Good morning, everyone. Welcome to the conference call for analysts and investors for Infineon's 2025 fiscal fourth quarter and full year results. Today's call will be hosted by Alexander Foltin, Executive Vice President, Finance, Treasury, and Investor Relations at Infineon Technologies. As a reminder, this call is being recorded. This conference call contains forward-looking statements and/or assessments about the business, financial condition, performance, and strategy of the Infineon Group. These statements and/or assessments are based on assumptions and management expectations resting upon currently available information and present estimates. They are subject to a multitude of uncertainties and risks, many of which are partially or entirely beyond Infineon's control. Infineon's actual business development, financial condition, performance, and strategy may therefore differ materially from what is discussed in this conference call. Beyond disclosure requirements stipulated by law, Infineon does not undertake any obligation to update forward-looking statements.
At this time, I'd like to turn the call over to Infineon. Please go ahead.
Good morning, ladies and gentlemen. Thank you for tuning in to our earnings call for the fourth quarter and the full fiscal year 2025. On this analog bellwether call, you have our CEO, Jochen Hanebeck, our CFO, Sven Schneider, and our CMO, Andreas Urschitz. Jochen and Sven will provide an overview on the market situation and divisional performance, key financials, and, of course, our long-awaited outlook for fiscal 2026. Our prepared remarks will also cover step-up, sustainability achievements, and the dividend proposal. After that, we will start our Q&A session. As usual, the illustrating slideshow is available at infineon.com/slides, and we will provide a PDF with Jochen's and Sven's introductory remarks in the course of the call on our website, namely infineon.com/investor.
As you know, this is your go-to place for a recording of this conference call, including the aforementioned slides, a copy of our earnings press release, as well as our investor presentation. At this point in time, Jochen, over to you.
Thank you, Alexander, and good morning, everyone. The turn of a fiscal year always calls for some look-backs as well as future perspectives. In each case, cyclical factors need to be distinguished from structural ones. The successfully concluded 2025 fiscal year formed part of a prolonged down cycle in most of our target markets. End customers and channel partners have undertaken a major destocking exercise intending to reach target levels. Geopolitical instability and tariff turmoil made them cautious on the direction of end demand and caused short-term ordering behavior. In this environment, we saw, as predicted, a slight annual revenue decline driven mostly by negative currency effects. We managed what we could control and maintained margins at a resilient level, supported by first meaningful benefits from our structural improvement program step-up, coming in ahead of the anticipated timeline.
The other structural initiative we are driving is to accelerate our innovation-to-customer value to strengthen Infineon's position in secular growth areas like software-defined vehicles or AI data centers. Taking these initiatives together, we will benefit from a more competitive setup and an unmatched portfolio in any return-to-growth scenario. In the just-started fiscal year 2026, such growth will be tempered by adverse currency movements. Furthermore, growth will be a function of the intensity and breadth of the recovery, hard to predict due to ongoing geopolitical and tariff-related uncertainties. Therefore, we are once again setting a prudent bar with our full-year outlook. From a more mid to long-term perspective, we are in an excellent position to continue to lead in our markets and nurture profitable growth based on our well-known structural growth drivers. Before looking ahead, though, let's take a look in the rear mirror.
Revenues in the September quarter came in at EUR 3,943 million, making the fourth and final quarter the strongest of our 2025 fiscal year. Sequential growth was around 6% once again, including a negative currency effect as the U.S. dollar/euro exchange rate weakened to $ 1.17 versus EUR 1.14 for the June quarter. At constant currencies, quarterly growth would have amounted to 8.5%. All segments contributed with positive revenue developments, in particular PSS. Furthermore, our Q4 showed a slight year-over-year growth at constant currencies, even of 5%. This is the first time within eight quarters. Revenues for the entire 2025 fiscal year amounted to EUR 14,662 million, 2% down from the previous year. At constant currencies, our revenue would have been essentially flat year-over-year, given that the U.S. dollar average exchange rate weakened from $1.09 in fiscal 2024 to $1.11 in fiscal 2025.
This is quite remarkable for a year characterized by substantial inventory corrections by customers and unprecedented tariff disputes. Using the product categories I explained in the last call, our full-year revenue for fiscal 2025 breaks down as follows. Around 35% relates to power discretes and modules covering silicon, silicon carbide, and gallium nitride. Around 30% pertains to analog and sensors. Finally, control and connectivity grew to around 35%, supported by the success of our automotive microcontrollers, which will be further strengthened by the acquisition of the automotive Ethernet business from Marvell. This balanced portfolio is illustrating the healthy degree of diversification we have achieved over the past few years. Regarding our operating profitability in the September quarter, the segment result amounted to EUR 717 million, corresponding to a segment result margin of 18.2%.
The slight improvement compared to the previous quarter is reflecting volume growth, which overcompensated the adverse currency development as well as some internally incurred opportunistic low-margin business for consumer-related products to avoid idle cost. More on this in the PSS section. For the full 2025 fiscal year, the segment result margin was 17.5%, landing in the high teens territory as predicted, in line with a lower range of our target operating model. Compared to one year earlier, annual price declines, a negative currency impact, and rising idle costs could be offset in part by contributions from our step-up program. Our order backlog increased by around EUR 2 billion, quarter- over- quarter, to close to EUR 20 billion at the end of September, a first proof of the recovery materializing. Now to our divisional review, beginning with automotive.
In the final quarter of 2025 fiscal year, the automotive segment achieved revenues of EUR 1,921 million, a further uptick compared to the previous quarter. The most notable volume growth contributors were smart power components, microcontrollers, and XEV-related solutions. The latter we ascribe to temporary pull-ins due to subsidy reductions in the U.S. and in China. Both the segment result of EUR 430 million and the segment result margin of 22.4% increased sequentially, driven by volume growth and improved mix, as well as some smaller favorable non-recurring effects. Independent of near-term market developments, we continue to shape the future of mobility with our leading product portfolio across power, analog and sensors, and control and connectivity, based on our P2S Approach. A few recent design wins examples underscore this. An established North American OEM will use AURIX TC4 Microcontrollers, along with several analog components for an upcoming ADA system.
