ams-OSRAM AG

Good morning, everyone. This is Juergen speaking.

Welcome to today's call on first quarter 2026 results. Aldo, our CEO, will comment on business performance and our strategic progress.

Rainer, our CFO, will then walk you through the financials. Please refer to the Q1 earnings call presentation that is available on our website.

Aldo, how did we perform in the first quarter following the launch of our Digital Photonics strategic realignment?

Thank you, Juergen, and good morning, everyone. Turning to Slide 3 here.

Overall, we delivered a very strong first quarter and made further tangible progress towards our ambition of becoming a focused Digital Photonics powerhouse. On a like-for-like basis, our semiconductor core portfolio grew by 9% year-on-year, clearly underlining that the strategic focus is the right one.

Revenue came in well above the midpoint of our guidance range. Adjusted EBITDA reached the upper end.

Design win momentum continues unabated across all end markets. From a Digital Photonics perspective, we achieved 2 important milestones in this quarter.

First, we are in the process of extending our portfolio of optical components that are decided for the system performance of AI-enabled augmented reality smart glasses covering key functional building blocks. Second, in AI photonics, we signed a development agreement for highly parallel micro emitter array-based so-called slow and wide optical interconnects targeting hyperscaler AI data centers.

In parallel, we advanced on execution topics. The simplified transformation program is well underway, and our balance sheet deleveraging plan progressed as planned.

The sale of the Entertainment & Industrial lamps business to Ushio closed in early March and cash proceeds were received. The divestment of our non-optical sensor business to Infineon remains well on track with unchanged timing for mid-'26.

Finally, we delivered positive free cash flow in Q1. As expected, divestment proceeds offset the seasonally high interest payments that typically occur in the first quarter.

With that, let look at the details. Turning to Slide 4.

Q1 performance came in stronger than initially expected. Group revenue came in with EUR 796 million, well within the upper half of the guidance spend.

Adjusted EBITDA reached 16.5% at the upper end of the guidance, driven mainly by the OS division and a very strong automotive lamps performance. Year-on-year, revenues declined slightly, entirely due to the weaker U.S.

dollar with a top line impact of roughly EUR 15 million. On a like-for-like basis, at constant currencies, the group would have grown by approximately 8%.

Adjusted EBITDA recently declined modestly year-on-year solely due to the deconsolidation of the specialty lamp business despite the FX headwinds. Let's turn to segment performance on Slide 5.

OS held up very well in a typically soft first quarter. Revenues were almost flat quarter-on-quarter.

We experienced supply constraints in select product lines due to short-term order increases. Without those, even the sequential growth would have been possible.

Margin declined sequentially due to higher gold prices, annual price downs effective January 1 and FX effects. However, it was 2 percentage points higher year-on-year, reflecting higher production volumes that are not fully visible in reported revenues due to the weaker U.S.

dollar. CSA delivered a solid performance in the seasonally weakest quarter.

Results were driven by continued strong demand for custom sensor products in consumer handheld and a recovery in Industrial & Medical. Revenues were slightly lower year-on-year solely due to declining contribution from exited noncore portfolio activities.

Profitability follows typical revenue fall-through dynamics, however, was down year-on-year, which is due to higher R&D expenses to fund growth projects and FX headwinds on top. Lamps & Systems again delivered a very strong quarter.

Aftermarket demand remained elevated, including short notice ordering following financial difficulties at a major competitor. Specialty Lamps contributed for only 2 months.

Please keep that in mind. The deconsolidation explains why reported revenue did not increase year-on-year.

Strong production loading in Q1 supported profitability. Overall, it was mostly a strong quarter across the portfolio.

Turning to Slide 6. Adjusting for the weaker dollar and the exited noncore portfolio contribution, the clean core portfolio grew 9% year-on-year.

The noncore portfolio is now largely wound down with only residual contribution in the order of magnitude of EUR 10 million. Looking at the markets, Automotive was broadly flat versus a typical seasonal slowdown.

