高通 (QCOM) 2026會計年度Q4財報電話會議：資料中心放量與車用成長

重點摘要

• 高通公布 2026 會計年度第四季 Non-GAAP 營收為 104 億美元，Non-GAAP 每股盈餘（EPS）為 2.17 美元。QCT 營收達 90 億美元，季增 6%，達到財測高標。
• QCT 車用業務單季營收首次達到 20 億美元，年增 89%。高通在 2026 會計年度結束時，車用年化營收運作率達到約 80 億美元。
• 管理層重申 2027 會計年度資料中心營收目標為 50 億美元。預計第一季營收將超過 6.5 億美元，並於 2027 會計年度結束時達到約 70 億美元的年化營收運作率。
• 高通預期在 2027 會計年度，涵蓋車用、物聯網與資料中心的 QCT 非手機業務營收將年增逾 60%，並占 QCT 營收超過 50%。
• 公司已與蘋果、華為及傳音完成長期授權協議，未改變 QTL 的營收運作率。預計下一個主要換約週期將於 2030 會計年度展開。
• 2027 會計年度第一季財測預估營收為 104 億至 112 億美元，Non-GAAP 每股盈餘為 2.15 至 2.35 美元。

關鍵財務數據

業務與營運表現

資料中心

高通的資料中心業務正步入初步營收放量階段。在向兩大超大規模雲端客戶出貨客製化晶片的推動下，預計第一季營收將超過 6.5 億美元。管理層表示，2027 會計年度 50 億美元目標的大部分將來自與這些客戶（包括亞馬遜）的客製化 AI 加速器合作專案，而連接產品則是第二大類別。

公司預期資料中心營收將在 2027 會計年度持續擴大規模，並在年度結束時達到約 70 億美元的年化營收運作率。管理層指出，相較於需求，供應端更有可能是限制營收超越 50 億美元目標潛力的主要因素。

高通維持 2029 會計年度資料中心營收達到 150 億美元的目標。該目標預計涵蓋客製化晶片、擴展連接產品、CPU 以及市售 AI 加速器。公司亦指出，其與 AWS 的策略協議代表在 10 年內高達 600 億美元的潛在營收機會。

高通表示，其 AI250 高頻寬運算晶片的表現優於原先的能效目標，在實驗室測試中達到約 0.3 皮焦耳/位元（picojoules per bit），而原設計目標約為 0.5。管理層指出，AI 解碼工作負載是客戶展現高度興趣的領域。

車用與物聯網

在客戶採用高通第五代運算平台整合模組的帶動下，第四季車用營收達到 20 億美元。第五代 Snapdragon 數位底盤已於 9 月當季開始商業出貨，預計將於 2027 會計年度持續放量。

管理層將車用業務的成長主要歸因於客戶開案設計轉化為營收，以及半導體含量提升。高通表示，第五代平台的晶片價值含量是其第三代平台的八倍。

物聯網營收年增 4% 至 19 億美元。工業、網路及機器人產品成長 20%，但部分被受記憶體限制影響的消費性產品所抵銷。Dragonwing IQ10 機器人處理器已排定展開商業放量。

手機與技術授權

QCT 手機業務營收為 51 億美元。管理層表示，6 月當季為中國 Android 營收的低谷，隨後在 9 月及 12 月當季實現季增長。

高通預期 2027 日曆年全球手機出貨量將持平或下滑。公司表示其營運計劃已納入蘋果數據機晶片轉移的影響。至於三星，管理層表示未來各代 Galaxy 機型的市占率規畫基準將由過去的 50% 提升至 70% 以上。

在技術授權方面，高通完成了與蘋果、華為及傳音的長期協議。管理層表示，換約維持了 QTL 的權利金費率與營收運作率，同時消除了近期主要手機代工大廠（OEM）的續約不確定性。

連接與基礎設施

SpaceX 選擇高通支援 Starlink Mobile 在全美建置的地面 5G 網路。高通將提供射頻晶片、大規模 MIMO 技術、數位單元加速及 Layer 1 軟體。預計將於 2027 年開始部署，並在 2028 會計年度帶來更顯著的營收貢獻。

高通亦在擴展其光通訊連接產品組合，包括 400 Gb、800 Gb 及 1.6 Tb 產品。預計對 AWS 的連接產品出貨將放量，且公司已爭取到第二家光通訊連接客戶。

管理層財測

手機營收預期的下滑，反映了 Android 的持續成長被蘋果產品營收下滑所抵銷。管理層表示，將資料中心納入 QCT，以及在 AI 加速器和 CPU 營收放量前進行的投資，使第一季 QCT EBT 利潤率降低約 3 個百分點。公司仍維持 30% 的長期 QCT EBT 利潤率目標。

高通正在實施漲價以抵銷更高的半導體投入成本。管理層預期效益將於 2027 會計年度第二季開始顯現，並在下半年度完全反映於 QCT 毛利率中。若不計資料中心產品組合，公司預期其基準毛利率將回升至歷史 48% 至 50% 區間的低標水準。