The total design win volume is a mid-triple-digit million euro amount, covering not only the MCU but also power management ICs, PROFET smart switches, and NOR Flash memory. Our newly acquired BrightLane Ethernet product family is very well received by customers. They will be a crucial building block for software-defined vehicles and, combined with our MCU portfolio, enable even more comprehensive solutions for our customers. As an example, they are currently ramping in the most recent platforms of two European premium OEMs. Lastly, we have secured an additional cumulative triple-digit million euro design for our OptiMOS 7 MOSFETs across all regions. These cover key applications such as power distribution in several upcoming software-defined vehicle architectures, as well as safety-critical applications such as steer-by-wire systems, braking systems, and active suspension. This success is underpinned by our differentiating technology, which enables leading-edge devices with superior quality, highly appreciated by customers.
Let's now move to green industrial power, where revenues in the September quarter came in at EUR 463 million, an increase of 7% quarter-over-quarter. The sequential improvement was strongest in the areas of power infrastructure, comprising renewable energy generation and grid infrastructure, as well as rail systems and commercial electric vehicles. On the back of higher revenues, GIP segment result edged up to EUR 69 million in the fourth quarter of our 2025 fiscal year, equivalent to a segment result margin of 14.9%. Besides the typical seasonal pattern, the market situation for the various industrial applications we serve shows a mixed picture. Macro uncertainty is prolonging the path to recovery for automation and end drives. Similarly, there are not yet clear signals for an upswing in heating, ventilation, air conditioning, or home appliances. Regarding renewable energy generation, market conditions throughout the delivery chain continue to show some signs of weakness.
That said, we see structural drivers on the power infrastructure side getting stronger. A higher share of renewables in the overall energy mix and the proliferation of AI data centers at gigawatt levels in various parts of the world cause the need to significantly upgrade and strengthen the power grid. Critical applications providing attractive evolving content opportunities for us are, for example, large-scale energy storage systems, uninterruptible power supplies, and solid-state transformers and circuit breakers. To advance the latter, we are partnering, amongst others, with SolarEdge to develop highly efficient next-generation solid-state transformer technology for AI and hyperscaler data centers. The collaboration focuses on combining advanced silicon carbide technology from us, Infineon, with SolarEdge power conversion and control topology to enable direct medium voltage conversion. The solid-state transformer technology will play a crucial role in future 800-volt direct current AI data center power architectures.
The technology enables highest end-to-end efficiency and offers several key advantages, including a significant reduction of weight and size, a reduced CO2 footprint, and accelerated deployment of power distribution, among others, when connecting the public grid with data center power distribution. The power and sensor segments recorded revenue of EUR 1,189 million in the September quarter, 13% up sequentially. The main contributors to this significant growth were our power solutions for AI servers. In addition, we saw strength for products going into smartphones and accessories like MEMS microphones. Notwithstanding the strong revenue pickup, the segment result of PSS decreased to EUR 179 million, leading to a segment result margin of 15.1%. Besides adverse exchange rate effects, the main factor were so-called FAP fillers, intentionally addressing low-margin market segments on the consumer side to utilize otherwise idle production capacity.
We plan to fade out such business in the near term, being replaced more and more by strongly going margin-accretive AI volumes. On the market side, consumer and general compute communication continue to see a tepid recovery. In stark contrast, the AI boom remains a powerful growth engine for us, with data center buildouts accelerating dynamically. Managing the power flow in line with the exponential compute growth across all power conversion stages from grid to core becomes mission-critical to scaling AI. Said differently, there is no AI without power. With the unmatched breadth of our product portfolio in silicon, silicon carbide, gallium nitride, various package concepts, speed of innovation, quality, and delivery capability, we have become the leading partner of all relevant GPU and ASIC providers. This success is clearly visible in our numbers.
In the 2025 fiscal year, our AI data center-related revenue nearly tripled to more than EUR 700 million. We thus were able to beat our originally predicted numbers of around EUR 600 million, despite the adverse currency development. For fiscal 26, we are raising our growth expectations significantly, from around EUR 1 billion to around EUR 1.5 billion, more than doubling year over year. The strong momentum of our business is coming from the unabated market momentum for AI, strongly rising power requirements of new processor generations, reconfiguration, and data center architecture, as well as expected further content gains. By the end of the decade, we expect the addressable market for us to be in the range of EUR 8 billion-EUR 12 billion, depending on factors like module share, customer structure, and speed of AI buildout.
Combining the expertise and unrivaled product portfolio of our GIP and PSS segments is putting us at the forefront of both the energy transition and the AI revolution. This creates highly attractive idiosyncratic business opportunities for us. More on these in an investor deep dive that our two division heads of GIP and PSS will host on 26th of November. On our investor relations homepage, you will find details about this event. Now let me touch on wide band gap. Our unique setup of having the world's most innovative, cost-efficient, and scalable in-house manufacturing of silicon carbide and gallium nitride is fully recognized by automotive, industrial, and AI data center customers. Talking about the latter, power systems of 800-volt AI server power supplies are only conceivable with best-in-class silicon, silicon carbide, gallium nitride, and package product combinations. In the overall silicon carbide market, supply is currently outstripping demand.
In this competitive environment, we could keep our silicon carbide revenues in the 2025 fiscal year at the level of the previous year, at around EUR 650 million, despite headwinds from pricing and currency. Given the favorable design win trend, amongst others linked to rising usage in the aforementioned AI data centers, we anticipate again growth for our 2026 fiscal year. Gallium nitride is at an earlier stage of the market uptake, and system topology changes are necessary to reap the full benefits of GaN. We are encouraged by an increasing design win momentum at automotive, industrial, and data center customers, which value our power systems expertise combined with in-house manufacturing. Looking ahead, GaN is poised to play a pivotal role in areas like automotive, AI servers, and robotics. Let's complete the divisional review with Connected Secure Systems. CSS recorded quarterly revenues of EUR 369 million, 6% up from the June quarter.
