple of things. I can give you a view of the nameplate. There are three major storages. There is an Air Liquide storage here in Europe. There is one cavern owned by Linde and the other cavern owned by Air Products. If you add the full capacity, assuming that they are full of helium or have a certain pressure… By the way, the helium is still there in a gaseous form, so there is an upper pressure, which you cannot exceed because you start losing it, it will vent. If you assume that the three facilities are full to the brim with helium, that will be about half of the one-year demand. If you think of demand about 6.5 billion, half of that is going to be 3.2 billion, 3.3 billion cubic feet, but the storage, the caverns will never be full. First of all, they are pretty new. Secondly, right now, I can assure you they are almost empty because I think maybe Linde might want to store some there. Once again, if the choice is between storing and selling, you go selling because it's money in the bank today, not to mention that storing is a very expensive proposition.

Helium is being produced in liquid form. Then you take that liquid in ISO containers. You bring it to the cavern. You vaporise it. You pressurise it underground. The ground, it gets slightly contaminated. When you need it, you take it back, you have to repurify it, reliquefy it before you send it to the market. The helium from a cavern adds about, depending on the cavern, the Air Liquide one is more expensive, but I would say anywhere between 20% and 40% additional cost at spigot when the product is back in an ISO container ready to be shipped to a customer. Caverns and storage are good for security of supply. Industrial, we are selling that feature, based on security of supply, telling customers, "Don't worry, if there is a shortage, we always have the cavern and we got you covered," but the reality is that helium from the cavern is actually going to be very expensive. When you get a lot of helium from the cavern, that is going to be dilutive to your margin significantly.

[00:28:02] To your point around margin dilution in terms of the earnings headwind for helium players, at the start of 2026, Air Products was expecting around a 4% earnings headwind, but it's now expecting around a 2-3% headwind. Do you expect 2027 to pose earnings headwinds to helium players, or could we expect some positive momentum?

DB [00:28:35]: I was going to say, I think Air Products is a bit conservative here. I think that's probably Eduardo talking to the investment community. Right now, the pricing is up to the roof. What is happening, even with Air Liquide, which is the worst affected, whatever helium you have, you sell it for a significant premium. My expectation is that your sales are going to go down, but your margin is going to go up significantly. I would say helium is actually going to be a contributor to the EBITDA margin percentage because of the pricing. Now, you might have a reduction on sale of, let's say, 10% or something like that, but you could have a significant increase in the margin that you make with that helium in the marketplace. By the way, most of the retail contracts are relatively flexible. First of all, they are short term, so you can change pricing with the market quite easily. I know right now, if I was still employed by Linde, I would be out there increasing pricing. I'm sure Linde does that very well and so does Air Products and so does Air Liquide for that matter. Everybody who has helium wants to make a profit.

[00:30:24] A lot of supply shocks were happening in March and April 2026, when a relatively small number of contracts were up for renewal and negotiations around pricing agreements were also happening. Do you expect most of the surcharges to flow through now? Even for contracts that are already underway and not up for renewal, when would you expect pricing to flow through?

DB [00:31:06]: I was going to say there is always a way to increase pricing. Some contracts have openers, so both buyer and seller are allowed to adjust pricing or to renegotiate should the market conditions at the time of renegotiation are totally or significantly different than the market conditions at the beginning of the contract. I would say it's quite common in retail contracts as well as wholesale contracts. There are surcharges. There is language… There is price escalation formula that can help out. My expectation is that even the ongoing contracts have a potential to increase pricing at this point. There is a little bit of a difference that you have to be aware of. For instance, in the United States, we are not limited to the length of the… The length of the contract is between seller and buyer. Typical contract in the US is five years. Maybe sometimes you'd find here and there seven-year contracts, but five years is the norm. In Europe, we are limited to three years.

Retail contracts, which cannot prove that the seller is making a customer-specific investment, cannot be longer than three years. Now, shorter contracts actually help because arguably, if you have 100 contracts, maybe 60-70 of them are coming up for re-signing this year or next year, and then of course, the supplier is taking the advantage of re-signing that at a much higher price. One other maybe I should add. The helium customers have a certain degree of anxiety related to helium not being available. In some of those processes, helium is irreplaceable. I know cases where they actually pay significant amounts of money even on a black market to get the product that they needed to keep their operation running.