風險與關注領域

• 隨著營收規模擴大，客製化資料中心晶片預計將支撐 QCT 營業利益率，但會稀釋毛利率。
• 2027 會計年度 QCT 手機業務營收將持續反映蘋果產品營收逐步萎縮的影響。
• 管理層預期全球手機市場將維持持平或下滑，同時記憶體價格與供應限制將持續影響手機與消費性物聯網需求。
• 高通預期供應端將限制資料中心營收超越 2027 會計年度目標的潛在上行空間。
• 資料中心在 AI 加速器與 CPU 平台的投資正在營收放量前展開，為近期 QCT 利潤率帶來壓力。
• 漲價旨在抵銷上升的半導體成本，但預計要到 2027 會計年度下半年才能完全體現利潤率效益。

分析師問答焦點

分析師高度關注資料中心放量的步調與獲利能力。管理層表示，兩大超大規模雲端客戶將貢獻 2027 會計年度 50 億美元目標的大部分，並應在 2028 及 2029 會計年度持續作為重要貢獻來源。後續加速器、CPU 與客製化晶片的前景可望進一步支撐長期成長。

在利潤率方面，高通重申客製化晶片應有助於提升營業利益率，但會稀釋毛利率。管理層預期自第二季起，產品調漲將部分抵銷半導體成本壓力。

關於手機需求的問題主要聚焦於近期 Android 的成長是否反映了漲價前的提前拉貨。高通表示，通路庫存仍處於低檔，且漲價主要透過較新的平台與裝置導入，從而降低了潛在的提前拉貨效應。

管理層將車用業務的加速成長歸因於半導體含量提升與開案設計轉化為營收，而非提前拉貨。管理層亦堅稱，儘管近期有資料中心投資以及部門財報呈現方式的變更，長期 QCT EBT 利潤率目標仍維持在 30%。

完整財報電話會議逐字稿

完整財報電話會議逐字稿

管理層陳述

Operator

Ladies and gentlemen, thank you for standing by. Welcome to the Qualcomm Fourth Quarter Fiscal 2026 Earnings Conference Call. [Operator Instructions] As a reminder, this conference is being recorded November 4, 2026. The playback number for today's call is (877) 660-6853. International callers, please dial (201) 612-7415. The playback reservation number is 13762625.

I would now like to turn the call over to Brett Simpson, Senior Vice President of Investor Relations. Mr. Simpson, please go ahead.

Brett Simpson

Thank you, and good afternoon, everyone. Today's call will include prepared remarks by Cristiano Amon and Akash Palkhiwala. In addition, Alex Rogers will join the question-and-answer session. You can access our earnings release and a slide presentation that accompany this call on our Investor Relations website. In addition, this call is being webcast on qualcomm.com, and a replay will be available on our website later today.

During the call today, we will use non-GAAP financial measures as defined in Regulation G, and you can find the related reconciliations to GAAP on our website. We will also make forward-looking statements, including projections and estimates of future events, business or industry trends or business or financial results. Actual events or results could differ materially from those projected in our forward-looking statements. Please refer to our SEC filings, including our most recent 10-K, which contain important factors that could cause actual results to differ materially from the forward-looking statements.

And now to comments from Qualcomm's President and Chief Executive Officer, Cristiano Amon.

Cristiano Amon

Thank you, Brett, and good afternoon, everyone. In fiscal Q4, we delivered non-GAAP revenues of $10.4 billion and non-GAAP earnings per share of $2.17. QCT revenues of $9 billion, up 6% sequentially, were at the high end of our guidance. Licensing business revenues were $1.3 billion. For fiscal '26, we reported non-GAAP revenues of $43.2 billion and QCT revenues of $37.2 billion, highlighted by year-over-year revenue growth of 24% across combined automotive and IoT.

We closed fiscal '26 with many of the uncertainties that have been weighing on our business now behind us, specifically QTL license renewals, the loss of Apple as a customer and the negative impact of memory pricing on the size of the mobile industry. Regarding the latter, the June quarter marked the low point for our China Android business, and we have seen sequential growth this quarter. Going forward, our plan assumes global handset units will remain flat to down in calendar '27. The timing and shape of the Apple modem transition is fully incorporated in how we're planning the business and is well understood at this point. We also renewed important QTL licensing agreements, including Apple and Huawei.

Now we enter fiscal '27 as a broader, more diversified Qualcomm with strong momentum and significant opportunities ahead. First, non-handset revenues in QCT will grow by more than 60% year-over-year. Second, our data center business begins its ramp with revenues expected to exceed $650 million in the December quarter. Third, automotive revenues reached an $8 billion annualized run rate exiting fiscal '26, $1 billion ahead of expectations. Fourth, we established a solid foundation in how agentic AI transforms the edge and leads to new product categories for Qualcomm. And finally, we reached a landmark agreement with SpaceX to power their nationwide terrestrial 5G network in the U.S.