The main driver were payment solutions, in part due to the fulfillment of CIA orders. The segment result of CSS improved to EUR 45 million, corresponding to a segment result margin of 12.2%. Macroeconomic risk continued to dampen consumer confidence in corporate spending, causing sluggish demand for IoT and security solutions. Looking beyond the near term, we continue to invest into innovation to foster profitable mid-term growth. To further strengthen our position in edge AI application, we introduced DEEPCRAFT AI Suite, optimized for our PSOC Edge Family of microcontrollers. To unlock the full potential of edge AI, we provide a comprehensive set of hardware and software solutions to seamlessly integrate AI and machine learning capabilities in the next generation of IoT edge devices. Our solutions empower customers to either develop their models from scratch or to integrate off-the-shelf models and solutions into their products, thereby shortening time to market.
AI in its various shapes and forms will bring significant business opportunities to Infineon across all our segments. Beyond AI, we are actively helping shape the quantum area. Based on our innovation strengths and excellence in volume manufacturing, we have developed trapped ion quantum processing units, so-called QPUs. Together with strategic partners such as IonQ and Quantinuum, we are focusing on scaling qubit counts and fidelity. Quantum and AI are complementary. AI enhances calibration, control, error mitigation of quantum systems, while quantum computing can generate high-precision data sets that accelerate AI-driven discovery, for example, in materials, pharmaceuticals, and logistics optimization. In parallel, we are winning business at customers, preparing for the security implications of the quantum age with post-quantum cryptography-enabled products, including our common criteria certified implementation on a security controller, helping safeguard today's data against tomorrow's threats. Now over to Sven, who will comment on our key financial figures.
Thank you, Jochen, and good morning, everyone. In my part, I will mostly focus on the September quarter, occasionally adding comments about full-year numbers. Let's start with our gross margin development. The adjusted gross margin for the final quarter of our 2025 fiscal year came in at 40.7% after 43% in the quarter before. The reported gross margin decreased quarter-over-quarter from 40.9% to 38.1%. The temporary FAP fillers on the consumer side of PSS that Jochen mentioned burdened margin levels by around one percentage point. Furthermore, mixed effects, the weaker U.S. dollar, and slightly higher idle costs were headwinds quarter-over-quarter. As you know, the background was our goal to manage on-books inventory, as previously indicated. With a DIO figure of 153 days as per the end of September, we have achieved our target for the end of the fiscal 2025.
Keeping stock levels under control will remain a focus area of our cycle management. However, the uneven recovery will stand in the way of further near-term inventory reductions. For the full fiscal year, which was the second consecutive year with a declining top line, we managed to keep the adjusted gross margin well above the 40% mark at 41.4%. Therein, idle costs amounted to close to EUR 1 billion, the vast majority of them reflecting cyclical underutilization. These correspond to around 600 basis points of margin potential without considering the fall-through from additional volumes. On the positive side, we have been reaping first material benefits from our step-up program, focusing on improving our structural cost competitiveness. In fiscal 2025, step-up has contributed about 200 basis points to our gross margin. Now turning to the OpEx side, looking first at reported quarterly numbers.
Research and development remained practically flat at EUR 565 million in the September quarter. Also, our selling, general, and administrative expenses moved sideways to EUR 401 million. On an annual basis, and net of non-segment result charges, R&D as a percentage of revenues was around 15%. Bringing out innovation to address customer issues is key to our differentiation and hence value creation. The SG&A percentage of revenues net of non-segment result charges was just under 10% in our 2025 fiscal year. Our step-up program entails also structural improvements of our operating expenses, and here we see positive contributions coming through nicely too. Overall, with step-up, we have so far recorded about half of the total targeted impact for the first half of our 2027 fiscal year, when all the measures will become fully effective.
Non-segment result charges for the fourth quarter amounted to EUR 263 million, bringing the total for the 2025 fiscal year to EUR 1.045 billion. The biggest component therein was acquisition-related amortization with around EUR 400 million. Other major parts were impairment and other charges related to the sale of our Austin manufacturing site and step-up-related one-timers. The financial result for the September quarter amounted to EUR 64 million after EUR 40 million in the quarter before. Contained therein are, among others, funding costs related to the debt-financed acquisition of the Ethernet business from Marvell, as well as non-cash interest expenses on uncertain tax positions. Income tax expense for the September quarter amounted to EUR 152 million after EUR 95 million in the quarter before. Valuation effects related to deferred taxes have led to a noticeably quarterly increase and influenced the effective tax rate for the entire 2025 fiscal year, which came in at around 27%.
Cash taxes for our fourth fiscal quarter were EUR 106 million. Adjusting for PPA effects, the quarterly cash tax rate stood at 21%. Going forward, for modeling purposes, a tax rate between 20% and 25% is and continues to be a reasonable assumption. Our investments into property, plant and equipment, other intangible assets, and capitalized development costs increased slightly quarter-over-quarter from EUR 442 to EUR 451 million. For the entire fiscal year 2025, investments amounted to around EUR 2.1 billion, coming in slightly below our guidance. Depreciation and amortization expenses, including acquisition-related non-segment result effects for the September quarter, were EUR 484 million, leading to an annual total for fiscal 2025 of EUR 1.9 billion, as predicted. Our free cash flow in the September quarter was significantly influenced by the closing of the acquisition of Marvell's Ethernet business. The reported number of -EUR 1.276 billion would have been EUR 904 million without the acquisition.
The strong quarterly improvement of our organic free cash flow was positively influenced by increased volumes, the predicted receipt of public fundings, in particular for the ramp of module four at our Dresden site, as well as the quarterly reduction of our own inventories. For the full year, the free cash flow amounted to -EUR 1,051 billion. The adjusted free cash flow, considering acquisitions and expenditures for large front-end buildings, in particular our analog mix signal power fab in Dresden, came in at EUR 1.8 billion, corresponding to 12.3% of sales, fully in line with our target operating model. The Marvell Ethernet acquisition also had a noticeable impact on our liquidity and leverage figures. At closing, we drew down the committed acquisition facility from our banks, consisting of a EUR 1 billion and a EUR 1 billion U.S. dollar tranche. The remaining purchase price was paid from our own liquidity.
