Good morning, ladies and gentlemen, and welcome to the Elmos Semiconductor SE Conference Call regarding the results of Q2 2021. The floor will open for questions following the presentation. Let me now turn the floor over to your host, Dr. Arne Schneider, CEO. Go ahead.
Good morning, ladies and gentlemen, welcome to the Elmos Q2 Conference Call. I can once again present to you a solid performance with strong increases in sales and EBIT in the past quarter, despite a challenging market environment characterized by an ongoing high demand for semiconductors and the corresponding allocation challenges. As usual, you have the opportunity to ask questions at the end of my presentation. Before I'm going to explain the current market situation and the financial results of Q2, I would like to give you a short update about the COVID-19 situation at Elmos. Since the outbreak of the pandemic in January 2020, we have introduced extensive protective measures at all Elmos sites. This has enabled us to maintain our own production and business operations until today without any major disruptions.
Starting as early as April this year, Elmos has coordinated three vaccination rounds for our employees and their relatives. This campaign was very successful. At our Dortmund site alone, almost 600 people have been vaccinated by external physicians. Every Elmos employee received a vaccination offer with the vaccine of choice. I think this is a great success and important step for us in the fight against Corona. Since July, we have now introduced first small steps to ease some of the Corona-related measures at Elmos concerning small group meetings or business travel, and we also started the gradual and partial return to the offices. We are prepared to tighten the protective measures immediately should it be necessary. Turning to the market environment we currently see. Since Q4 last year, we have witnessed a very high demand for semiconductors across all industries.
The ongoing boom of office and consumer electronic products and the faster-than-expected recovery of the automotive market resulted in soaring demand levels for semiconductors, which could not be satisfied by the global IC manufacturing capacity. This led to a longer and also more severe than expected allocation situation, which is also by far from not over. The entire semiconductor supply chain, including wafer production, assembly, and testing, is still under a great deal of pressure. In addition to the increasing real car demand, all automotive OEMs and Tier 1s naturally want to rebuild their inventory levels and increase safety stocks after the reduction during the corona crisis last year.
To secure as many ICs as possible and to avoid potential line shutdowns, the current order levels of some automotive customers are significantly higher than their real production demand, adding even more stress to the already tight value chains and leading to some heated discussions concerning how much can be obtained. More and more market participants now expect that the current situation most likely may get worse before it will get better. There's also extensive coverage on the topic in the newspapers. The capacity bottlenecks, especially for eight-inch wafers, will continue in the next year, which could temporarily dampen growth in 2022. All IC companies, including Elmos, are faced with the limited fab and foundry capacities, supply bottlenecks for key components, significant material price increases, and logistical challenges, all leaving traces on the business performance. I will further comment on this later in my presentation.
In this dynamic market environment, Elmos once again had a successful quarter. Group sales increased significantly by 34.2% in Q2 to EUR 78.9 million. Again, a new record in quarterly sales. You may have noticed that a part of that sales, of course, came out of our inventory. The EBIT improved to EUR 12.5 million in the second quarter of 2021, based on higher volumes. The EBIT margin improved significantly to 15.9% compared to 5.4% in the previous corona-ridden year. Overall, the Q2 results are in line with our quarterly guidance. However, as I mentioned before, allocation-related effects, especially material price increases, started to impact our results in the second quarter. Our R&D expenses continue to remain on a high level, supporting the numerous serial launches and new development projects in all application fields.
After six months, we are still fully on track with our new design win activities, and we are able to acquire new projects in all of our product segments. Despite the temporary allocation situation, we have significantly intensified our own CapEx program in the back-end area as already announced earlier this year. In Q2, we have invested EUR 17.3 million, mainly in the testing area to install the necessary capacities at the Dortmund sites and at our partner sites in Asia to secure our delivery capabilities. Although we have further reduced our own inventories by around EUR 3 million in the second quarter, the cash flow from operations was lower compared to Q1, mainly due to seasonal effects.
As a result of the current allocation and the lack of external foundry capacity, we have reduced our inventory by more than EUR 10 million to fulfill our delivery obligations in the first six months of 2021 alone. The significantly higher CapEx spending led to a slightly negative adjusted free cash flow of - EUR 3.4 million in Q2. The higher spending will continue in the next months, influencing free cash flow in the further course of the year. Accordingly, our net cash position decreased to EUR 47.3 million at the balance sheet date due to the negative free cash flow and the dividend payment in May. Looking ahead to the third quarter, based on the current order situation, we expect another successful quarter, which will continue to be notably influenced by the allocation situation.
