Qualcomm (QCOM.US) Q4 conference call: CFO responds to storage price increases dragging down gross margin and expects the share of mobile phone revenue to fall below 50%

The Zhitong Finance App learned that Qualcomm (QCOM.US) held a conference call on results for the fourth quarter of fiscal year 2026 (ending September) after the market on November 4, EST. Non-GAAP revenue for the quarter was $10.44 billion, higher than analysts' expectations of $10.11 billion; adjusted earnings per share (EPS) of $2.17, slightly exceeding expectations of $2.16. However, core mobile phone chip revenue was US$5.14 billion, down 26% year on year, lower than the expected value of US$5.25 billion, and the median EPS guidance value ($2.25) of the company for the first quarter of fiscal year 2027 was lower than the market forecast of $2.27. After the earnings report was released, the stock fell by more than 6% after the market.

The focus of this conference was on the impact of the increase in memory chip prices on the mobile phone business and gross margin, as well as the impact on the pace and profitability of the data center business. CFO Akash Palkhiwala (Akash Palkhiwala) responded that the company is raising product prices across the board to reflect the rise in semiconductor costs across the industry. “All revenue will be reflected in QCT gross margin in the second half of fiscal year 2027, and the underlying gross margin will return to the historical range of 48% to 50%”; the data center business “increases operating margins but dilutes gross profit margins”. At the same time, management determined that as uncertainties such as license renewal and loss of Apple's baseband are settled, Qualcomm will enter the 2027 fiscal year with a “broader and more diversified” attitude. It is expected that this fiscal year's non-mobile phone revenue will increase by more than 60% year on year, and the share of mobile phone revenue will fall below 50% for the first time.

Focus and Response

The core contradiction of this conference is a tug-of-war between “data center volume” and “gross margin pressure”: the custom silicon (custom silicon) business merged into the QCT division and began climbing in fiscal year 2027, compounding the increase in storage prices across the industry. The median value of the company's EPS guidelines for the December quarter fell short of expectations, which became the main reason for the post-market stock price decline. A number of analysts inquired about gross margin trends and the prospects of the mobile phone market. CFO Parshiwara responded that the full benefits of the price increase will be reflected in the second half of fiscal year 2027, when the underlying gross margin will return to the lower end of the 48%-50% historical range; the data center business increases operating profit margins but dilutes gross profit margins, and the partial adjustments themselves will drag down the QCT profit margin before tax (EBT margin) by about 3 percentage points. CEO Cristiano Amon (Cristiano Amon) added that price increases are mainly for new platforms and models, and are less affected by advance prebuy (prebuy). Market opinions are divided: the Financial Services Association and other media interpret weak guidelines as showing cost pressure, but there are also opinions that focus more on the space implied by management's statement that “data center upside is limited by supply rather than demand.”

Uncertainty settled, and “broader and more diversified” entered the new fiscal year

Anmon characterized FY2026 as a year of concentrated uncertainty: the completion of QTL license renewals, loss of Apple baseband orders, and the impact of storage price increases on the mobile phone industry have all been included in the business plan. Non-GAAP revenue for the year was $43.2 billion, of which QCT revenue was $37.2 billion, and total automotive and Internet of Things (IoT) revenue increased 24% year over year. The company returned $2 billion to shareholders this quarter, with repurchases and dividends of $1 billion each.

Data Center: Reiterates $5 billion target, upside is “limited by supply rather than demand”

The data center is the biggest incremental message of this conference. The company reaffirmed its revenue target of $5 billion for fiscal year 2027. It is expected to climb quarterly and withdraw from this fiscal year with an annualized operating rate of 7 billion US dollars (run rate), and target 15 billion US dollars for the 2029 fiscal year. Most of the $5 billion comes from two hyperscaler (hyperscaler) custom silicon projects, one of which is Amazon AWS — the potential revenue opportunity for both parties is as high as $60 billion in ten years. Anmon said that the upward space exceeding $5 billion “will be limited by supply rather than demand”; the High Bandwidth Compute (HBC) laboratory measured energy efficiency of about 0.3 picojoules per bit (picojoules per bit), which is better than the 0.5 design target, and the yield exceeds expectations.

Auto business: Breaking $2 billion for the first time in a single quarter

The automobile business's revenue exceeded 2 billion US dollars for the first time, up 89% year on year, and withdrew from the 2026 fiscal year with an annualized operating rate of about 8 billion US dollars, higher than the previous forecast of 7 billion US dollars. Commercial shipments of the fifth-generation Snapdragon Digital Chassis (Snapdragon Digital Chassis) began in the September quarter, and the silicon content of bicycles was eight times higher than that of third-generation platforms. Management expects Qualcomm to rank among the world's largest automotive semiconductor suppliers in terms of revenue in fiscal year 2027.