As a result of this and the strong organic free cash flow in fiscal Q4, our gross cash position at the end of September stood at around EUR 2.1 billion, equivalent to around 14% of sales, in line with our liquidity target. Our gross debt amounted to EUR 6.8 billion, the resulting net debt to EUR 4.7 billion. Our gross leverage is now at our self-defined threshold of 2x. Net leverage is corresponding to 1.4x . We view this situation as temporary, as increased financial debt following our Ethernet acquisition is meeting a cyclically subdued EBITDA. Our after-tax reported return on capital employed remains at a depressed level. For the complete 2025 fiscal year, our return on capital employed came in at around 5%, clearly falling short of our aspirations.
This is due to an increase of capital employed driven by organic and inorganic investments over the last years, compared with a temporarily lower NOPAT. Finally, to our dividend proposal for the concluded fiscal year 2025. To our next annual shareholders meeting in February, we will propose a dividend of $0.35 per share, unchanged versus the prior year. This would result in a payout of about EUR 460 million. Keeping the dividend constant in a second downturn year with lower net earnings compared to the previous year shows our strong commitment to shareholder remuneration. We are again striking a balance between shareholder remuneration and sufficient financial flexibility for the company. Now back to Jochen, who will comment on our outlook.
Thank you, Sven. As of today, it is challenging to predict a one-year business trajectory in an environment where short-term ordering is limiting visibility. We see the still cautious ordering behavior of customers primarily as a reflection of geopolitical and tariff-related uncertainties. Inventory levels have generally normalized throughout supply chains. That said, there are lingering pockets of digestion and further working capital reductions, and on the other hand, also a few instances of supply tightness. In addition, the underlying demand will determine the magnitude and pace of the recovery. As a base case, we are expecting volume growth to return in 2026 amid a gradual upcycle. For the currently running December quarter, the first of our 2026 fiscal year, we predict revenues of around EUR 3.6 billion, based on a new assumed U.S. dollar/euro exchange rate of $115.
This would equate to a sequential decline of about 9% above our typical seasonality because in the near term there's a risk that some automotive customers are driving down inventories to critical levels below target into calendar year end. Also, from industrial customers, we expect pronounced destocking towards the end of December. By segment, revenue at GIP and CSS is anticipated to decrease more than group average, whereas sales of ATV and PSS should decline less. Compared to the first quarter of our previous fiscal year, we expect 5% nominal growth despite a significantly weaker U.S. dollar. Adjusting for this, like-for-like growth would be 11%. For the December quarter segment result margin, we expect a mid to high teens percentage level resulting from the lower sales volume. Now to our outlook for the full 2026 fiscal year.
In the light of low market visibility, we have to allow for a certain range of outcomes and act swiftly. In our base case, we expect our full year revenues to be moderately up compared to the previous fiscal year. Besides an assumed typical price decline, an unfavorable currency development is masking a stronger underlying volume growth. We assume a U.S. dollar/euro exchange rate of $115 compared to the average of $111 realized in our 2025 fiscal year. This translates into about EUR 400 million of top line headwind using our rule of thumb. In line with the assumed uneven and overall gradual recovery, we expect our business segments to follow markedly different patterns. Regarding automotive semi market, several headwinds make us cautious. While the trade and tariff situation remains fluid, we expect it to have an impact on vehicle affordability and thus weigh on underlying car demand.
The latest production forecast for 2026 S&P sees only a very small decline to about 91 million vehicles. We see this as a best case. Regarding automotive content dynamics, we anticipate the Chinese XEV market to moderate significantly in terms of growth now that it has surpassed 50% penetration, also caused by subsidy reduction and rising efficiency requirements. Europe should see some improving momentum, whereas XEV sales in the U.S. will suffer from the lack of government support. Accordingly, we are seeing push-outs of several new EV platforms across the Western OEM landscape and our production mix shifts in favor of ICE cars. In light of this, we have taken action to restructure our frame-based module production in Warstein in Germany and concentrate it at a more cost competitive Hungarian site in Csiglet. The picture is brighter for the other big auto content driver, the proliferation of software-defined vehicles.
Here we expect market momentum to accelerate from the second half of 2026 onwards, driven by more and more STV launches, further supported by the automotive Ethernet business from Marvell. In total, we expect our automotive segment to grow below group average. In contrast, we expect PSS to grow significantly faster than corporate average, driven by buoyant demand for our AI power solutions, for which we assume revenues now of EUR 1.5 billion. We enable this revenue growth in a very capital-efficient way by converting idle IGBT capacity from ATV and GIP to advanced MOSFET capacity for AI throughout the year. For GIP, we assume revenues to grow moderately year over year based on CapEx restraints by industrial customers, some remaining inventory digestion, moderating growth of renewable energy build-outs and accelerating investments into grid infrastructure. For CSS, we expect only slight yearly growth as IoT demand remains sluggish.
Regarding profitability, we expect our full year adjusted gross margin to come at low 40s and our segment result margin to land at high teens percent level. The adverse currency development, together with assumed typical price decline, will work against the positive fall-through effect from volume growth. Idle costs are expected to go down only gradually to an annual level of around EUR 800 million. They will therefore still constitute a considerable margin drag of around 400 basis points without considering the positive fall-through from additional volumes. Further benefits from our step-up program will support our 2026 margins as more and more measures are becoming effective. Investments are a key variable within our key variable within our control, and the consecutive reductions throughout 2025 show how we take related decisions in an agile manner.
For fiscal 2026, we predict investments, including capitalized development expenses, to amount to around EUR 2.2 billion, a level very similar to the previous quarter year. A focus area will be the finalization of the construction of the fourth module in Dresden and equipment, just in time to match strongly growing customer demand for our AI power solutions. For depreciation and amortization, we anticipate a value of around EUR 2 billion in our 2026 fiscal year. This includes amortization of around EUR 400 million resulting from purchase price allocations, mainly in connection with the acquisition of Cypress and Marvell Ethernet business, which will be recognized in our non-segment results. For the free cash flow, we expect a level of around EUR 1.1 billion for fiscal year 2026. Our adjusted free cash flow net of investment into major front-end buildings is expected to come in around EUR 1.6 billion.