Our technology investments are aligned with the direction the industry is moving. We've made key acquisitions to build on our hardware and software capabilities as we evolve into a full stack semiconductor company. We announced strategic agreements with industry leaders, including Meta, Amazon, BMW and Stellantis with more to come, and our design win pipeline across the company grew 70% year-over-year. This is the backdrop that positions Qualcomm to deliver on our fiscal '29 targets of $40 billion in non-handset revenues with over $15 billion coming from data center.

Now I will focus on the areas that are top of mind for investors. Let me start with licensing, where we reached a milestone this quarter. We're extremely pleased with our renewals with Apple, Huawei and Transsion. These agreements provide stability for QTL as we move to the 6G era. They also demonstrate the enduring value of our IP and patent portfolio. And just as in every prior generation, we expect our inventions to be foundational to 6G. Importantly, our QTL revenue run rate remains unchanged. The next renewal cycle begins in fiscal 2030 and runs through fiscal 2033. Licensing gives Qualcomm a durable, highly profitable foundation as we scale our next set of growth opportunities.

Data center is another area where we made considerable progress this quarter. It's one of our top strategic priorities and a major focus of my own time as we build this business around long-term customer relationships, a broad product portfolio and an open software ecosystem. AWS is a good example. Our agreement spans multiple product generations across custom silicon for AI inference and optical connectivity. Custom accelerators began shipping in October and our connectivity solutions with AWS ramp later this month. We also secured a second optical connectivity customer. With Qualcomm High Bandwidth Compute, our AI250 silicon is exceeding our expectations for power efficiency and bandwidth and early yields are on track. These results are generating significant interest from AI customers and memory partners, and we added 2 major prospects in the quarter.

Our Dragonfly C1000 CPU remains on track to ship to Meta as scheduled, and we engaged with a large new potential customer. We're optimistic about the opportunities we have today with our high-speed connectivity solutions such as optical DSPs, and we are investing in bringing optics directly alongside our XPUs and custom silicon as we demonstrated with Lumentum and Corning at ECOC 2026. In software, our full stack integration with Modular is progressing well. Modular is delivering impressive performance as an open multi-vendor stack running on NVIDIA and AMD GPUs, Amazon Trainium and Google TPU. Optimization for Qualcomm accelerators is also underway. Last month, Modular made the full MAX source code available, and this builds on open sourcing the Mojo language in August. We're also excited that Modular Cloud went live during the September quarter and is already processing tens of trillions of tokens per month. We'll share more details soon about how our ecosystem partners are using Modular Cloud.

Our data center business is off to a strong start. We are reiterating our $5 billion revenue target for fiscal '27 with healthy order coverage supporting continued revenue growth in fiscal '28. In automotive, we exit fiscal '26 at an annualized revenue run rate of approximately $8 billion, well above the $7 billion expectations we provided last quarter. Additionally, fiscal fourth quarter automotive revenue grew nearly 90% year-over-year. Commercial shipments of our fifth-generation Snapdragon Digital Chassis began in the September quarter and will continue to ramp in fiscal 2027. As a reminder, this platform is our most advanced to date, delivering a significant increase in performance and represents our largest generation-to-generation content increase in Qualcomm's history.

Our products also evolve from SoCs to integrated compute modules. Customer demand remains strong across our portfolio, and Qualcomm is positioned to become one of the largest automotive semiconductor suppliers by revenue in fiscal 2027. In robotics, our premium tier Dragonwing IQ10 processor is set to begin its commercial ramp and continues to generate strong customer momentum. During the quarter, ecosystem partners demonstrated AI models of up to 80 billion parameters running directly on IQ10. In fiscal 2027, we will continue to expand the technologies and capabilities across our robotics platforms. As we anticipated, the proliferation and capability of AI agents is increasing rapidly, and we're seeing growing consumer adoption as a result. A prime example is Meta Muse, which reached over 10 million downloads in just 40 days.

We're also seeing more companies designing experiences and devices around agents. This includes agentic smartphones, smart glasses and other new personal AI form factors. Meta's audio AI glasses, Muse Charm, Gemini eyewear and Alibaba's Qwen AI glasses are recent examples of intelligent wearables with more to come over the next year. The smartphone is an essential device for agentic experiences as it's the device that knows people best. It's always with them and has their context, preferences and permissions. The device is also evolving to simultaneously support what the user is doing and what agents are doing on the user's behalf. Delivering these experiences will require greater computing performance, power efficiency and connectivity, which play directly to Qualcomm's strengths. More of our customers are already building agentic features into their flagship devices, and we're pleased to share that the world's first agentic smartphones were launched on Snapdragon.

Heading into fiscal 2027, we're pleased with our premium tier Android share position, and we expect it to strengthen as the transition to agentic smartphones drive an increase in computing content per device. For perspective, leading AI companies are targeting smartphones by 2028, capable of running 100 billion parameter models locally with an all-day battery life. Reaching this goal will require a step function change in on-device compute performance. Finally, I'm very proud to share that SpaceX selected Qualcomm to power Starlink Mobile's nationwide terrestrial 5G network in the U.S.