After a lackluster start early into the year, we saw a clear ordering uptick in February and March. Given the declining underlying vehicle production outlook, we interpret it as a partial restocking after a prolonged period of very limited inventories, combined with some level of precaution due to the turbulences in the Middle East.

All regions performed sequentially better, except China, where end market demand remains softer and competitive intensity is elevated. Industrial & Medical showed a clear recovery.

Horticulture had a seasonal low point, but professional lighting demand was solid. Order intake improved materially and order patterns at the end of the quarter point to a solid seasonal upswing into Q2.

Consumer followed typical seasonal patterns sequentially. Year-on-year, the decline is explained by FX and the phaseout of noncore portfolio elements.

Turning to Slide 7. Q1 is typically the weakest quarter for design win activity, yet momentum remains solid.

Total design wins amounted to around EUR 850 million. Naturally design wins are geared towards automotive, but the other verticals also contributed well.

In our classic semiconductor core business, automotive remains the backbone with triple-digit million euro contribution across the portfolio and strong momentum in forward lighting. Industrial showed very good traction, particularly professional lighting with customers in the U.S.

and Europe, while horticulture performed materially better year-on-year. Consumer continued to see recurring sensor design wins in Android-based smartphones, particularly in display management.

On the Digital Photonics side, progress was equally encouraging. EVIYOS continued to add platforms, taking the number of awarded platforms to well above 60.

And interest for new designs remain strong, especially in China. Augmented reality, several of our existing components such as ambient light and spectral sensors are already designed into smart glass models available in the market.

AI photonics, well, product development for micro-emitter arrays for highly parallel AI optical interconnects has started. And we are not doing this alone.

We have signed a collaboration agreement with a strong AI infrastructure partner. We will now look at these Digital Photonics themes in more detail.

Turning to Slide 8. Augmented reality, smart glasses are a key Digital Photonics growth theme.

While the category is still at an early stage, adoption is accelerating even with today's limited functionality. AI is a game changer, making the glasses potentially in the midterm a replacement of our smartphones.

Some of our sensors and LEDs are already designed into several commercially available smart glass models today. Our current and future portfolio covers key functional domains, health and well-being, sensors enabling measurement of parameters such as medicine levels via blue light, heart rate and UV exposure.

Privacy and camera performance, spectral and flicker sensors as well as high-performance LEDs. Display engine, today, our LEDs eliminate LCOS displays.

Going forward, micro LED arrays can enable sustainably higher brightness resolution and better power efficiency. World sensing compromise gesture and 3D Time-of-Flight sensing.

HMI, today, we supply our proven proximity sensors. Tomorrow, we have super tiny optical for sensing buttons in store.

And eye tracking can be done with our integrated optical sensing solutions. This illustrates our strategy of focusing on the size of system components built on our core technologies.

Content estimates naturally vary depending on volumes, life cycle stage and computer -- customer implementation choices. For this, however, we see content potential between EUR 50 and EUR 100 per device, which underpins the triple-digit million annual revenue opportunities we outlined when launching our Digital Photonics strategy.

On to the next highlight today, turning to Slide 9. Our progress in AI photonics is accelerating.

I have 3 slides for you. First, where our products will sit in the data center; second, how do we fit in our structure; and third, which components are we targeting.

We believe that the so-called slow and wide optical interconnect based on highly parallel micro-emitter arrays can play an important role in future AI data center architectures. And here, slow is relative as we're talking about 8 gigabit switching speed and hundreds of parallel channels.

Initially, the focus is on short distance scale-out interconnects achieved between the racks, then scale-up connections within the rack, replacing copper over distances of up to several tens of meters. Over time, chip-to-chip connections, for example, between GPU and high-bandwidth memory could become addressable as well, a really great market potential for us.

Turning to Slide 10. It's important to distinguish between integration content and the optical engine technology itself.

On the integration side, today's solutions on the upper right rely on pluggable transceivers or active optical cables with energy consumption of up to 30 picojoules per bit. And these solutions, not only longer -- the long copper traces, but typically also signal shaping chips consume quite a lot of power.

ToF sensor near port optics can reduce this to roughly 5 to 10 picojoules per bit. The optical engine moves much closer to the ASIC.