Given the limited visibility, we view our current guidance for fiscal 2026 as the appropriately prudent base case with upside and downside scenarios. The base case already factors in certain headwinds, including persistent tariff impacts, and compared to some market researchers, a more cautious near-term view on XEV adoption. Further downside could stem from escalating geopolitical tensions or unresolved supply chain disruptions. Upside could come from stronger recovery in end markets, less rigid inventory management by customers, even higher AI revenues, and a more benign outcome of tariff bargaining. Let me spend a few words on sustainability metrics. We are very well on track towards year two neutrality by 2030, covering direct and indirect energy-related emissions scope one and two. In all our sites, we are procuring electricity from renewable sources only.
We have exceeded our stated milestone of reducing CO2 emissions by 70% by 2025 compared to the base year 2019. As per end of September, the reduction amounted already to more than 80%. Furthermore, our climate target was validated by the science-based target initiative. This includes our scope one and two emissions, and additionally, it also covers our scope three emissions in close collaboration with our supply chain partners. Before going into Q&A, ladies and gentlemen, let me summarize. We close our 2025 fiscal year fully in line with guidance, both in terms of revenue as well as profitability. The latter was supported by cycle management and step-up benefits. As we move into our 2026 fiscal year, geopolitical events and macroeconomic volatility are tempering cyclical dynamics and cause customers to order on site. Our base case is a gradual, uneven market recovery. AI is standing out as a bright spot.
Our market-leading solutions for the grid-to-core power flow provide us a powerful idiosyncratic growth opportunity. We will more than double our revenue to around EUR 1.5 billion in fiscal 2026, and further significant growth is to come in the subsequent years. Expectations for automotive are bifurcated, muted for short-term XEV demand and currency headwinds, more than compensated by software-defined vehicle trends supported by further growth in analog microcontrollers and the Ethernet business from Marvell. We expect moderate growth for our 2026 fiscal year, including a significantly negative currency impact and assumed typical price declines. Continued implementation of step-up will support a high teen segment result margin. With our unparalleled portfolio of power analog sensors and controlling connectivity solutions, combined with our P2S approach, we are addressing secular growth drivers while focusing on accelerating innovation to customer value.
Thank you very much, ladies and gentlemen. This concludes our introductory remarks, and we are now opening the call for your questions. We kindly ask you to limit yourself to one question and one follow-up. Operator, please start the Q&A session now.
Thank you. Our question- and- answer session will be conducted electronically. If you would like to ask a question, simply press the star key followed by the number one on your telephone. If you are joining us today using a speakerphone, please ensure that your mute function is turned off. And we'll take our first question from Didier Scemama, Bank of America . Please go ahead.
Yes, good morning, gentlemen. Thank you for taking my question. I've got a couple, maybe starting with the big picture, Jochen. I think you very kindly gave us the mix earlier for 2025, which is showing a sort of progressive evolution away from power discrete into analog and microcontrollers and sensors, obviously recently augmented with the acquisition of the Marvell Ethernet automotive switch business. So I just wondered, is it the direction of travel for the company given the headwinds you're seeing in, let's say, silicon power devices and the reshaping of manufacturing you've announced away from high-cost location to lower-cost location, the slower growth in EVs? So is that something you are thinking about? And if that's the case, should we expect over the medium term a lower capital intensity for the business? And I'm going to follow up. Thank you.
Okay. So you know that we want to lead in power systems and IoT i n that regard, we need a broad portfolio which, combined with the P2S approach, will also differentiate us from pure component competitors, for example, also from China. So I believe that the portfolio ranging from power to analog to sensors to compute and connectivity, combined with software, combined with the P2S approach, is the way to differentiate us from low-price competition. Having said this, of course, we need to be flexible in each of these categories and follow closely market developments. And for the time being, as I mentioned also in my summary, we are less hopeful for the XEV market. In fact, we believe that the development will be showing less growth than many market participants anticipate.
And in the IGBT domain, for XEV, as much as for certain applications in photovoltaics, we believe that it's better to use this capacity, as I outlined in my intro, to convert it to leading-edge silicon MOSFETs and use it for the fast-growing margin-accretive AI business. So in principle, I would like to be leading in all three categories, but in detail, we have to then react to market development and focus on profitable growth.
Okay, thank you for that. And then on the AI server, just wanted to look at the puts and takes within that. So first of all, congratulations for getting to one and a half billion, or at least guiding to one and a half billion for 2026. So if we peel the onion a little bit, what's your assumption for market share in 2026? Is that flat, up, down? And then if you could give us a sense of any initial contribution from either GaN or Sikh in that one and a half billion, that would be extremely helpful. Thank you.
Thank you everyone. Andreas will start, and then I'll go ahead.
Yeah, so hello, Didier, Andreas speaking. So in terms of market share, 2026, what we foresee is an increase of our market share given the projects that we are currently working on with major suppliers of server racks and GPU builders and so on and so forth. So definitely here, we are going northwards based upon all the portfolio offering and all the undertaking that we are having with those guys in terms of joint innovation and us providing most efficient and the world's best semiconductor solutions, if you will, to power AI efficiently. Jochen, if you built on this.
Yeah, maybe let me add a qualitative personal comment on this. In my 30 years in the company, and I think Andreas would add just the same, I've never seen such a growth momentum coming on us. It's amazing. Our positioning in the market, ranging across the whole grid to core power flow, customers coming to us for performance, for quality, and also now for delivery capability. So I think the 1.5 billion is a step in this year, but based on the platforms that will ramp at the customer, I think there will be also strong growth into the following year. GaN and silicon carbide are already, of course, included in that number. If you ask me for the first application of GaN, it's likely going to be beyond the parts which are right now already in the PSU.