For Q3 2021, we expect sales of EUR 79 million, ±EUR 5 million, and an EBIT margin of 15.5%, ±2 percentage points. The wider range of our sales and EBIT guidance reflects the increasing uncertainties about the future course of the allocation situation, the pandemic, and related challenges such as logistical delays. That's also why a reliable forecast for the full year is still not possible at the present time. As I mentioned before, we expect that the challenging market environment will continue for the time being. Looking ahead into the next year, we are currently in negotiations with our foundry partners to secure sufficient wafer capacities for 2022. As you can imagine, these negotiations are very tough as it has become pretty obvious for everybody that the capacity tightness will continue into the next year, especially for eight-inch wafers.
Although some foundries have announced that they will allocate more capacity to automotive products, we do not expect a significant increase in overall eight-inch capacity next year. As you may know, all Elmos ICs are manufactured on eight-inch wafers only. Since our own fab in Dortmund is fully loaded and cannot be expanded any further, our growth potential critically depends on external eight-inch foundry capacity. Our view regarding potential capacity constraints in the next year is supported by various announcements made by foundries or OEMs in the last weeks.
For example, DigiTimes Asia published an article in early July stating that, and I quote, "Foundry capacity supply could become even tighter in 2022, particularly at eight-inch fabs." That, I quote again, "Capacity promised for 2022 can satisfy only 70%-80% of demand." This is, status today, a realistic reflection of the actual situation and the current state of our negotiations, as more and more market players expect that the global semiconductor shortage will continue into the next year. We, of course, work closely together with our wafer factories to ensure that Elmos is not the cause of any line downs. We were not the cause up to now, and we will do our utmost to keep it that way. However, despite these temporary growth dampers, the semiconductor market will continue to develop positively in the long term.
Based on our innovative product portfolio and strong financial position, Elmos will be able to successfully participate in the positive development of our markets in the future. Far, thank you very much. I'm opening the floor for questions.
Ladies and gentlemen, if you'd like to ask a question, please press nine star on your telephone keypad. Please press nine star if you'd like to ask a question. The first question comes from Stéphane Houri from ODDO BHF. Your line is open.
Yes, good morning, Arne. Good morning, everyone. I had a few questions on the outlook, obviously. What prevents you from now to guide for Q4 already? What kind of order book do you have? We had some comments yesterday from Infineon saying that basically they had 2 years of order book. I assume this is probably a problem of capacity and supply chain, but is, at this stage, a flat Q4 a good assumption? That's the first question.
Very long question ahead of me. Also for 2022, how should we translate what you are currently saying? Are you calling for a no-growth year in 2022, or do you think you can gain productivity or maybe price increases will provide you some growth? Yeah, if you can also describe the discussions with the foundry. You said it's very tough, what can we expect at this stage? Thank you very much.
Thank you, Stéphane. On your first question, we do see that we come into a situation where our inventory is basically empty. At the same time, we see the coronavirus pandemic is not over. We see partial or full factory shutdowns, sometimes for days, sometimes for weeks, of some suppliers in Malaysia, in Thailand. At the same time, logistics gets less plannable. Also, there are some supply constraints, such as lead frames at assembly houses.
Overall, the key thing why there is volatility is not that our customers are not willing to take any amount of ICs and all amount of ICs that we can deliver. The key thing is that since we are living from hand to mouth, basically, in the value chain, that is the hard thing to predict. Our order book, just as Infineon commented yesterday, I believe, is super full.
Given on the orders, I guess, or I feel 120 or 140 million cars could be built, but these are wish orders. They are reduced to the allocation level that the customers get. Orders these days are nothing. You get an allocation, and that is what you can expect to get delivered. The key in the volatility lies more in the value chain, lies more in actually bringing it out. It's stronger than it was in the past, because, as I said, we cannot just say, "Well, we have the inventory. We are sure that we have." No, we get it, we test it goes out. It's all a very continuous flow. If we lose the week due to whatever, COVID-19 or material shortage, then it's just not there. It's no revenue with all the things linked to it.
That is the key concerning why we have a little difficulty being as precise as we have been over the years with our forecast. On 2022, we are in a very unclear situation. This is not finished. In the current state, we have basically a flat wafer allocation from our foundry partners. Given that we delivered substantial volume out of inventory this year, this would lead not only to no growth but to worse next year. Obviously it cannot stay that way, and I do expect it will not stay that way. However, these are intense, difficult discussions that are going on, and they are key for our course in 2022.
We just want to make you aware, and also since you can read it in the newspaper basically every second day, I believe it would be foolish that we don't talk about it and just ignore the existence. These are discussions where we do not know the end of it. We are very much pushing for a good end for Elmos, but status today is that it's unfinished.