Co-packaged optics shown on top left, promises further reductions towards 1 to 5 picojoules per bit over time. The optical engine moves as close as possible to the ASIC.

Put simply, the closer the optical engine sits to the chip, the lower the electrical losses and the associated thermal load. The slide illustrates this distance comparisons.

Independent of the integration approach, optical engines can be implemented either as fast and narrow or slow and wide. Fast and narrow is today's established technology based on indium phosphide lasers, often EMLs and silicon photonics integration concepts.

We believe in future slow and wide architectures, highly parallel micro-emitter array-based optical engines that transmit light pulses at chip speed without need for power hungry serializers and deserializers. Key advantages include substantially higher bandwidth density, very low power consumption per bit and inherent redundancy through parallelism.

If one micro-emitter fails, no problem, there are enough channels for backup, an important consideration for hyperscale customers. Turning to Slide 11.

On the left, you see our prototype, which helped accelerate the signing of a development agreement with our ecosystem partner, a leading AI infrastructure supplier. The table in the center illustrates the simplified technology stack for highly parallel optical interconnects.

In essence, you can think of the transmitter side, the receiver side and advanced packaging technology that glues everything together. Our current development focus is on the transmitting side, micro-lens and micro-emitter arrays.

Given our CMOS and sensor capabilities, we're also evaluating opportunities on the receiver side. We'll keep you updated as development progresses.

With that, let me hand over to Rainer for an update on selected financial aspects.

Thank you, Aldo. Good morning, everyone.

I'm on Slide 12. We generated EUR 37 million free cash flow in Q1, which includes EUR 90 million divestment proceeds.

The cash inflow from the sale of the specialty lamps was received early March. Operating cash flow was a breakeven, reflecting the seasonally high interest payments on our senior notes.

Higher than a year ago after the EUR 500 million tap we did last summer. CapEx remained disciplined and well below our full year guidance of 8% of revenue.

With that, let us take a quick look at the Simplify program that we launched with Q4 announcement in February 10. Turning to Slide 13.

Last quarter, we reported that the Re-establish the Base delivered savings 1 year early. The implementation of the remaining measures that had been identified continued.

Re-establish the Base program delivered EUR 237 million savings, really a great success. Now in February, we launched a successor Simplify program, which is a broader transformation program aimed at reshaping our operating model and delivering another EUR 200 million of additional annual savings by '28.

Happy to announce that all saving measures have been identified and at least 90% of those have already today a high maturity level. Cost, speed, agility are our guiding principle as we reshape our operating model.

Implementation started immediately. And after just 1 quarter, the teams have already delivered EUR 5 million of savings, demonstrating disciplined execution continuity.

Now, let's have a quick look at liquidity and capital structure on Page 14. In Q1, the interest payments for senior notes were due with the cash proceeds from the sale of the specialty lamps, free cash flow was positive such that our cash on hand position only reduced because we paid back EUR 200 million nominal of the convertible note.

And the cash position now stands at EUR 1.3 billion at quarter end. With that, the available liquidity position closed accordingly at EUR 2 billion.

It is backed by a diversified mix of instruments, cash revolver and bilateral lines. The sale and leaseback value moved up in line with currency swings and the quarterly interest accrual.

And with that, let us zoom in on the coverage of the upcoming short-term maturities on the next slide, which is Page 15. Now we have EUR 1.3 billion cash on hand, and that will be enriched with EUR 570 million from the Infineon deal upon closing somewhere midyear.

That gets us to a total of EUR 1.9 billion pro forma cash, and that completely recovers all the near-term maturities, and that is the remaining outstanding convertible bond, which as we paid EUR 200 million is now sitting at EUR 560 million. When we received the money from Infineon, we have 120 days to offer the amount related to the guarantor assets as part to noteholders approximately EUR 130 million.

And second, there are some business needs for the transition effects in '26. Basically saying the cash flow will be negative.

The EBITDA will be somewhat lower because we are selling business, and there will be some stranded costs that we will be cleaning up. And then there's quite a bit of transformation costs and talking cash flow, the cash outflows from the Simplify program.