It's going to be the 48-volt IBC, where we see high potential for GaN adoption.
Okay, brilliant. And maybe just one final quick follow-up. When I do the math on your full year guide and taking into account what you said on ATV, on GIP, and CSS, as well as AI server, it feels like your PSS revenue growth for 26 XAI is a decline of about 8% over the year, which sounds pretty bad. So are you perhaps overly cautious, or is there anything we should be aware of, like loss of sockets in smartphones or anything like that that would justify the decline year over year in the PSS revenue XAI?
No, maybe Sven you take.
Yeah, Didier, hello. So I mean, I'm comparing now, of course, the actual numbers for AI, so no longer the 600, but more than 750 to 1.5, which gives you, let's call it 750 AI growth. If you now look at the rest of the non-AI business and you take into consideration currency and negative PV, there is also growth on the non-AI PSS side. We see good traction on the sensor side. The weaker elements right now are really e-mobility and photovoltaic. And again, we will reuse or repurpose these capacities in a capital-efficient manner to grow in AI.
Yeah, thank you so much.
The next question comes from François Bouvigny from UBS. Please go ahead.
Thank you very much. I have actually a follow-up from Didier's question, more at the group level. So moderate growth, if we assume 3%- 5% growth, so that's 500 million, 800 million incremental revenues. If we take into account the currency, so 400 million, you said, Jochen, so it's a 1 billion constant currency incremental revenues in your guide, roughly 1 billion- 1.2 billion. Now, AI is 800 million alone, and Marvell have 200 billion- 250 million. So all in all, Marvell and AI, it's 1 billion, which is basically what you guide for incremental revenue. So in other words, without currency and without AI, you guide flat at the group level revenues. And I absolutely understand that you need to be cautious in this time, a lot of uncertainty around macro, but flat would be well below what you would expect to be with content growth, what you have laid out for many years. And I would have thought that 2026 would be more normal years, less inventory correction.
So is that fair to say that you take extra conservatism and your guidance implies still significant inventory correction into 2026?
So François, let me take that one. I expected it. So first of all, I start by saying we have the honor and the onus to give annual guidance, as you know, in times of low visibility. We provide, like in the past year, and I think looking back, this worked pretty well, as also many investors have reconfirmed a qualitative outlook. And we said it in the script that it is a prudent base case given where we are in the cycle and in the geopolitical framework. Moderate growth for us means it's a mid-single-digit percentage territory. So you mentioned 3%- 5%, that would be a bit too low in my mind.
So if you then run the numbers again and you take the AI out, you take the FX out, you take Marvell out, there is still some growth, but it is a tad lighter than in the past for the reasons we have mentioned and also for the reasons that this is the start of the year and the guidance contains some kind of conservatism in it.
Right, thank you very much, Sven. 2020And maybe on the profitability side, I mean, it's a kind of a similar question. In fiscal year 25, so you just reported, you had 17.5% segment margin and 41% gross margin, and you guide low 40%s and high 10%s for both for fiscal year 2026, so basically the same. So despite the growth that you see, I would have expected as well lower under-retention charges and also the cost-saving program coming through.
It seems very conservative again to me. So maybe, Sven, could you maybe quantify what you assume in the under-loading charges and cost savings pricing kind of a bridge for those margins to understand?
Yeah, I'm happy to do so, François. And I think it's a question which probably is relevant to many of you, so I will be pretty explicit here. So first of all, there is a kind of an idle assumption included. We are coming from a year where idle cost went up to a billion. We are now forecasting around 800 million. The question comes, why is it still at that elevated level and why is it not going down quicker and further? So here, the two answers are, we are super strongly growing, as you have heard, in AI. So here, nothing is idling. We have to build new capacities to support the growth.
So that does not help on the idle front, number one. And number two, we also see nice growth opportunities on the microcontrollers, on the software-defined vehicle-related parts, which is mostly outsourced. Again, no real help on the idle front. That's why we are not reducing idle to the extent probably some of you had in the models, which is then relevant for the year-over-year comparison in terms of gross and segment result margin. And some more flavor. So if you look at the bridge year-over-year, there are puts and takes. On the positive side, there is a positive volume effect. I mean, our volumes are growing in the low teens. So that is, I think, a good sign. So of course, there's a positive contribution. There is a positive contribution from Step-Up.
I said it in the last call, and happy to repeat, 50% last year, 2/3 this year included in our numbers. So yes, that's positive. On the other hand, there is a single to mid-digit decline in PV included. There is 400 million of FX included. And if you add all that, you end in the high teens. And here again, we do not want to be more precise starting this year from today's perspective. So these are the factors. And maybe last comment, gross margin in the low 40s is, in my mind, pretty resilient in this environment. And that's why you can see here very nicely the 200 basis points contribution from the Step-Up program, which are three-quarters gross margin accretive and only one quarter goes into OpEx. So very long and detailed answer, but I hope it helps all of you to build the right bridges.
Let me add one aspect because maybe it was otherwise confusing because I said we convert IGBT capacity for leading-edge MOSFETs for AI. That is a true statement. But what Sven also said is that we build up new capacity for AI that is mainly on the AMS side. And the conversion of the IGBT capacity, which of course then reduces idle, will take several quarters to become effective.
Very clear answers. Thank you very much, Hanebeck .
The next question comes from Andrew Gardiner from Citi. Please go ahead.
Good morning. Thank you for taking the question. Sven, perhaps first one for you. Just to clarify some of what you described in the last answer to François, you talked about the volume growth of low teens. Can you also sort of give us the detail on what the price pressure is going to be, so ASP pressure into next year as well? It feels like if you're growing at low teens, then given the FX headwind of 400 million, the Marvell boost, that the pricing is perhaps down a bit more than we would have normally anticipated. So can you just specify where that's coming out? And then secondly, perhaps for you, Jochen, you mentioned a couple of times delivery capability when it comes to AI. So is that helping you to gain share in the near term relative to what you might have expected just because of your size, your ability to get those parts out in what is a very rapidly growing market relative to perhaps some of the smaller players that have previously led in the power space? Thank you.