Okay. Can you just remind us how much of your inventories you sold in your revenue this year so far?
Well, so far we had EUR 10 million less inventories in the first half. That is probably, at a very rough rule of thumb, double that in revenue. Because, of course, in inventories are counted at cost, and with our gross margin, it's around double that in revenue.
Sure. All right. Just a housekeeping question. Can you update us on the tax rate you are expecting for this year and for the coming years?
Well, I would guess that our tax rate will fluctuate around 30%. Not only this year, but also for the time being. We actually, some years ago, did a tax project to see whether we can somehow reduce that. In Germany, it was the time of Mr. Schäuble, but under Mr. Scholz, things have not changed in any way. It is very hard to get to lower tax rates. We would have to shift substantial business operations to Ireland or Singapore or somewhere where we just cannot shift very substantial operations to. We could shift a little bit, but then this has not the effect that we wish for. I cannot give any hope there. We will likely fluctuate around the 30%.
Okay. Thank you very much. I have other questions. I will go back in the line. Thank you.
Thank you, Stéphane.
The next question comes from Johannes Ries from Apus Capital.
Yes, good morning. Also some questions, like always. Maybe following the question from Stéphane, maybe you can to repeat it, what about pricing? You talked about price increases in your supply chain, but I think you are also able, and Infineon has talked about also yesterday that it's able and the customers accept it and know about the situation to increase the prices at your customers. Maybe with some time lag, but could that help in some regard to get at least stable or maybe revenues or growth next year? Like you said, also, it's not key if you stay with a stable capacity. I have a follow on, but first maybe let's start with this pricing question.
Yes. Concerning pricing, of course, we do increase prices. We do negotiate around that. It's not all finished, but we have to ask our customers for their share of the burden concerning allocation-related costs, because these are real, and these are large. We cannot just take that on our own shoulders. I believe it's a very fair approach to do a cost share when it comes to allocation-related additional cost. Meaning that we have some burden, but we for sure have not all the burden, and we also have the growth, which somehow goes against that burden when it comes to profitability. That are, of course, price increases, but they are not huge. Theoretically, of course, yes, they do help to grow, but it's small growth.
They should help to at least hold the gross margin now. Far, to surpassing your cost increases in the prices now. That's the idea.
Yeah. Well, the idea is, of course, to find a fair solution with our customers. If that means a really stable gross margin or slightly decreasing, but then we have OpEx effect and scale effect in the OpEx such that we can somehow keep the overall profitability, that would probably be okay. We don't want to get rich by charging our customers in these difficult times unfair amounts of money. This is not our policy.
Our policy is to share the burden with our customers in a fair way, and that's what we ask of them, and that's where we also have to insist on, because there is no other way.
Maybe on your own fab, how much of your sales are coming percentage-wise around from your own fab? You said you are fully loaded, it's clear, but are not some efficiency measures possible to a little bit, maybe in the low percentage-wise, maybe increase the output next year? Is it totally impossible because you can't increase your efficiency anymore?
Well, yes. First of all, it's around half-half internally.
50/50, okay.
I believe end of last year or beginning of this year, I asked Guido Meyer, my head of production, the board member from production, for an additional maybe 15 or 20 wafers. Of course he said, "Oh, no, this is way not possible. Maybe five a day or maybe eight." That's where you can see this is really limited, but we try to squeeze out as much as we can.
It is not a surprise, but thanks for the confirmation. Maybe another question, there is Erfurt fab, which will still have capacity because it shifted to, it's a foundry site now to more advanced solutions. It's close related to your main competitor, the X-FAB. Could therefore maybe more general, are there other maybe potentials to go to a third foundry partner to get additional capacities, especially for new projects? I know you need this certification, therefore it's hard to move an old project to a new source. Is that a potential opportunity, and is this X-FAB solution totally out of any discussion because it's so close to Melexis?
Well, the X-FAB discussions we wouldn't entertain for the reason you mentioned, but also for a timing reason. This would become online in one and a half or two years, this is no immediate relief.
That would also concern every new foundry. However, what we do, we try to ramp our 130 nm projects a little quicker.
We ask some customers to take the newer product, because then we get overall a little bit more wafers for next year. This is a very limited thing, but still, we try to find all the measures that there are to ensure that we deliver to the best of our abilities, including all additional measures one can think of.
That's clear. Finally, if I look to your CapEx in the back end, it's really high. So far, you don't invest in your testing capacities if you not see a very interesting growth maybe after the year 2022. Is that right, this assumption?