Starlink is building one of the most ambitious new cellular networks in the country, and our Dragonwing infrastructure platforms will bring together advanced massive MIMO purpose-built RAN processing and commercial-grade Layer 1 software. Under a multiyear agreement, deployment begins in 2027 and is expected to be one of the largest 5G deployments in the United States. Together, we have an opportunity to advance United States wireless leadership, starting with 5G and the convergence of satellite and terrestrial connectivity. This new architecture establishes a foundation for next-generation connectivity and network intelligence. This is a significant win for our infrastructure business and the beginning of a strategic relationship with tremendous potential.

Before I close, I would like to briefly address our litigation with Arm. We're gratified that the jury found Arm breached its long-standing Architectural License Agreement, ALA, with Qualcomm, failed to honor its commitments, intentionally acted to disrupt Qualcomm's business and engage in unfair competition, both as a supplier and a competitor. Together with Qualcomm's prior court victory, this outcome further exposes Arm's intentions to harm Qualcomm and its ecosystem. This outcome is a victory for innovation and competition, and we remain focused on developing and delivering industry-leading semiconductor products, including our Oryon CPUs to customers and consumers worldwide from mobile to data center.

In summary, we entered fiscal 2027 having made important progress against the ambitious growth plan we outlined at Investor Day. Since then, we've signed major strategic agreements and positioned our data center business to ramp at scale. We remain focused on execution as we build the next chapter for Qualcomm, and we're confident in delivering on our fiscal 2029 targets.

With that, I will turn it over to Akash.

Akash Palkhiwala

Thank you, Cristiano, and good afternoon, everyone. Let me begin with our results for the fourth fiscal quarter. We delivered non-GAAP revenues of $10.4 billion and EPS of $2.17 with revenues at the high end of our guidance. QTL revenues of $1.3 billion and EBT margin of 70% were in line with our expectations. QCT revenues of $9 billion were at the high end of our guidance and EBT margin of 24% was in line with expectations. QCT handset revenues of $5.1 billion reflected the impact of industry-wide memory dynamics on the global smartphone market.

In QCT automotive, we delivered our first $2 billion revenue quarter with 89% year-over-year growth, driven by the growing adoption of integrated module products for our fifth-generation compute platforms. QCT IoT revenues of $1.9 billion were up 4% year-over-year, led by our industrial, networking and robotics category of products, which grew 20% versus the year ago period, offset by the impact of memory constraints on consumer products. Total non-handset revenues in QCT, including automotive and IoT, grew 36% year-over-year, underscoring the continued execution of our growth initiatives. Lastly, we returned $2 billion to stockholders, including $1 billion each in share repurchases and dividends.

Before turning to guidance, I'd like to provide an update on a couple of factors reflected in our financial performance. First, in QTL, we have now completed long-term license agreements with Apple, Huawei and Transsion and do not have any upcoming license renewals with major handset OEMs for the next several years. Following these agreements, QTL revenue run rate and long-term revenue forecast remains unchanged. Second, as we outlined last quarter, we are implementing price increases across our products to reflect the higher semiconductor input costs. We expect the full benefit of these actions to be reflected in QCT gross margins in the second half of fiscal '27. Finally, we're excited that fiscal '27 marks an important milestone for our data center business as we transition from customer engagements to revenue generation.

Our first fiscal quarter guidance reflects the commencement of shipments to 2 hyperscaler customers for custom silicon engagement. As we have discussed previously, we expect these custom silicon engagements to be accretive to QCT operating margins but dilutive to gross margins. These factors are reflected in our fourth quarter performance and first quarter outlook. I'll now provide our guidance for the first fiscal quarter. We are forecasting revenues of $10.4 billion to $11.2 billion and non-GAAP EPS of $2.15 to $2.35. In QTL, we estimate revenues of $1.3 billion to $1.5 billion and EBT margins of 72% to 76%, reflecting normal seasonal trends. In QCT, we expect revenues of $9 billion to $9.6 billion and EBT margins of 20% to 22%. Please note that QCT EBT margin guidance includes an impact of 3 percentage points due to the addition of Data Center in the QCT segment starting in fiscal '27 and reflects investments for AI accelerator and CPU platform ahead of revenue ramp.

Importantly, this segmentation change has no impact on total company guidance or on our long-term QCT EBT margin target of 30%. We forecast QCT handset revenues to decline by a mid-single-digit percentage sequentially as growth in Android revenues is offset by the continued ramp down in Apple product revenue. In QCT Data Center, we expect revenues of greater than $650 million, driven by commencement of shipments to custom silicon customers. We expect combined QCT Automotive and IoT revenues to be approximately flat sequentially. Lastly, we anticipate non-GAAP operating expenses to be approximately $2.8 billion in the quarter.

Before I conclude my prepared remarks, I'd like to outline the progress in the execution of our growth strategy. In fiscal '27, we expect to achieve 2 key milestones for revenue diversification. First, non-handset revenues in QCT across automotive, IoT and data center will grow by greater than 60% on a year-over-year basis starting in the first fiscal quarter. Second, non-handset revenues will represent more than 50% of QCT revenues in the fiscal year. In data center, we're executing to the revenue targets we outlined at Investor Day, including $5 billion in fiscal '27 and $15 billion in fiscal '29. We expect Data Center revenues to scale throughout fiscal '27, exiting the year at a $7 billion run rate. We're also expanding our customer engagement pipeline across additional global hyperscalers, including the long-term strategic agreement with Amazon Web Services with a potential revenue opportunity of up to $60 billion over the next 10 years.