Mobile phone and storage price increases: full reflection of price increases in the second half of the year

Mobile phone chip revenue of 5.1 billion US dollars was squeezed by rising storage prices across the industry. Management expects the global mobile phone market to naturally decline from two digits to the middle double digits in 2026 and remain flat to a slight decline in 2027; China's Android revenue bottomed out in the June quarter, and has since rebounded sequentially for two consecutive quarters. Qualcomm's share baseline among Samsung's high-end models has moved from 50% to over 70%. In response to the question of “whether the price increase triggered early stocking”, Anmon responded that the price increase was mainly aimed at new platforms and models. The impact was low, and channel inventory was still low.

Authorized business: Three major contract renewals settled, operating rate unchanged

QTL completed a landmark conclusion this quarter: long-term licensing agreements with Apple, Huawei, and Voice have all been renewed, and there are no renewal arrangements with major mobile phone manufacturers for the next few years. The CFO clearly stated that “the operating rate of revenue remains the same” should be understood as the renewal of the contract maintaining the original tax rate; the next renewal cycle will begin in FY2030 and continue through FY2033. Management said that the licensing business provided a lasting and highly profitable foundation for Qualcomm to move into the 6G era.

SpaceX and Infrastructure: Starlink 5G Big Deal

Qualcomm and SpaceX have reached a multi-year agreement to provide a Dragonwing (Dragonwing) infrastructure platform for Starlink Mobile (Starlink Mobile)'s nationwide terrestrial 5G network, covering large-scale MIMO (massive MIMO) RF silicon, digital unit accelerators, and Layer 1 (Layer 1) software. The deployment will begin in 2027, or one of the largest 5G network constructions in the US; financially, related revenue will be shown starting at the end of the 2027 fiscal year, mainly reflected in the 2028 fiscal year.

Arm lawsuit: jury finds Arm in breach of contract

At the meeting, Anmon responded to the lawsuit with Arm: the jury found that Arm violated the parties' long-term architecture license agreement (ALA), failed to fulfill promises, deliberately interfered with Qualcomm's business, and constituted unfair competition.” The result is a victory over innovation and competition.” He said that the outcome of the lawsuit will not affect the CPU roadmap. Qualcomm will continue to deliver Oryon CPUs from mobile to data centers, and new CPU customers are all upside for existing revenue targets.

Next Season Guidelines and Segment Adjustments

The company expects revenue of 10.4 billion to 11.2 billion US dollars, non-GAAP EPS of 2.15 to 2.35 US dollars (median value lower than LSEG's estimate of 2.27 US dollars); QCT revenue of 9 billion to 9.6 billion US dollars (profit margin before tax of 20%-22%, including 3 percentage point impact of data center mergers), QTL revenue of 1.3 billion to 1.5 billion US dollars (profit margin before tax 72% - 76%), data center revenue of more than 650 million US dollars, and non-GAAP operating expenses of about 2.8 billion US dollars. In fiscal year 2027, non-mobile phone revenue will increase by more than 60% year-on-year, accounting for more than 50% of QCT revenue.

The full details of Qualcomm's FY2026 fourth quarter results conference call are as follows:

Management statement

Brett Simpson (Qualcomm Senior Vice President of Investor Relations): Welcome to Qualcomm's 2026 fourth quarter results conference call. Today, Anmon and Parshiwala will prepare to speak, and Rogers will participate in the Q&A. This conference call will use non-GAAP (non-GAAP) financial indicators as defined by Regulation G, and the relevant GAAP reconciliation statements have been published on the company's website; the conference contains forward-looking statements. Actual results may differ materially from expectations; please refer to the company's latest 10-K document.

Ammon (Qualcomm President and CEO):

Thanks Brett and good afternoon everyone. In the fourth fiscal quarter, we achieved non-GAAP revenue of $10.4 billion and non-GAAP earnings per share (EPS) of $2.17. The chip business (QCT) revenue was $9 billion, up 6% month-on-month, and is at the upper end of the guidance range. Authorized Business (QTL) revenue of $1.3 billion. For the full fiscal year 2026, we reported non-GAAP revenue of $43.2 billion and QCT revenue of US$37.2 billion, of which total automotive and Internet of Things (IoT) revenue increased 24% year over year.