Okay, Andrew, I'll start with your question. So again, I do not want to be more precise now than guiding comparatively, but if you look at the numbers or the guidances which we have given you, I mean, there is this 400 million of FX that's clear on the pricing front, although it's very early stages because we are in the midst of the negotiations with the customers as usually. We are anticipating low to mid-single digit. So that's something you have to deduct. And if you deduct these two and you add then the volume of low teens, then you come into the moderate growth territory, which is, as I said, low sorry, mid-single digits. So that's all I can say at that moment in time. We'll, of course, specify that then over the quarters.
And then in terms of the delivery capability, you're right. In the past, I stressed our differentiating factors in the AI powering AI domain is performance and quality. I would now add delivery capability. I mean, there will be always small players around, but the industry needs players that can really scale. And I think here we are in the prime position and we see customers coming to the table and trying to secure capacities for this ramp to happen. And of course, that is then a great opportunity on our side. The facility in Dresden then comes perfectly in time. But of course, we will take here also an approach that we will not take the whole risk on our books.
Thank you, Jochen.
The next question comes from Stephane Houri from ODDO BHF . Please go ahead.
Yes, good morning, everyone. Actually, I wanted to come back on the AI opportunity because in the press release, you gave a targeted market of a potential market, I would say, about EUR 8 billion-EUR 12 billion by the end of this decade. So the question is, what kind of market share are you targeting at this target? And can you maybe repeat what is your market share at the moment? Thank you.
Yeah, I think we always gave you the indication that in the classical data center, we have today a market share of 30%- 40%. That should be our minimum achievement. Honestly, on this market outlook towards the end of the decade, it's very, very difficult to come up with good numbers. As I said in my intro, it depends on the module share, the customer structure, and of course, the speed of AI build-out, which may depend also on other factors other than, of course, the power part. So very, very difficult to predict, but we wanted to give you some sort of mid-to-long-term view. But I would not be surprised if that number is changing also over time. In general, I would say, if you ask me about the risk entailed, it's up to us to execute. So it's all about us to get the volumes in quality, in time for the various REMs to the customers.
Okay, and another question is on the automotive segment where you seem to be pretty conservative at the start of this new fiscal year. And you notably talked about the kind of price war in China, c an you maybe come back on the potential upsides and downsides to this market and if it is spread, I mean, the conservatism on China, but also on the rest of the market? Thank you.
I can give you some insights. We expect clearly our microcontroller business to continue to grow also into this fiscal year. So that is clearly on the upside. Also the Ethernet business. So everything related, so software-defined vehicles is on the upside. On the XAV side, we take a very cautious note. I mean, look at the BEV sales in the U.S. in October. They tanked as predicted. In Europe, yes, there is some growth, but in China, there is not only a price war taking place, especially on this power inverter that sticks out in the regard that, yeah, I always say each procurement department will find that component and rather beat 10x on that one compared to microcontrollers where the switching efforts are much, much higher. So we take a very cautious note on that. And that, of course, is driven also by the price walls of the OEMs. And therefore, I expect even a revenue decline in power modules in the automotive division. If we are wrong, fine, but that one we are clearly cautious.
And is there a question about competition? That's my last one. Sorry. Competition in China, I meant.
Yeah, again, I mean, there are Chinese competitors, for example, in these IGBT modules as well, t here is less competition or no competition yet to be seen in the microcontroller or the Ethernet or the radar components, which we are selling. So it really depends. But again, our defense line for that is we are not a component supplier in China, particularly in China where customers are eager to adopt solutions very quickly to get their products out in time. And therefore, I think our overall position in automotive China is strong, but we are cautious on the XAV power modules. And the whole thing, of course, doesn't come now unexpectedly. We expected this. And again, we are converting capacities as we are not willing to follow any price war to any unreasonable level.
Just quickly jumping in here at this point, Alex Fulton speaking from Infineon IR. Great to see so much interest. We still have a pretty extensive queue of analysts waiting to ask their question. As a reminder, please restrict yourself to one question in order to be fair to everyone. I think the board here is generous with its time. We can do a little bit of overtime, but please be crisp.
The next question comes from Janardan Menon from Jefferies. Please go ahead.
Hi, good morning. I'll be quick. I've got one question, and it's on your capacity planning for AI. I was actually interested to hear that where you are adding capacity for AI, if I heard you right, is on the ANS side. So what exactly is the ANS component in your AI solution? How much is pure MOSFETs, whether that is silicon, GaN, or SIC, and how much is any other kind of component in that? Any idea?
And when you talk about the conversion from IGBT, which, as you've just been saying, is a low-margin product for you because of price pressure in China, etc., and you're going into the MOSFETs, I know that FY 2027 is quite far away, but if I look at FY 2027, should we be seeing a lot of that effect coming through and get a big step up in margin, both from underutilization and the effect of some of this conversion? Any kind of qualitative answer, that would be great. Thanks.
Yeah, we take every challenge, but of course, we'll focus this now on fiscal 2026. I said that given the platform introductions of the AI guys, which are looming second half of 2026, I think we will see there another growth step, clearly a growth step in our AI business. I would not like to allude now in general on margin on 2027, but the AI business is margin accretive. You were asking about where is AMS playing a role. In particular, it plays a role in the power stages. Each power stage comprises of an analog mix signal part plus two advanced MOSFETs, then combined in more or less complex packages ranging all the way up to vertical backside delivery modules, which, of course, command a much higher price than a simple power stage. But it's really there and some more analog mix signal parts to come if you think about hot swap control and other ASICs. But the biggest part is, of course, MOSFETs. Then I would say the next part is the assembly stack required to package the MOSFETs and the analog mix signal part to the module level.
And would the analog mix signal be about 10% of your total revenue in this area?
This I would need to check, but probably in that range. But here, I would rather like to have a check.
Understood. Thank you.