[00:33:34] You mentioned that you expect helium to be a positive contributor to Air Products' margins, despite the company expecting an earnings headwind in 2026. Do you expect a positive margin in helium in 2026 or 2027?

DB [00:34:04]: I think both, but once again, let me clarify what I said. The dollar margin might be going down, might be shrinking. What I expect to increase is the EBITDA over sales, so the profitability of the business, basically.

LD [00:34:26]: I imagine low-single-digit margin for that, given it's been a low-single-digit negative margin for the last couple of years.

DB [00:34:38]: No. Helium has never had a negative margin.

LD [00:34:46]: Air Products was expecting negative margins or at least an earnings headwind, such as an earnings per share headwind, for 2026.

DB [00:35:00]: You mean dollar margin?

LD [00:35:03]: No. Percentage headwind, like earnings.

DB [00:35:13]: I have a hard time to believe that the percentage margin will go down. The percentage margin, if anything, will be stable until they manage to increase pricing on the limited volume they have, which then will contribute to higher pricing. I'm trying to rationalise that. It might have something to do with their costs as well. Again, pulling from the cavern is actually going to increase their cost to deliver. One cubic foot of helium from the cavern is 30-40% more expensive than one cubic foot of helium from their production facility in the United States. That could be a contributor to having pressure on the percentage margin, but that is only temporary because once you factor it in, there is no way but up to go because of the price increase.

[00:36:15] I understand that spot comprised a small percentage of volume in the helium space. In the current environment, we'll see a slight uptick in the volume mix for spot across contracts. What might be the drivers for spot to gain more of the helium volume mix moving forwards and even after the conflict is resolved?

DB [00:36:44]: What the helium suppliers do, and I'm referring to the smaller ones like Helison Marketing and the likes of PGNiG in Poland and so on, they have contracts, but they also allow for a certain percentage of their expected supply to be available for spot. In a time of oversupply, that spot is actually not doing very well. I've seen spot going down below contract price because there was so much product in the market, nobody was interested in spot. In times like now, the spot is gold. My understanding is that PGNiG here in Poland, they are selling spot at 2-3x the price of contracts. If you say the price of contract is USD 500 per, say, 1,000 cubic feet, the price of spot would be USD 1,500-2,000 per 1,000 cubic feet. Right now, spot is very lucrative. The challenge is that for these companies to actually be able to carve a portion of their production available for spot. If I'm contracted, then they need to deliver me five containers per month, and I know their production is 100%, I do expect five containers per month. If they cut me to four, then we can have a legal problem there. Right now, the helium suppliers would want to sell as much spot as possible, but the limitation to that is their contractual obligation.

[00:38:30] In terms of end market mix and growth rates across aerospace, electronics and medical, how should we break out the drivers in those spaces? What does the historical volume mix across those markets look like for helium? What should we expect moving forwards?

DB [00:38:55]: The growth driver right now is electronics by far. Electronics started some 10 years ago started from a very small base, but with the advent of the new fab shops and the increase of the circuit from 150mm to 300mm to 450mm, the surface that needs to be cooled with helium increased significantly, almost asymptotically, we know, with the size of the wafer. The driver is electronics by far. Space is growing, but it's growing from a small base and is very much focused on the US and China. In Europe, in my last couple of years in the job, I sold a couple of contracts to space, but nothing compared to SpaceX, nothing compared with NASA in the US. Again, driver is electronics, I think, with the new fab shops and so on, followed by space. The medical business, which is primarily MRI, liquid helium going to cooling down those coils and so on, that's shrinking. It's shrinking because the OEMs, the likes of Siemens, Philips and GE, they are making more and more efficient machines that require less and less liquid helium to cool them down. When you add this all together, you put the package together, I think the market growth assuming that product is going to be available, is still in line with the global GDP, so I'm thinking by 3-4%, in that range.

One comment I was going to make on the electronics, and I know you have something very specific on there, but let me just say that the demand for electronics is still an expectation, not a reality. You make all these fab shops, and we'll talk in a second about the investments that Linde and Air Products are making in electronics, but this is all on spreadsheets happening in the years to come. Nothing is concretised right now other than what they are now selling to the likes of TSMC, Samsung, Infineon, etc, currently. All these new fab shops, which is, basically, the engine for growth in helium, is still an expectation rather than a reality.