As we close the 2026 fiscal year, many of the uncertainties that have been suppressing the business have settled, including the renewal of QTL licenses, the loss of Apple as a customer, and the negative impact of memory pricing on the scale of the mobile industry. Regarding the latter, the June quarter marked a low point for our Android business in China, and we have seen month-on-month growth this quarter. Looking ahead, our plan assumes that global mobile phone shipments will remain flat to decline in 2027. The timing and shape of Apple's baseband transition has been fully incorporated into our business plans, and all parties have fully understood this. We've also renewed important QTL licensing agreements, including Apple and Huawei.

We are now entering the 2027 fiscal year with strong momentum and significant opportunities as a broader and more diverse Qualcomm. First, QCT's non-mobile phone revenue will increase by more than 60% year-on-year this fiscal year. Second, the data center business has begun to expand, and revenue for the December quarter is expected to exceed 650 million US dollars. Third, the automobile business withdrew from the 2026 fiscal year with an annualized revenue operating rate of 8 billion US dollars, exceeding expectations of 1 billion US dollars. Fourth, we have laid a solid foundation on the edge of intelligent AI (agentic AI) transformation and spawned new product categories. Finally, we reached a landmark agreement with SpaceX to support the latter's nationwide terrestrial 5G network in the US.

Our technology investments are in line with industry trends. We have made a number of key acquisitions to strengthen our hardware and software capabilities and gradually evolve into a full-stack semiconductor company. We announced strategic partnerships with industry leaders such as Meta, Amazon, BMW (BMW), and Stellantis (Stellantis), and there will be more in the future; the company's overall design win (design win) pipeline increased 70% year over year. This is the foundation for our confidence to meet our FY2029 target — $40 billion in non-mobile revenue, of which more than $15 billion comes from data centers.

Now let's focus on what investors are most concerned about. First is the licensing business, and we reached a milestone this quarter. We are extremely satisfied with the renewal of our contracts with Apple, Huawei, and Voice. These agreements provide stability for Qualcomm's QTL business moving into the 6G era, and also demonstrate the enduring value of our intellectual property and patent portfolio. As with every previous generation of technology, we expect our invention to once again be the foundation for 6G. Importantly, our QTL revenue run rate remains the same. The next renewal cycle begins in FY2030 and continues through FY2033. The licensing business provides us with a long-lasting, high-profit foundation for cultivating the next batch of growth opportunities on a large scale.

Data centers are another area where we've made significant strides this quarter. It's one of our top strategic priorities and something I personally put a lot of effort into; we're building this business around long-term customer relationships, a broad product portfolio, and an open software ecosystem. AWS is a great example: our agreement spans multiple generations of products, covering custom silicon (custom silicon) and optical interconnections for AI inference. Shipments of the custom accelerator began in October, and the optical interconnection solution with AWS will be launched later this month. We've also targeted a second optical interconnect customer. With Qualcomm High Bandwidth Compute (HBC) technology, our AI250 chip exceeded expectations in terms of energy efficiency and bandwidth, and the early yield was in line with the plan. These results are attracting strong interest from AI customers and storage partners, and we have added two significant new prospects this quarter.

Our Dragonfly C1000 CPU is still shipping to Meta as planned, and we've reached out to a new major potential customer. We are optimistic about the commercial opportunities for existing high-speed interconnect solutions, such as optical DSPs, and are investing in co-deploying optical devices directly with our XPU and custom silicon, as we demonstrated with Lumentum and Corning (Corning) at ECOC 2026. On the software side, our full stack integration with Modular is progressing smoothly. As an open multi-vendor technology stack, Modular has been running on NVIDIA and AMD GPUs, Amazon Trainium, and Google TPU, and its performance is impressive; optimizations for Qualcomm accelerators are also progressing. Last month, Modular opened up the full MAX source code, which builds on the August open source Mojo language. We are also happy that Modular Cloud went live in the September quarter and currently processes tens of trillions of tokens every month. We'll share more details on how our ecosystem partners use Modular Cloud later.

The data center business is off to a strong start. We are reaffirming our revenue target of $5 billion for FY2027, with good order coverage to support continued revenue growth in FY2028.