And, Janardan, just quickly again, I will definitely not guide for 2027 now, but I mean, you are on the right track because if you just think about the cyclical idle costs being a headwind of 400 basis points in this fiscal year without fall-through, then I think this is one of the key answers to a further margin progression in the years to come.
Understood. Thank you.
The next question comes from Joshua Buchalter from TD Cowen . Please go ahead.
Hey, guys. Thanks for squeezing me in, and I'll follow Alex's clear instructions and just ask one. So congrats on raising the AI power number again. I was just hoping you could maybe help us with both the 1.5 billion this year and the longer-term TAM that you gave. How much do you expect that to be coming from stage two power on the GPU board versus other areas in the rack and also out to the grid? Any details you're able to share on that side, again, specifically on the stage two power within the contribution would be very helpful. Thank you.
Yeah, it comes really all across, and it's very difficult to dissect this because on the second stage, the power stages, it all depends on whether it's a lateral power supply still or already a vertical power supply. But you remember maybe my saying that we are on a good track to convince all customers to go for vertical over time. And that, of course, then changes this ratio dramatically.
And maybe another comment on the analog mix signal part, which was just brought up before, it's, of course, one of these parts that also makes it very differentiating because to control these MOSFETs in this environment is a little piece of art where competitors seem to struggle. At this moment in time, I think stage two is a very significant chunk of the money of the revenue. But that can increase for module. And we also see in all the other components power-related now a demand coming towards us because we will be able to deliver.
Thank you.
The next question comes from Sandeep Deshpande from JP Morgan. Please go ahead. Mr. Deshpande, your line is open.
Yeah, hi. Thanks for letting me on. My question is regarding the auto market in terms of your FY 2026 guidance. You are saying that it is below the moderate growth that you're seeing for the firm in the year. How does that break out into your expectation on volume growth and content growth given the points you're making on China on the EV market, etc.? Thanks.
Yeah, not sure whether I can give you now on the top of my head different structures, but again, microcontrollers will grow volume and revenue. Ethernet obviously will grow. Power modules will not grow. MOSFETs in very good demand. Power or the analog parts also, I would say, reasonable. So content-wise, STV is clearly helping us. The power module, as we said, is the one weak element in the automotive story as we speak. Does that answer it, or what would?
In terms of your business in the auto market, how much would you say that the power modules is your biggest market, or today now the mix has shifted much?
No, no. The roughly 7.5 billion, more than 3.5 billion in the automotive division is microcontrollers. So the power module in terms of automotive revenue for Infineon is in the range of 10%.
Understood. Thank you.
The next question comes from Alexander Duval from Goldman Sachs. Please go ahead.
Yes, many thanks. Just one quick question. You talked about various factors that can help you win in AI power, for example, your ability to deliver at scale. But you also touched on GaN as well as the packaging that you can offer. I wondered if you could provide some further color on how important those are in securing a strong market share in the AI domain. Many thanks.
Yeah, Andrea speaking. So going forward, GaN is accelerating in terms of its importance to deliver outstanding and shiny efficiency and power factor, if you will, or size possibilities, which are very much important to power these server racks going forward. So having said that, again, where do we see that this comes and becomes reality? You find it in the power supply units for secondary site rectification. But we also see now first concepts being evolving, then also in the area of the intermediate bus converters. Beyond this, while we are speaking, hard to say, but on the long run, given that Infineon pursues its strategy of offering GaN performance through our 300 mm scaling possibilities, and scaling means then also cost performance. So we want to offer GaN performance at silicon pricing on the long run.
We believe it will strongly proliferate, so to say, server power flow at scale where technologically it makes sense.
The next question comes from Jakob Bluestone from BNP Paribas . Please go ahead.
Great. Thanks. I just got a quick question on the phasing of your revenues through the year. You obviously guided for below normal seasonality for Q1. And I think you mentioned that your orders were up by 2 billion and also that STV would sort of step up in the second half of the year. And so I just wanted to confirm, should we expect a sort of very H2 skewed 2026, or how should we think about the phasing over the course of the year?
I think if you take the last year's number, this quarter is 200 million above last year. And then if you just take every quarter 200 million more, that's it.
Very clear. Thank you.
We now take our last caller for today, Jochenes Schaller from Deutsche Bank. Please go ahead.
Thanks for taking my question. Good morning. Jochen, you made some comments already on the last call, and I think again this morning on the press call about the inventory situation at your auto and industrial customers, and that we're kind of destocking to maybe unsustainable levels. How big is the incremental risk you see from this continuing beyond the December quarter, or are we really going to get so lean that this is then finally over? And has the Nexperia situation, in your view, changed anything in terms of how customers are thinking about holding inventory? Thank you.
Yeah, typically, it's a year-end effect. Obviously, there's a balance sheet reporting then. To be on the fair side, I mean, there are several players, Tier 1s, OEMs, which are really in difficult territories, and they are, of course, cash-strapped. And then the question is, where do you spend your cash? But I can only warn that situations like what we have seen in the last weeks, which hopefully now are resolved, may occur again if significant players are reducing their inventories to unsustainable levels. For us, I would expect the main impact to happen this quarter like in the past as of this year-end effect.
So is it fair to assume that in your full-year guidance, you have not begged in a kind of restocking from these customers, but more kind of?
No, exactly. If all these customers raise their stock to reasonable target levels, we can debate reasonable, but I think that would be beyond the base case.
Very clear. Thank you.
Jochenes, maybe quickly to add, I mean, you heard me saying that on some conferences and some roadshows already. I fully understand, and I'm part of the community who is really watching to manage working capital in an efficient way. But overdoing it is a risky thing, and we are just seeing it again with the case you were just referring to. But we can only encourage our customers to look at that.
Super clear. Thank you very much.
Excellent. Thanks, everyone. With a quarter of an hour of overtime, we are now wrapping up the call. Thanks for the questions. Thanks for the interest. This concludes the fiscal year-end outlook 2026 conference call. Of course, you can reach us in the IR team here in Munich every time. My colleague on the press side said it feels a bit like pre-Christmas time. find that a tad early, but still. Have a good time. Take care and talk soon. Bye-bye.