During the quarter, we also successfully validated our high-bandwidth compute technology in lab trials, marking a key milestone in delivering an innovative AI inference accelerator solution designed to address the memory bandwidth challenge in data centers. We are also well positioned to lead other key industry trends, including the transition to agentic AI in consumer devices and the adoption of physical AI across automotive, industrial and robotics. Finally, we would like to thank our employees for their exceptional execution and innovation throughout fiscal '26. Their commitment enables us to deliver industry-leading products to customers around the world.

This concludes our prepared remarks. Back to you, Brett.

Brett Simpson

Thank you, Akash. Operator, we are now ready for questions.

Operator

[Operator Instructions] The first question comes from the line of Joshua Buchalter with TD Cowen.

分析師問答

Joshua Buchalter

Maybe to start, I mean, last quarter, you guys kind of telegraphed the gross margin step down that was expected from the ASIC business starting to ramp. How should we think about it through the rest of the year, though? Because just the simple math of having to get to your $7 billion run rate exiting the year implies the data center business should grow close to 3x. Like should we expect gross margins to continue to trend down through the year as a result of the larger data center business and this being the year where custom silicon is ramping aggressively?

Akash Palkhiwala

Josh, it's Akash. So let me kind of frame the overall gross margin framework for the company. As you know, we mentioned this last quarter, we are implementing price increases to reflect the higher semiconductor cost across the industry. And so we start realizing the benefit of that in our second fiscal quarter, and we'll see the full benefit of that show up in second half of '27. So the way it plays out is as these price increases come into effect, our baseline gross margin goes back to our historical range of 48% to 50% at the low end of that range.

And then separately, on your question on data center, we -- as we said previously, we expect that to be something that is accretive to our operating margins, but dilutive to gross margins. And so that calibrates the gross margin down from the framework I just outlined. So that's no different than what we told you last quarter, and we're executing on the plan.

Joshua Buchalter

And then maybe for my follow-up, it's good to see the HBC demos that you outlined earlier this week. Can you maybe speak to what type of workloads you're seeing the most traction for your accelerator parts initially? And then any updates on the commercial traction as we think about that? I think I believe that's supposed to start contributing and ramping in the fiscal second half of this year, which is kind of only a couple of quarters away at this point.

Cristiano Amon

Thanks, Josh. This is Cristiano. Look, we're very happy with the progress we're making in HBC. We're actually doing better than what we said at Investor Day. We had a target on the design to about 0.5 picojoules per bit. We're doing better in the lab. We're doing about 0.3. Yield is exceeding our expectations. So we're very happy with the technology progress. The main workloads that we're getting a lot of traction, it's incredibly efficient for decode. So decode, it's where we're getting a lot of the positive feedback.

We are engaged with customers. It changes the conversation when you have a PowerPoint to where you have silicon. And I think that continues to build confidence in the projections we have, not only for our custom ASIC, but for our merchant solutions coming up as we think about 2029 targets. What is also interesting, it now became part of the conversation of custom silicon with HBC, combining our HBC with some of our custom accelerators. So I would just say customer engagement is very positive.

Operator

The next question is from the line of Joseph Cardoso with JPMorgan.

Joseph Cardoso

Maybe just wanted to start off first with happy for you guys to give us the disclosure around how to think about the Data Center revenue ramp through the year. Maybe can you just provide a little bit more color on how you're thinking about your customers and how they get paced into that revenue ramp through the year itself? And then maybe the second aspect there, too, is, I think in the past, you've talked about 2 customers. So can you provide any color on the third customer here that's mentioned in the slide deck and how you're thinking about that getting intertwined in the guidance that you provided for this year? Or does the third customer ramp at another point? And then I do have a follow-up.

Akash Palkhiwala

Okay. Joseph, this is Akash. Let me address the financials and then Cristiano will comment on top -- on customer ramp. So for fiscal '27, we are looking at $5 billion revenue forecast that we've previously provided. We're guiding greater than $650 million in the first quarter, and we expect this revenue to ramp throughout the year as we go through the quarters and we'll exit the year at approximately a $7 billion exit rate going into '28. And so very happy with that. We will -- we're very well positioned from a supply perspective to execute on the $5 billion target. And really, we expect to be constrained by supply rather than demand for upsides beyond the $5 billion.

When you look at fiscal '28, we are looking at a $7 billion starting point. And then the customers that we have in '27 will extend to '28 as well, and it gives us really strong design win coverage to deliver very strong year-over-year growth in fiscal '28 as well. And then as you know, we have set a target of $15 billion for '29, which the customer engagements that Cristiano will outline right after this will position us very well to have a very high coverage of that number as well.