And then we will be repaying USD 100 million in customer prepayments, and we will reduce because we have so much cash, roughly [ USD 100 million ] in factoring. Now excluding the disposal proceeds, expect the cash flow to be something triple-digit million negative.

However, and once we are through that in '27, the free cash flow will be substantially better. And if business remains strong as it is, we expect it to move to positive territory, excluding any additional disposal proceeds and even that we also next year, still have to repay a similar amount of customer prepayments.

So excluding disposal proceeds, we expect that next year to be in positive territory. And third, what we will be paying from the existing cash on hand is the tendering of the Osram minority shares after the final verdict, there's no news, but we still have it -- assume that it will come this year.

And that should then leave once everything is taken care of at least EUR 500 million cash on hand. And that is the very important point.

All upcoming near-term maturities are fully covered. And that obviously then now that is covered, we now have started to focus conceptually on optimizing the cost and the maturity profile of our '29 senior notes, where, as you know, the interest rates are higher than I would love to, and we will keep you posted with what our plans are.

And with that, let me hand back to Aldo for the summary and the outlook.

Thanks, Rainer. And let me summarize today's call.

I'm on Slide 16. In Q1, we beat again our revenue and profitability guidance.

The core semiconductor business grew 9% like-for-like. Free cash flow was positive at EUR 37 million.

We completed Re-establish the Base with EUR 237 million savings a year early and started executing Simplify. In Digital Photonics, we continue to progress on the comprehensive component portfolio for AI-enabled smart glasses, giving us a content opportunity between EUR 50 and EUR 100 per smart glass.

We initiated the product development of micro-emitter array-based AI optical interconnects together with the commercialization partner. We also progressed in balance sheet deleveraging the Specialty Lamp transaction closed and proceeds were received and the Infineon transaction remains on track.

No changes to the indicated closing timeline of mid of this year. Now to the outlook for the second quarter.

We expect revenues between EUR 725 million and EUR 825 million, with adjusted EBITDA around 15.5% plus/minus 1.5 percentage points based on euro-dollar of 1.17. The traditional auto lamps business will show the usual seasonal slowdown in view of the overall rate in the aftermarket.

Remember, all non-automotive business was transferred to Ushio. We still have EUR 10 million revenue in Q1 and 0 in Q2, obviously.

Semis will make a step forward in Q2 more than typical seasonality. We see strong order intake and book-to-bill higher than previous quarters.

Our full year 2026 outlook remains unchanged at the moment. Group revenues modestly softer given the divestment and FX.

Adjusted EBITDA of around 15.5%, plus/minus 1.5 percentage points, assuming euro-dollar of 1.17. Adjusted EBITDA will be negatively impacted by several one-offs, the divestments, stranded costs, precious metal prices and other factors.

Free cash flow, we expect above EUR 300 million, including divestment proceeds. Into 2027, we see a path to positive free cash flow without any divestment proceeds even with repaying a similar amount of customer prepayments.

And with that, we are through the presentation, and we're happy to take your questions.

The first one is just on the micro LED optical networking opportunity. I think previously, you said it's kind of a 2030 plus revenue opportunity for you.

But given the announcement today and then also kind of industry announcements about commercializing solutions as early as 2027. How are you thinking about the timelines on when you could see revenues?

Yes. Thanks, Harry.

That's a very good question. The industry is hyper quick and really waiting for solutions here.

We're working very hard on those solutions together with our partners. And I think timing-wise, we are definitely before 2030.

Whether it's as early as 2027, I can't tell you yet, but we definitely see this as an opportunity in the next few years, not in the far future.

Okay. Great.

And just maybe a follow-up on that. It was good to see that you're addressing kind of multiple elements of the full solution.

If you look at kind of the elements that you have in development and consideration, so the emitter, the lens and the diode array, how -- are you able to give kind of like a rough split of like what your -- like what percent -- so yes, the relative kind of content size for you of each of those? What would be the -- yes, what would be the biggest content for you?