Yeah, then even in 2022, I want to caution the community. The only thing that we can currently say, it's an unsolved problem. However, with the lead times that exist on testing equipment, if I were to decide today that I only invest in testers where I have the wafers for, then we would need to invest a lot less. These are open discussions, and they may be open for some time, 2022 for sure will not be closed quickly. We have to go out on a limb there and make an assumption. The thing that we, as the board of Elmos, cannot accept to happen is that Elmos is unable to deliver because we failed to invest in testers. That should never happen.
Great. Super. It's also for the years after 2022, because you said very good design wins, a very full order book, and so on.
If we do this some months later, because then these are things that run for 15 years. It's not that the first few months. In these days, we must be able to deliver the key. Then it comes to optimization and efficiency. The first thing is really the key that we are able to deliver.
Great. Thanks a lot. I wish all the best in the negotiations with the foundries.
Thank you, Mr. Ries. We use it. We keep it.
Okay, thanks.
The next question comes from Malte Schaumann from Warburg Research.
Good morning. The first question, unsurprisingly, again on 2022. From a timing perspective, for the first quarter next year, say, wafer production has to start maybe in two or three months. Would you say that this is more or less fixed, the volume, and that the discussions are rather open then for the second to the fourth quarter? Is there still any scope for even more capacity for the revenues to then require in the first quarter?
Well, I think we will get decisions on our wafer allocation and what we finally get as wafers, not what we currently have, but what we finally get, which I hope is more than today, because otherwise the chaos will be large. It must be more than today, but we won't get it for the full year, generally, I would expect. I would expect we kind of have decisions that are relatively close to the time where you actually have to start the wafers at the foundries.
Okay, that will be kind of a monthly or quarterly thing, probably?
Well, there's no clear process written down for these kind of proceedings.
Okay. Still for the first quarter of production late in Q4, still things are pretty fluid ongoing.
As you said, in the next month, we have to get some sort of a result out of that. That is what everyone is aware of. It's not today, but time is flowing forward, so yeah.
Yeah. Okay. Do you have any insight in the decision-making process in the sense that is your supplier keeping the capacity and then deciding on very short notice who's offering the highest price? Is it then a scarcity issue, to avoid line downs, as you said, and who needs it the most from an urgency perspective, to be flexible in that respect? Any insight, which are the criteria your supplier makes basis his final decision on finally?
Yeah, it's all those and with various weights at different suppliers. Some are less concerned about their reputation and more about money. Others are more concerned about doing the right thing. Of course, if you allocate all the capacity in eight-inch globally only to automotive, and then the car guys unconstrained, sit on high inventories after 2022, which would happen if you let it go totally unconstrained, you also did the wrong thing. There is no other way than to discuss what is really needed, to keep everything rolling. These are, of course, difficult discussions. We hope that everyone is reasonable, and we keep on rolling as good as we can.
S ure. Okay. On the CapEx, the Q2 number, was that kind of a spike, or should we expect kind of a similar level somewhere in the third or fourth quarter?
W e will actually invest quite a lot this year. There is potential volume, at least if we look at the demand from our customer set that we need to prepare for. As I mentioned, I think it would be, given that we got longer and longer lead times on machinery, we have to be careful that we're not caught in allocation there again, such that we ourselves are the bottleneck to delivery. That is currently not the case, we want to really make sure that it will not be the case in the future.
Yeah. In the engineers anyway, so doesn't matter. Strategically, is the current situation kind of accelerating your plans to move products to 12-inch wafers? I understand that this is a lengthy process, is there something in volumes that's for the next generations, is able to be moved to 12-inch?
Today, no. Concerning the next generation technology after the one we just started with, the 130 nanometer, you can discuss many things, until that comes into serious delivery, that is a long time. We will, for the time being and for the next years, many years, we are on eight-inch. We asked some foundries whether for them it makes sense to transfer processes we are on right now to 12-inch, they declined that. We are for the time being on eight-inch.
Yeah. Okay. Maybe a quick word on design wins, everything according to plan? Strong we go?
Strong and on track. These are also the days where when you ask the customer for written confirmation of their design win, which may only come online in 2024 or 2025, but still, we have a climate where semiconductors are valued. If you want to close the deal, then it's not only the supplier that wants to close the deal, it's also the customer that at some point wants to close the deal and say, "Okay, now we agree. We sign it, that's it." That's, of course, the requirement for design wins. Design wins develop nicely.
Okay, good. Thanks so far.
The last question comes from Robert Sanders from Deutsche Bank.
Yeah. Hi, good morning. My first question is about redesigning by OEMs and Tier 1s. Can OEMs or Tier 1s redesign with alternate suppliers if you're not able to deliver products? Which of your products are substitutable and which are just not because they are so specific and qualified that that couldn't happen? I have two follow-ups for you.