[00:41:36] Is the electronics sector showing double-digit YoY growth for helium? You mentioned that it's due to the increased complexity of the circuitry. Is it a mix of the complexity and the massive investment this space is seeing?

DB [00:42:03]: There are two engines of growth. One, as I said, is technological. They move from smaller wafer to larger wafers. 150 was 10 years ago. Then they moved to 300. Now, they are moving to 450. You've seen those wafers. They are like disks. When you move the diameter from 150 to 300 to 450, the surface of that disk is actually growing asymptotically, is growing a lot faster than 150, 300, 450. The more surface you have to cool down, the more helium you need. Now, in the last… One of the shortages, I shouldn't say the last, one of the shortages that we had back in, I think it was 2013-14, some of the wafer manufacturers, and I was involved personally with Samsung in Korea, they found out that the load lock where the cooling takes place does not necessarily require 100% helium. They could do with a mix of helium and argon. Of course, argon is 10x cheaper than helium and more abundant. At that point… Actually, there was a dip in the helium demand for electronics, but once this technology built in and everybody is doing more or less the same using a little argon into helium to replace some of the helium they need, the helium is actually going to grow. Expectation is that helium is going to grow double digits and will go in tandem with the new fabs that are being built.

[00:43:59] What is your growth expectation for the large electronic gas contracts that Air Products and many others are signing? How would you characterise the boom that the company has seen in these large contracts in the electronics space? Its backlog is two-thirds electronics.

DB [00:44:39]: In my opinion, it's very risky. Not only my opinion, if you look at the… I just read recently something about Linde and these investments of USD 1bn in Arizona and another USD 1bn with TSMC over in Taiwan. It's proverbial putting your too many eggs in one basket. The issue with electronics, to me, is relatively simple. The electronics developments, these fab shops, the new fab shops, they require a basket of products. They require air gases tonnage, so you need to build ASUs, etc. Then they require a portfolio of other industrial gases, which are speciality gases, electronic gases and helium. Now, these are two different animals. On the one hand, you have a long-term investment. Let's say, you invest USD 30m or USD 50m in an ultra-high-purity ASU. That ASU is going to be depreciated over a longer period of time. You need a longer contract. You need 15-year or maybe 20-year contracts.

The gas companies are so anxious to get into the AI and so on. I know that the target is 15% IRRR [IRR]. By the time everything is said and done, specifically on the investment, you probably sign it at 12% return on investment, but then you have the other gases that are producing a lot better margin. When you add them all together, they use helium and electronic gases to hide, I'm going to use the term hide, but you know what I mean, to hide the lack of profitability in an investment. The problem with that, again, is relatively simple. The investment in an ASU that provides this customer with gases via pipeline has a backstop, and that's called take-or-pay. No matter what, the customer, to the extent that they remain business viable, they will have to pay that take-or-pay. You cannot go below that. Now, with the other gases, they very seldom accept take-or-pays. Therefore, the electronics companies are notorious by overstating their demand. If they say, "Oh, we need 200," your expectation should be that that 200 described is going to be about 120 or 150 by the time everything is said and done.

What's going to happen? The profitable product could potentially be, the volumes that you sell could potentially be less than what you expect. Then that is actually going to uncover the lack of profitability in your tonnage business. That's a thing that everybody fears. If you look at the new valuation of Linde, Linde is suffering now. The stock price went from USD 550 to, what is it, USD 100 less. One of the reasons behind that is because analysts like yourselves are very concerned with the profitability of these projects going into the future, and the fact that you are investing USD 2bn, not knowing what you're going to make in five years from now.

LD [00:48:16]: Do you think that's going to force electronics customers to dual-source their industrial gas supply in this space, given it's all ASUs [air separation units], or will it be a mix between ASU and merchant volume?