Well, let me answer it this way. I think we have said that by 2030, we see this as a triple million dollar opportunity, and we continue to see that at least.

And that was a statement based on just the micro-emitter array and the micro lens array on top. So if we actually add more to that, if we also do the photodiodes or the amplifier or the driver underneath the LED, that would be further revenue potential for the group.

Super helpful. And then just maybe a quick one on the short term.

The guide for Q2 of typical seasonal decline in the semiconductor business, given the portfolio changes over the years, it's kind of tough for us to know what exactly seasonal is. Wondering whether we could get a little bit of help on that.

I think automotive will do quite well in the second quarter. Also horticulture will start to pick up ahead of the season.

So those are positives. And also with the little precaution that we also mentioned in the text before that at the moment, we are still a bit uncertain on how to interpret this good order intake at the moment.

The number of cars, of course, are not going up, it's going down probably being built globally. Yes, content per vehicle is rising, but the order intake at the moment is quite strong.

So we do think that it's part of restocking and not taking risk out of the supply chain given the global uncertainties. But at the moment, encouraging, and we hope it continues that way.

Obviously, consumer goes down quarter-on-quarter, that is normal and then Q3, of course, being usually the strongest quarter for consumer. Professional lighting, pretty steady, I would say, and lamps, automotive lamps, in principle down.

I mean the lighting season is over. Obviously, that is always Q4, Q1.

So Q2 will be lower there. But here, we benefit from the financial troubles of our major competitor and also see here additional orders coming in that might limit the negativity of the normal seasonality in this.

You mentioned in the press release some tough competition in the automotive business in China. At the same time, we are seeing capacity getting a bit tighter in many areas of the market and some peers raising prices.

I'm just curious about how do you see prices trending at AMS for the coming months? And the second one is on the free cash flow.

You mentioned for this year free cash flow to be significantly negative excluding divestments. Could you please help us quantify the kind of cash burn we can expect for this year?

Do you have some building blocks to understand a little bit the magnitude of the range of negative free cash flow we can expect for 2026?

Yes. On the pricing side, I would say on the one hand, competitive pressure in China is increasing.

Our customers are under a lot of pressure. Vehicle volumes are going down, so that puts them even more under pressure.

And of course, they're looking to the supply chain as well to contribute. And that has somewhat limited our ability to pass on the cost increases that we also spoke about last quarter.

At the moment, we see in the LED space, neither from us nor from others price increases, but we do see more limited price declines than we otherwise would have seen. So kind of an indirect effect that you see there.

In the sensor business, actually, we are selectively raising prices like others as well. And also in the automotive lamp business, we are selectively raising prices for increased input costs.

So that's kind of the overall pricing dynamic, I would say. Towards China, perhaps important.

The remark I made before on the vials, I think that's really nice to see. Yes, of course, on the more standard parts, there is competitive pressure.

However, China is also now really an innovation-driven market for us and our innovations like these high-pixelated headlamps are really designed in much more widely also in China now, and that gives us good hope that we will continue to see a strong market position for us in China. On top of that, by the way, we, of course, are also grabbing share, mainly from our international competitors.

You probably know that the Samsung has exited the market. We've got quite a bit of share there.

Also, other American competitors are struggling, and that helps us to expand our position. In that sense, we look with confidence in the future.

But it's -- yes, at the same time, not an easy market.

Yes. And Sebastien, on the cash flow.

So maybe if you try to build a bridge compared to last year, first of all, I mean, the public funding will be significantly less. Last year, we got like 2 tranches.

This year, we only get one. Number two, we're repaying customer prepayments, which is roughly $100 million or EUR 80 million.

Number three, from the new Simplify savings program, where we want to achieve EUR 200 million of savings. I mean that also comes with a onetime cost of more than EUR 100 million.

I mean trying to pay as much as possible of that this year, so to have as much as possible behind us. And finally, then the factoring, which I mean cost a lot of money also given the high cash position, we want to reduce that.

Let's say the EUR 100 million, and that is obviously also then an element of the free cash flow. Then when it comes to the business, yes, the EBITDA will be down, and that also certainly then ends in a bit lower cash flow.