Well, I think it is rare that you get on anything. Of course, you can, and you can do for everything. It only depends on timing. Within reasonable timing such that it has any effect, say, on 2022, this is a very hard thing to do. You can only do that if you're willing to take substantial risk of failure of such a new product. You will be unqualified, you will have no tests done. You will just shoot in the dark. The thing that you can discuss is leaving things away, to restrict your customers or the end customers of a car, that you just say, "Look, this is a feature, but it's unavailable." Say you have a four-zone climate control and you say, "Look, this is a feature currently unavailable.
You can only get it with two-zone climate control because you save ICs with that." That is a more realistic option. Although, of course, I believe the car manufacturers hate it. We haven't seen that to any noticeable degree up to now.
Got it. If you get a flat wafer allocation in 2022, at what point do you think your OEMs would start to have the issue around line downs? I guess this year there'll only be 80 million production units, but your demand today is probably 95 million. There must be some inventory out there that they can absorb, some kind of buffer inventory. At some point, that will be exhausted. Do you think the line stop issue would materialize in '22, or it might crystallize a year after?
No, we are pretty sure it materializes in Q1 2022. Because we did the exercise with the OEMs, and this feeds into our discussion with foundries.
Got it. The last question is, on the OEMs, they're all profit warning on the second half at the moment. BMW just yesterday, They've exhausted their inventory, now semi shortage is really biting. The thing I don't really understand is in first quarter, there was a lot of attention on NXP and on Renesas because of the microcontroller supply issues. That led to the foundries going public and saying they were prioritizing automotive microcontrollers at the 40 to 65 nanometer nodes. Why has there not been a proper escalation process in these smaller value parts like sensors and power management chips? Why has there been a kind of breakdown of communication here in the industry, given, as you say, there could be a line down in Q1 2022 caused by Elmos, heaven forbid?
There is a lot of communication to avoid that right now. I believe we are kind of going forward as we go. There is no great all-encompassing map of who needs semiconductors and how many and where they come from, because it's just too complex. As we go on, we need to discuss month by month, quarter by quarter, however it will happen, what needs to be done. That is what is currently happening. It's a lot of dynamic in that situation. We have a view on what needs to be done, and the OEMs have a view on what they definitely lowest level possible need. The foundries, of course, need to protect their other customers.
Also, I believe fundamentally, they make their point, "Yes, we want to support." Of course, you also see that there's limited transparency concerning the whole world at the same time.
What incentive is there for a Tier 1 to be transparent about their inventory when having inventory is now a badge of honor that now enables them to take market share? Why would I be transparent if I'm a Tier 1, for example?
Well, there is just not that transparency from all sides. If you have a certain product. We have a lot of product, and we are a small player. We have a lot of customers, and they again have a lot of customers. This is, even for the pure Elmos thing, this would be a gigantic Excel. If you then add some of the others, because it's an optimization not only for the little Elmos share, but also for all of our competitors that require foundry capacity. That is just not there. There is no complete halfway real-time stock level, plus IC level, plus lead times, plus end demand map. We try to give as much transparency as we can get, as the OEMs can provide in addition, and then that's the basis for a discussion.
In the end, there have to be judgment calls, what can be done and what can't be done.
Got it. Thanks a lot for your insight.
Thank you.
There are no more questions. Ladies and gentlemen, if you'd like to ask a question, please press nine star on your telephone keypad. There is another question from Johannes Ries. The line is open.
Yes, a short follow-on to a more general topic. We discussed it even in the last call very briefly. Have you heard anything from politicians from the EU, as they are all talking about maybe to build up own European capacities and to talk with the industry? Have you heard anything about them?
Yeah, we will be part of an interest group just to be close to what goes on. However, I think these are more kind of very long-term things, because building a fab, choosing a location, building it, getting it online, getting processes transferred, we're not talking 2022 or 2023 there. We may talk 2025 or 2026, if it's applicable to Elmos at all. What you sometimes read in the newspaper, that the European fab will then immediately start to solve all the problems. I think this disregards the timing perspective of such efforts.
It's clear. There is some movement, but slowly, and it's something for the long term, if it happens. I have the feeling the Americans are much faster, seeing the recent announcements of Intel. It's not only my person, there is also the U.S. government a little bit behind it. Okay, they know more about the industry, so that's a big advantage. All right, thanks a lot.
Thank you, Mr. Ries.
There are no more questions.
Very good. Thank you very much for your participation and your interest in Elmos. I would also like to remind you that we will publish our Q3 results on Thursday, November 4, 2021. Finally, I would like to wish you all.