DB [00:48:34]: In the United States, you can have the all requirements. My expectation is, for instance… I led the investment in Arizona with TSMC. My expectation is that there is going to be an all-requirements contract. The tonnage cannot be replaced. Either provide the tonnage business, the pipeline business to the customer or you don't provide it. There is nothing in between. The other gases, usually, they package them with the tonnage, and they expect the customer to sign on all requirements, meaning that they will be buying everything from one supplier. The challenge there is that those requirements might actually be less than originally thought. In Europe, we cannot do all requirements. There's some electronics coming up here. Some of them are delayed. Intel was supposed to invest in Germany and in Poland, and we are not seeing sign of that just yet, I guess, because of their financial situation, but in Europe, it can sign a tonnage contract, and that's all requirements. It's you supplying all the gases via pipeline to this particular customer. The other portfolio of gases, electronics and helium cannot be all requirements. Here, you can have two suppliers or more.

LD [00:50:05]: In terms of the durability of the backlog, you mentioned that, if electronics customers are expecting 200 million cubic feet or tonnage for a project, they might only get 120 million or 150 million. Is that fair to say?

DB [00:50:32]: No, I used it as an example. The number is absolute value. Again, the tonnage has a backstop on take-or-pay, so the customer cannot buy less than the take-or-pay, but the other gases, the volume is flexible. I know if the customer says, "I need 200 million," at the end of the day, they might end up with 120 million, 150 million. When you are diminishing the amount of product you are selling at a premium, then the other product, which is less profitable, is going to show up real quickly.

[00:51:23] How durable might Air Products' backlog be through 2030 or 2035, given that two-thirds of it is electronics? Is there a risk that most of it won't be realised in the next five years?

DB [00:51:49]: I was going to say, it all depends on the demand for AI at this point. Imagine if the market crashes because, all of a sudden, people pull their plug on AI, all this investment that's made with the AI in mind and the need for additional computing and so on is going to crash with it. Now, I'm sure this is a bit of a doomsday scenario, but imagine that something like this happens, I would suspect that some of those investments will be put on hold. Again, very unlikely, but it's a possibility.

[00:52:37] Within helium and other gases, where do you see market share gains or losses for Air Products relative to Linde, Messer Group or Air Liquide? Where do you expect volatility?

DB [00:52:59]: I was going to say Air Products.

LD [00:53:01]: Within helium and other gases?

DB [00:53:04]: Within helium, Air Products might be losing some helium business to Linde. The reason is simple. Linde has more products. Air Products has less products. They might be trimming some less profitable customers, and Linde would have the opportunity to pick them up. Messer is private, and Matheson is Japanese-owned. Both Messer and Matheson have, how shall I put it, different way of looking at profitability compared to publicly traded companies, compared to US-based companies like Linde and Air Products. The only problem is that neither Messer nor Matheson has sufficient helium product to be aggressive in the market. Had they had the product, they would be the usual suspects. They would be the ones going out there and signing lower-profitability contracts just to increase their market share. I think Air Products might be losing some helium business to Linde, but some of the business is vulnerable to both Messer and Matheson. Again, I'm sure Eduardo is pushing this increase in profitability. Air Products has a certain profitability expectation that they need to achieve, while Messer doesn't really care much about that, nor does Matheson. My expectation is that on the merchant side and maybe on a tonnage side, they might be losing some contracts or not winning as many contracts as they should have otherwise.

[00:54:49] Do you expect Air Products to recapture any market share losses once the supply shock normalises?

DB [00:54:57]: On helium, yes, absolutely. As soon as they get the product and those contracts become available again, that's another story, I'm sure that they will be going out there and trying to re-sign that. That's normal.

[00:55:18] We talked about some of Air Products' investments into TSMC, China and the Arizona plant, but the company cut back on its 2026 CAPEX outlook by about USD 500m. Most of that was going towards maintenance. Do you expect an acceleration in CAPEX in 2027 and 2028? What would be the risks if CAPEX doesn't accelerate next year?

DB [00:55:54]: I was going to say, with Eduardo Menezes being the Chief Executive of Air Products, I did expect their CAPEX to go down significantly, and it did. Eduardo is taking a very critical look at the CAPEX-intensive projects that Seifi Ghasemi signed before him. My expectation is twofold. Number one, that as the numbers say, he is going to lower the CAPEX spending, and number two, he is going to write off projects that don't make financial sense. Some of them are related to investments that Seifi made in Saudi, I think somewhere in Bahrain and some in that part of the world. Some of those projects are not particularly profitable. I think Eduardo is going to take a financial look at them, and if they don't make sense, they are significantly dilutive to his margin expectation, he's going to get rid of them.

LD [00:57:05]: Maybe one of those two is the Neom Green energy project.