Cristiano Amon

Just real quick, Joseph, on the customer. The majority of the $5 billion in fiscal '27 comes from the 2 hyperscaler customers, so one of them being Amazon. I think the details of those were made public on the silicon plus connectivity. We do have another customer on the accelerator and an advanced discussion also on subsequent custom silicon that is part of the ramp towards '28 and '29. And we had said in the script, I think we have new prospects, especially on the CPU as well, I think, beyond Meta, and we're working with the customers.

Joseph Cardoso

Got it. I appreciate the color there. And then maybe just wanted to double-click on how you guys are framing handset revenue kind of exiting the year and into next year. Specifically, are you guys thinking about the growth here in terms of pricing layering on? I know it sounds like you're not expecting that to be material or to kind of tranche here over the subsequent quarters. But any more color you can help us think about how that pricing ends up influencing kind of the numbers here and the magnitude that we should be thinking about?

And then as a second question, just bigger picture. Obviously, when we look at the market data points on the handset market, it does seem like a contraction on kind of the memory prices has been something we've been witnessing through this year. Just any early thoughts on how you're thinking about the health of the market going into next year? I would appreciate it.

Akash Palkhiwala

Yes. Joseph, when you look at the overall handset market, '25 to '26, as -- like other industry analysts, we are projecting a low teens to mid-teens decline in the market. And then as we go into '27, we're thinking flat to slightly down for the handset market. And that's how we've always projected our business based on that assumption, and it's playing out consistent with expectations. Specifically for QCT handset revenue in Android, as you'll recall, we called the June quarter the bottom for QCT handset revenues and Android in China. What has happened since is we've seen revenue grow into September and December quarter. And so we are encouraged to see that. We've been conservative in our planning, and it's playing out exactly as we'd expected.

Operator

The next question is from the line of Joe Moore with Morgan Stanley.

Joseph Moore

In terms of the royalty rate, you sort of said that the revenue run rate of QTL is unchanged. Is that the rate per device kind of similar as you look at these new contracts that you signed? Is there any change there? I assume that's right, but I just want to make sure.

Akash Palkhiwala

Yes, Joe, you should -- we obviously don't specifically talk about the details of the contract. They're confidential, but we are indicating that the revenue run rate remains the same, which you should take it as we were able to hold the rates and the royalties as we renewed the contracts.

Joseph Moore

Great. That's helpful. And then separately, on the Amazon deal, the warrant structure that you have there, do you view that as kind of the model for approaching other customers? Do you expect to have those types of warrant arrangements? Or is that unique to that specific relationship?

Akash Palkhiwala

Well, if you look at our general set of engagements across data center, the engagements are very traditional, normal kind of customer engagements in the semiconductor industry. The Amazon warrant structure is the exception. And as you know, the warrants that we issued were at-the-money warrants. So that's the only exception that we have at this point. And I think as we consider data center engagements going forward, we'll look at other structures on a deal-by-deal basis, but you should think of our framework as being aligned with the traditional semiconductor model.

Operator

The next question is from the line of Timothy Arcuri with UBS.

Timothy Arcuri

Akash, I was wondering if you can provide a little bit of maybe outlook into March. I know that you don't want to guide March. But I think last call, you thought that revenue would be up a little bit in March. You have easy comps for Apple into March. Android should be a little better. So can you just give us kind of just a sense on March would be great.

Akash Palkhiwala

Yes. Tim, you're right. We are not guiding March at this point. But let me make a few comments on the year that I hope are helpful. The first is, typically in previous years, December has been our strongest quarter in the year. Now with Apple product revenue ramp down, we do not expect December to be the high quarter for us. When you look at Data Center revenue, we've talked about kind of greater than $650 million in the first quarter, but we expect that to ramp as we go through the year. So that kind of defines our quarterly profile to some extent.

The third is when you look at '26 to '27, we expect auto IoT plus Data Center kind of non-handset revenue to grow by greater than 60%. So this revenue would more than offset the decline in Apple product revenue. And then finally, I think if you look at to fiscal '27, we expect handset now to be less than 50% of our total revenue. And so obviously, that's a key milestone for our diversification plan, and so we are happy to achieve that this year.

Timothy Arcuri

Got it. So you don't want to -- I just asked because last call, you did sort of say that March would be up. I guess you don't want to commit to that now.

Akash Palkhiwala

Yes. I think it's the reasonable framework that I think the way people have been looking at the business is a reasonable way of looking at it.

Operator

Our next question is from the line of Stacy Rasgon with Bernstein Research.

Stacy Rasgon

I wanted to dig into your handset market outlook a little bit. I'm a little surprised still with the outlook for flat, especially given where we seem to be running at the end of the year, we're probably down more than 20% year-over-year in Q3 and things don't look better, don't look all that great in Q4. You also said your own China Android revenues were actually starting to grow in, I guess, June or September and December. You're also doing price increases. Like how do you know in the context of everything that's going on that, that's just not prebuys in front of the price increases in Android? I'm just struggling a little bit with the handset outlook. Any further color you could give on that would be super helpful for me.