On the automotive side, we withdrew from the 2026 fiscal year with an annualized revenue operating rate of about $8 billion, far higher than the $7 billion forecast provided last quarter. Furthermore, automobile revenue increased nearly 90% year over year in the fourth quarter of the fiscal year. Commercial shipments of the fifth-generation Snapdragon Digital Chassis (Snapdragon Digital Chassis) began in the September quarter and will continue to climb in the 2027 fiscal year. As a reminder, this is Qualcomm's most advanced platform to date, with significant performance improvements, and the biggest increase in bike content in our history. Our products have also evolved from SoCs to integrated computing modules. Customer demand remains strong across the product portfolio, and Qualcomm is expected to rank among the world's largest automotive semiconductor suppliers in terms of revenue in fiscal year 2027.

In the field of robotics, our high-end Dragonwing (Dragonwing) IQ10 processor is about to begin commercial climbing, and customer momentum is strong. This quarter, Eco Partners demonstrated an AI model with up to 80 billion parameters running directly on the IQ10. In the 2027 fiscal year, we will continue to expand the technology and capabilities of robotics platforms.

As we can expect, the popularity and capabilities of AI agents are rapidly increasing, and consumer adoption is growing accordingly. A typical example is Meta Muse, which has over 10 million downloads in just 40 days. We are also seeing more and more companies focusing on smart device design experiences and devices, including new forms of personal AI such as smart phones and smart glasses — Meta's audio AI glasses, Muse Charm, Gemini glasses, and Alibaba's Qwen (Qwen) AI glasses are recent examples, and there will be more in the coming year. A smartphone is an essential device for the smart experience: it knows the user best, is portable, and grasps the user's context, preferences, and authorization; devices are also evolving to support the user's own operation and act on behalf of the user. Delivering these experiences requires greater computing performance, energy efficiency, and connectivity, which is where Qualcomm excels. More customers have built smart devices into their flagship devices. We are happy to share that the world's first smart phones have been released on the Snapdragon platform.

Entering the 2027 fiscal year, we are satisfied with the share position of high-end Android, and expect that the share will further increase as smart phones transition to promote the improvement of stand-alone computing content. As a reference, leading AI companies are aiming to launch smartphones in 2028 that can run 100 billion parameter models locally and achieve all-day battery life. Achieving this goal will require incremental improvements in terminal computing performance.

Finally, I'm very proud to share that SpaceX chose Qualcomm to support its Starlink Mobile (Starlink Mobile) US nationwide terrestrial 5G network. Starlink is building one of the most ambitious new cellular networks in the U.S., and our Dragonwing infrastructure platform will combine advanced massive MIMO (massive MIMO), dedicated RAN processing, and commercial-grade Layer 1 (Layer 1) software. Under the multi-year agreement, the deployment will begin in 2027 and is expected to be one of the largest 5G deployments in the US. Together, we will advance America's wireless leadership, starting with the integration of 5G and satellite and terrestrial connectivity. This new architecture lays the foundation for the next generation of connectivity and networking intelligence. This is a major win for our infrastructure business and the beginning of a strategic relationship with huge potential.

Before I close, I'd like to talk briefly about our lawsuit with Arm. We are pleased that the jury found that Arm had breached its long-term architecture license agreement (ALA) with Qualcomm, failed to meet its promises, deliberately interfered with Qualcomm's business, and engaged in unfair competition — regardless of its status as a supplier or competitor. Combined with Qualcomm's previous court win, this result further revealed Arm's intention to damage Qualcomm and its ecosystem. The result was a victory for innovation and competition. We will continue to focus on developing and delivering industry-leading semiconductor products to customers around the world, including Oryon CPUs from mobile to data centers.

In summary, as we enter the 2027 fiscal year, we have made significant progress towards the growth plan set out on Investor Day. Since then, we have signed a major strategic agreement and positioned our data center business on a large scale. As we build Qualcomm's next chapter, we will continue to focus on execution and are confident of meeting our 2029 goals.

Akash Parshiwara (Qualcomm Executive Vice President, Chief Financial Officer and Chief Operating Officer):

Thanks Cristiano and good afternoon everyone. First, let me introduce the results for the fourth fiscal quarter. We achieved non-GAAP revenue of $10.4 billion and EPS of $2.17, and revenue was at the upper end of the guidance range. QTL's revenue of $1.3 billion and profit margin before tax (EBT margin) of 70% is in line with expectations. QCT's revenue of $9 billion is at the upper end of the guideline range, with a profit margin of 24% before tax, in line with expectations. QCT's mobile phone revenue of US$5.1 billion reflects the impact of industry-wide storage dynamics on the global smartphone market.