DB [00:57:11]: Yes, I read that. Air Products is actually advertising that as something major development, etc. A couple of things. First of all, on green hydrogen, green hydrogen is manufactured by electrolysis, by splitting water. Electrolysis require a lot of power. Hydrogen is only green if that power is either wind turbine or solar or so on and so forth. Now, imagine that the Kingdom of Saudi Arabia, which is the number two producer of oil worldwide, much of their electrical energy is actually provided via normal power plants run by natural gas or by some fossil fuel. The heavily advertised green hydrogen there, I'm sceptical that this is indeed green. Then they have the so-called grey hydrogen, which is power made in a traditional way using fossil fuel, but then use that power in an electrolyser and you split water and make hydrogen.

I internalise a little bit. Let's say, they are successful in actually making this, what they call, the green ammonia in Neom, and they have this JV with Yara, where Yara is being commissioned to market the product. Now, the market for that ammonia, the local market for that ammonia is very small. Everything that you make in Saudi, you have to export. We all know what's happening to the Saudi exports, whether it's natural gas or oil in particular. The Hormuz is closed. The Bab al-Mandab, they have problem with the Houthis there. The only place that they can ship something is to the Port of Jeddah, which is shallow water, so it's not particularly suitable for large container ships. I think you can make the product very cheaply and what have you in Saudi, but then you cannot sell it in Saudi. You have to bring it to a market that requires it, and those markets are very far away.

[00:59:42] Air Products is pulling out of some clean energy and hydrogen products, taking upwards of a USD 3bn tax charge for that. How would you characterise this exit? You talked about CEO Eduardo Menezes pulling out of the contracts that were seemingly not profitable, but should we expect more of these clean energy exits from Air Products and others?

DB [01:00:29]: Again, I think clean energy is still in the eye of the beholder. I'm sure there is money to be made in it, but it's not something that's going to happen overnight. The clean energy is something sexy, that looks good on your annual report, but making money with clean energy is still maybe a few years away. It's not going to happen immediately. Saudi, one point I wanted to make. With these Neom projects, Air Products and Yara are the beneficiaries of significant subsidies and tax breaks and money from the Saudi government and so on. Maybe by those subsidies, you can make a profitable business case, but remember, subsidies are pretty fickle. They are a question of the government supporting them. If the government changes its mind, and many governments do, and pull the plug on subsidy, then all of a sudden, a project that made a 30% return on investment is going to go below red line real quickly. The point I'm trying to make, I would not make a long-term investment based on a subsidy. If you can make something that goes quickly and you can have the local government commit to 10-15 years of tax breaks and subsidies and so on, yes, but if the subsidy is temporary, then those projects are going to be very problematic from a profitability perspective.

[01:02:08] Is there anything else you think will be important to address or keep in mind as we assess Air Products' positioning moving forwards?

DB [01:02:21]: I was going to say, I think the one point I was going to make, Eduardo Menezes is the best thing that ever happened to Air Products, that I can tell you. I worked with Eduardo. He was here in Europe. I know his style. He's a very pragmatic man, very short term, but then again, the US is a short-term economy. Eduardo is going to expect to bring the Air Products numbers in line with Linde, and in my opinion, and I looked a little bit at the Air Products numbers, I think they are very close to being that. If you look at the EBIT over sales and a few other financial metrics, they are very, very close to Linde. There are two big differences here that probably you have to keep in mind. Air Products of today is very similar to Praxair before the merger with Linde. It's a company easier to manage because it's limited from a geographical expansion perspective.

Linde, on the other hand, is a very large company. It's like a big container ship. When you move, it's not going to respond real quickly. It has a lot of inertia. My expectation is that actually Air Liquide eventually, within foreseeable future, say, 6-12 months, is actually going to overtake Linde from a financial perspective, profitability that is. Certainly, not in sales. They are less than half in sales, it's a much smaller company, but from a profitability perspective, Air Products has a lot of upward potential with a good manager in charge.

LD [01:04:11]: Awesome. I appreciate your time. I know we're just a bit over the hour here. Thanks so much for your flexibility. Really enjoyed the Interview today. Thanks so much for taking the time.

DB [01:04:24]: No problem. Anytime. Thank you.

Transcription ends at 01:04:29 of the recorded material