That is from the divested businesses, right? We said that the divestment of those businesses plus the stranded cost would be on an annualized basis, something like around EUR 75 million.

The stranded costs we are obviously working on. But then on the positive side, as Aldo pointed out, the business is certainly improving.

I think the guidance for Q2 is quite optimistic if you compare it to the last year Q1 to Q2 bridge, right? So business is running well.

We are getting NREs from customers. So there's also a lot of positives in there.

But I mean, if you count it together, free cash flow will be significantly negative this year.

I just have -- I have 2 questions, one on your AI-related photonics business and one on your AR glasses. On the AI side, on your press release, you -- on the front page, you had said that you have signed a development agreement with a leading AI data center infrastructure partner.

And on the second page, you've said that you've signed it with a leading AI photonics industry partner. I'm just a bit confused because when I hear the word AI data center infrastructure partner, I sort of think of a hyperscaler.

And when I think of an AI photonics partner, I think of someone like a Lumentum or maybe even [ CNM ] or someone like that. So can you give us some clarity on what exactly is the nature of this partner that you have got a development agreement with?

And secondly, in a development agreement, is it that they pay you money and you do all the development? Or is it that you're sort of sitting in their premises as well or they're doing half the development, you're doing half the development.

Just an understanding on how that works. And I have a follow-up on the AR glasses.

Yes. To start with the second part of your question.

I mean all these systems are still so new that it requires development on both sides. I mean we have to optimize and design our part of the system and develop that to the right performance and reliability levels.

At the same time, it needs to be well integrated into the larger solution to really be effective. So in that sense, both parties have to contribute here and be in close alignment.

And that's why in these new fields, it's so important that you find lead customers that you can together design and optimize the systems with that just tremendously increases your likelihood of success and your ability to be fast. Much more detail on the kind of partner that we have, we cannot give.

But obviously, it is about photonics. We are about photonics.

It's clear that the data rates and the energy they require is a major topic. So anything we can do to optimize that, this is highly welcome, and we expect also then this technology to scale with other partners as well in the future.

Yes. So you can't give a nudge on whether this partner is someone who builds the data center.

Sorry.

Okay. And then on the AR glasses, you're saying that you already are shipping some of these components into the market right now.

So I'm just wondering what needs to change when you are going to get that triple-digit million euro kind of revenue number? Is it that you need a big customer to start volume production with your components and today, you're selling to smaller customers?

Or is it that you need the microLED array itself to start shipping because that might be a high-value component in contrast to LCOS trajectory or something like that? I mean what is the change that we're waiting for to get that triple-digit million euro?

Yes. Both factors will kick in.

We will -- for the LCOS, we supply some high-power LEDs. That's nice, but it's not a major value driver yet.

The move to microLED-based engines will be the major step-up for us. And of course, I mean, these glasses are now in being utilized more and more, and there will be a significant demand increase, we feel with next-generation smart glasses.

They just are getting better and better from generation to generation, more and more attractive. So we expect the volumes to go up significantly.

And at the moment, the majority of AI glasses is sold without a display. Now, our first simple displays with LCOS are being sold with the microLED, we think that the glasses will gain a lot of more momentum with picture quality being better, bigger field of use and so on, smarter designs.

So we think this category will do quite well. And with the rising volumes and a higher content per phone, that will help.

However, at the same time, I think also beyond that, there's interesting potential as we outlined. It is not only the light engine.

Also other parts we feel we are quite well positioned in. And they, of course, will also then scale with more customers and higher volumes of the sales of the smart glasses in total.

And again, any kind of time horizon for the triple digit? Is that -- is there a chance that, that could happen in '27?

Or are we looking beyond that?

I can't really comment on that because as you can understand, our customers are always quite tight on timing of their launches. But it's going to be good this decade.

It's definitely closer than the AI.

Yes.

My first one was just a reminder, please, on your smartphone exposure. It seems to be an end market that could be challenged in '26, maybe '27 as well.