Akash Palkhiwala

Sure, Stacy. So let me break it down into 2 parts. Let's talk about the market and then separately talk about QCT product revenue within Android. Overall market, as I said earlier, our view is that '25 to '26, it's down low to mid-teens. And then '26 to '27, we are projecting flat to down for the total market. So we think that's a prudent assumption as we plan given the uncertainty overall in the environment.

Specifically on QCT revenue, as you'll recall, the impact we saw in '26 was not just from the change in the market, but also channel inventory drawdown by OEMs. And we don't expect that impact in '27 because we've been following channel inventory. As you know, we get QTL information as well that we can check against. And so the channel inventory continues to be thin. And so we expect kind of QCT revenue really to be aligned with the overall market, and that's how we are projecting it going forward.

Cristiano Amon

Maybe, Stacy -- it's Cristiano, just to add another thing. One of the things we had said before. And that's kind of because of our long-term partnerships that we have with our customers. A lot of the price increases, the reason that they have a delta versus the cost increase, and we talk about prices become more effective in the numbers in the second half, we kind of been pricing newer platforms, new phones. So less subject to prebuying effect. And as Akash said, I think based on the data we're seeing in the market, June remains, I think, the bottom, I think, for China Android.

Stacy Rasgon

If I could ask one follow-up, please. You talked about Samsung market share framework, 75%. Is that still the framework share you have in mind for their -- your share with them in their next-generation platform that ramps in '27?

Cristiano Amon

Yes, absolutely. So here's how we think about the Samsung business. And here's a simple way to do it. We have -- we -- if you think about our business in the past, the baseline was 50%. This baseline is no longer 50%. It's north of 70% as we have been doing. When you think about future generations of Galaxy, this is how we plan our business, and I encourage all of you to plan the same way. That's the new baseline with Samsung. The other thing with Samsung is we're expanding the business. So traditionally, we were not the main supplier of all of their wearables and other devices, and you started to see that associated with the premium Galaxy design as well.

Operator

Our next question is from the line of Ben Reitzes with Melius Research.

Benjamin Reitzes

I'm switching my question because Stacy just asked on the Samsung. So I wanted to ask about automotive with regard to the beat in the quarter versus expectations. I know you're talking about IoT, automotive and data center all growing 60% for the year. But specifically in automotive, it was particularly strong versus expectations, and you are increasing some prices that you mentioned. Is there a chance that there was prebuying there in the automotive segment? Or do you feel like this kind of performance is sustainable throughout the year? And then I have a follow-up.

Cristiano Amon

No, very good. Thank you for the question. Look, we have been saying -- I believe we have said that in last quarter about the transition of our fifth generation platform. Just to give an idea, for cars that you see now on the road that with our third-generation platform, the silicon content increase versus the fifth generation is 8x. There's a lot of more -- there's a lot more compute content into the automotive. And most of our automotive growth is really share where the design win pipeline translate into revenue. That's where we started to see step function increases on the SoC revenue, and that will continue.

Benjamin Reitzes

Okay. And then, Akash, with regard to the EBIT margins, the EBT margins for QCT, how is that going to progress throughout the year from the 20% to 22% that you guided for given the mix shift towards data center? I was a little unclear there. How should we model that as we go throughout the year?

Akash Palkhiwala

Yes, Ben, so maybe it's a good opportunity to clarify that one of the changes we have made in the way we segment the business is QCT now includes Data Center. In '26, it was outside the QCT segment. And just that change kind of reduces QCT gross margin -- operating margin by 3% as, of course, we are investing ahead of revenue in a couple of product lines. It doesn't change the overall Qualcomm numbers. It's just a change within segments that is included now in the QCT numbers. We're, of course, not guiding specific quarters or years for operating margins, but you should think of it as our operating margin target for QCT doesn't change. We are still looking at a target of 30%. And as revenue starts scaling across the next 2 to 3 years, we are confident in our ability to achieve it.

Operator

The next question is from the line of Vivek Arya with Bank of America.

Liam Pharr

This is Liam Pharr on for Vivek. On networking interconnects, you mentioned the AWS optical ramp and a second optical connectivity customer. Could you help us size how large connectivity is today and how you see it growing into fiscal '27 and '28 as you move to 224 gig and 1.6T? And does the ongoing consolidation in the optical interconnect space impact you in any way? Do you see any natural synergies between interconnects and lasers from an ownership standpoint?

Cristiano Amon

Very good. Actually, thank you so much for the question. I really appreciate that. So we're not breaking that down, but you should assume that there is interconnect connectivity revenue within the $15 billion that we talk about it and some of the $5 billion that we talked about in data center, $5 billion '27, $15 billion '29 . We have been very optimistic about this. We have been investing. This asset came to Qualcomm from the Alphawave acquisition. I think Qualcomm is probably a natural place for comms and we have been investing into this.