Just if you could give your revenue exposure to smartphone and then how that is trending? And then my second question, just on the Kulim-2 sale process and just any update there?

It seems like the industry is scrambling for capacity, and this seems like it could be an interesting asset for a buyer. But if you're able to give any update on the process, that would be helpful.

Yes. If you -- on the smartphone side, we're probably talking EUR 600 million, EUR 700 million revenue that is tied to that space.

We have to now at the moment, look very much of who is getting memory and who isn't. And you can imagine that the top cell phone makers that we often bet with are the ones that are getting the memory and the lower and midrange phones get less of it or struggle more with the prices that memory at the moment commands.

So yes, I agree with you, there might be somewhat of an impact. But so far, we're not really seeing it given the positive customer mix in this aspect.

On Kulim-2, yes, nothing really changed. We keep having inbounds.

We keep having good discussions, but we also cannot say that the timing is around the corner. So yes, we'll inform you as soon as something gets more tangible.

But yes, at the moment, no real change to versus what we've said last quarter.

And can I just -- sorry, a brief follow-up, but just can I ask on the Q2 guidance, whether that includes any divestment from the Infineon disposal? Or is that only impacting Q3?

Yes. Craig, the guidance assumes the closing of the transaction midyear.

So the impact of the disposal would then be seen in Q3. The Q2 guidance, though includes the disposal of the traditional lamps business, right?

So in a way, if you subtract the revenue that we are losing from the traditional business and you look at midpoint compared to Q1, you could say that it's more or less flat.

Maybe just a question on the AI opportunity. I mean, obviously, there's a lot of development happening on co-packaged optics at the moment.

I would say most of those solutions are relatively mature, so probably a bit more insertion for you guys. Is the idea to -- in the CPO side to insert in a later version?

Or is it more as we transition to optical IO that you're going to get sort of broad-based adoption? And then just secondarily, in terms of your thoughts about triple-digit million by 2030, is that based on a couple of hyperscalers that are clearly championing microLED?

Or is it based on a portion of the indium phosphide market that you think you're going to penetrate?

Well, I think the overall AI opportunity is huge. And there's a lot of, I would say, sub applications or spaces where optical technology will be helpful in getting higher bandwidth at lower energy consumption.

Obviously, CPO and indium phosphide lasers are the coming waves that are heavily being invested in. However, we feel that not all of the system -- not the whole system architecture can be optimally addressed with that.

So there will be significant pockets that we feel this technology that we are developing will find a place besides or beyond fast and narrow. So it's -- the estimate of the revenue contribution really comes from us finding the right spots in this overall market.

And the customer discussions show that clearly that they also see that indium phosphide is a good solution, but there are places where narrow and wide has a limit and where wide and fast actually can play out the advantages. And especially, I think in the longer term, the chip-to-chip communication, I think this is definitely a space where slow and wide can do a lot of good.

And would you be open to a kind of strategic relationship like you had with your microLED relationship in the past on the consumer side? Or is this something you just want to sell lots of merchant LEDs to different players like Avicena and others that are creating the module?

No, we want -- we clearly want those relationships ourselves. Besides my private life, I'm very open for multiple relations.

And here, we -- I think we'll see that different -- first of all, it's important that we find the right lead customer for the right type of application. In the AI space, I think we will see different parts of the application being potentially spearheaded by different people.

So also in that case, even within AI, I can imagine that at the end of the day, we will have one with several partners to develop specific solutions for specific parts of the whole chain. Once we normally develop with our partners, then of course, we first want to fully support them and scale with them.

But usually, we have a common interest to scale beyond that first partner because that brings overall volumes up and cost down, which is then good for everybody. So that's kind of the thought model that we have been using already in the early days in automotive, are now using it in the AR opportunities, and we'll do so also with the opportunities in AI in a similar fashion.

Yes. Thank you, speaker.

I'd like to thank everyone for dialing-in today. Thanks to the analysts for your questions.

And as always, we are ready for follow-on questions from the Investor Relations team, just reach out to us. And with that, we wish you a great rest of the week and speaking to you soon or latest in the next quarter.

Thank you, and goodbye.