We have very interesting technology demonstration or progress with optical and co-packaged optics in the quarter in the 2026 ECOC in Europe, we actually did the demonstration with Lumentum. And you bring a good point. I think the structure of the industry is going to change. My analogy is what happened on RF front end, and that creates opportunities. So we have been engaged in very interesting conversations with some of the laser companies, especially as different forms of partnerships and strategic partnerships are going to be formed. This is actually a very exciting opportunity for Qualcomm. And you're going to see over time a lot more products from Qualcomm on optics. And we have been working on the road map that we outlined in 400 gig, 800 and 1.6 tera.

Liam Pharr

And then on the Dragonfly CPU, you mentioned new prospects beyond Meta. How should we think about the relative upside in CPUs relative to custom accelerators? And with the Arm remedies phase still pending, does the outcome have any implications for your CPU road map or cost structure?

Cristiano Amon

So let me break that question into 2. No implications on our CPU road map. I think we've been very clear about where we are with the Arm trial. We're actually very happy and with our ability to continue to develop and ship Oryon products. I will look at that as probably upside in the projections that we made on $5 billion and $15 billion, we just had Meta as a customer. Any new customer above that is an upside to the revenue target.

Operator

The next questions are from the line of Jim Schneider with Goldman Sachs.

James Schneider

I was wondering if you can maybe just kind of clarify relative to the 3 distinct hyperscale customers that you're ramping through fiscal '27, will all 3 of those be at some kind of revenue materiality exiting the year? Or will it just be maybe 2 of them and 1 will kind of fall outside the year?

Akash Palkhiwala

Yes. You should think of 2 of those customers at a very significant scale, making up that $5 billion, and they will extend into fiscal '28 and then '29 as well. So it's -- think of it as a long-term multi-generation multi-product relationship and both the relationships are at scale.

James Schneider

And then as we model fiscal '27, you provided some helpful overall EBIT margin commentary in terms of the profile of the year. But is there any way to think about OpEx? And specifically, is there any kind of material OpEx step-up that we should be thinking about during the course of fiscal '27?

Akash Palkhiwala

Yes. I think you should just think of our OpEx approach as consistent over the last year. We're guiding $2.8 billion for the September quarter. And the focus on OpEx is really kind of bringing in new skills that don't exist in the company as we diversify, especially within Data Center, and that will continue to be the focus area for us going forward. Just maybe one timing thing is just we typically have calendar year employee-related expenses reset when we get to January. So that will happen, but that's kind of consistent with previous years that you've seen.

Operator

Our final question is from the line of Chris Caso with Wolfe Research.

Christopher Caso

Yes. I guess first question is with regard to the Starlink announcement you made this morning. Could you give some more color on that in terms of timing, magnitude and what Qualcomm is providing into that venture?

Cristiano Amon

Yes. Thanks for the question, Chris. So timing, I think we talk about '27, you should think about from a financial standpoint, it will start to appear as we exit fiscal '27. It's more of a fiscal '28 from a revenue standpoint. Magnitude, I think that what we can say at this point, and I think it's probably how we both articulated into the announcement is a nationwide network in the United States and probably one of the largest build-outs of a network. It's terrestrial network.

We provide silicon content for radio, including massive MIMO as well as the accelerator for the digital unit, and we provide software on the Layer 1. It's a very -- it's a great partnership with SpaceX for next generation, I think, infrastructure. The industry has been running on dedicated equipment. I think we have seen a lot of trends of technology ranging from Open RAN to AI native as an opportunity to innovate in this space. And as we talk about it, it's a multiyear agreement and including how we're going to be thinking about the next generation of wireless as well. We're very proud of this announcement. I think it's great also from the United States technology leadership standpoint, and we're going to be executing on it.

Christopher Caso

Great. As a follow-up question, I wonder if you could give some color. And I know you talked a bit about this in the Analyst Day, but in terms of what product categories are in the Data Center revenue in the $5 billion in fiscal '27 and when other products layer in? And obviously, for near term for this year, is it all connectivity revenue? Is there -- assuming that there's no accelerator revenue in the numbers for the $5 billion that comes in '27?

Akash Palkhiwala

Chris, so for the $5 billion in '27, most of that revenue is actually custom silicon engagement for the main AI accelerator at our 2 hyperscaler customers. The next category that's material in the $5 billion in fiscal '27 is the connectivity revenue. So you should think of those as driving the FY '27 revenue ramp. When you think about the $15 billion revenue in '29, that will be inclusive of the custom silicon engagement, expansion in connectivity, but then we'll also have CPU and AI accelerator both coming in. And that's why we feel confident about achieving the $15 billion and really going beyond that as we execute on the product road map given how good our technology is looking and the interest that we're seeing from various hyperscalers.

Operator

That concludes today's question-and-answer session. Mr. Amon, do you have anything further to add before adjourning the call?

Cristiano Amon

No, I just wanted to thank our employees and our partners. As I said in my prepared remarks, this is an important, I think, close of a fiscal year for Qualcomm. And it really positions us to start fiscal '27 with all of the things that are within our control and now focus on growing in a much more diversified company. We're very excited about that. We have a lot of things we're doing in parallel, but I guess we developed that capability as we diversify the company, and we're excited about our future. Thank you.

Operator

Thank you. Ladies and gentlemen, this concludes today's conference call. You may now disconnect.