QCT's automotive business delivered its first $2 billion revenue quarter, up 89% year over year, mainly driven by the growing popularity of integrated module products for fifth-generation computing platforms. QCT's IoT revenue of $1.9 billion, up 4% year over year, was driven by industrial, networking, and robotics categories — this category grew 20% year over year, partially offset by the impact of storage supply constraints on consumer products. QCT's total non-mobile revenue (including automobiles and IoT) increased 36% year over year, demonstrating the continued execution of the growth plan. In the end, we returned $2 billion to shareholders, including $1 billion each in repurchases and dividends.

Before moving to guidance, update a few factors already reflected in financial performance. First, in terms of QTL, we have completed long-term licensing agreements with Apple, Huawei, and Voice, and there are no renewal arrangements with major mobile phone OEMs for the next few years. After the agreement was signed, QTL's revenue operating rate and long-term revenue forecast remained unchanged. Second, as described last quarter, we are implementing price increases across our entire product line to reflect higher semiconductor investment costs; we expect the full benefits of these initiatives to be reflected in QCT gross margin in the second half of fiscal year 2027. Finally, fiscal year 2027 will be a major milestone for the data center business—moving from customer engagement to revenue fulfillment.

First-quarter guidance reflects the start of shipments of custom silicon projects to two hyperscaler customers. As we have discussed previously, these custom silicon projects will increase QCT operating margins but dilute gross margins. These factors are reflected in the fourth quarter results and the first quarter outlook.

First fiscal quarter guidance: We expect revenue of $10.4 billion to $11.2 billion, non-GAAP EPS of $2.15 to $2.35. QTL's revenue is expected to be $1.3 billion to $1.5 billion, with a profit margin of 72% to 76% before tax, reflecting normal seasonality. QCT's revenue is expected to be between $9 billion and $9.6 billion, with pre-tax profit margins of 20% to 22%. Note that the QCT profit margin guidelines include an impact of 3 percentage points, stemming from the merger of data centers into the QCT division starting in FY2027, and reflect investment in AI accelerators and CPU platforms before revenue is released. Importantly, this division adjustment had no impact on the company's overall guidance and QCT's long-term pre-tax profit margin target of 30%. We expect QCT mobile phone revenue to drop by a single digit percentage, and the increase in Android revenue will be offset by the continued decline in Apple product revenue. QCT data center revenue is expected to exceed $650 million, driven by custom silicon customers starting to ship shipments. The combined revenue of QCT Automotive and IoT is expected to be roughly the same from month to month. Finally, we estimate non-GAAP operating expenses of approximately $2.8 billion.

Before concluding my remarks, outline progress in implementing the growth strategy. In the 2027 fiscal year, we expect to achieve two key milestones in revenue diversification. First, QCT's non-mobile revenue (automobiles, IoT and data centers) has increased by more than 60% year-on-year since the first fiscal quarter. Second, non-mobile phone revenue will account for more than 50% of QCT revenue this fiscal year. In terms of data centers, we are implementing the targets set on Investor Day, including US$5 billion for FY2027 and US$15 billion for FY2029; data center revenue is expected to climb quarterly in FY2027 and reach an annualized operating rate of US$7 billion when we exit this fiscal year. We're also expanding our customer pipeline to more global hyperscale cloud vendors, including a long-term strategic agreement with Amazon AWS — potential revenue opportunities of up to $60 billion over the next ten years.

This quarter, we also successfully validated high-bandwidth computing technology in laboratory tests, a key milestone in delivering innovative AI inference accelerator solutions to solve data center storage bandwidth challenges. We are also well-positioned to lead other key industry trends, including the transition from consumer terminals to smart AI, and the adoption of physical AI in automotive, industrial, and robotics. Finally, thank our employees for their outstanding execution and innovation in FY2026, whose efforts enabled us to deliver industry-leading products to customers around the world.

Q & A session

Joshua Buchalter (TD Cowen): Let's start with gross margin. Last quarter, you predicted a step-down in gross margin brought about by the start of the ASIC business. But looking at Latong throughout the year — in simple terms, to reach an operating rate of 7 billion US dollars after leaving this fiscal year, it means that the data center business will nearly triple. As the scale of the data center business grows and custom silicon aggressively climbs during the year, should gross margin be expected to continue to decline?

Akash Parshiwara: Josh, I'm Akash. Let me first give the company's overall gross margin framework. As stated in the previous quarter, we are implementing price increases to reflect higher semiconductor costs across the industry; related benefits began to be reflected in the second fiscal quarter and fully realized in the second half of fiscal year 2027. The mechanism is like this: as the price increase takes effect, our underlying gross margin will return to the historical range of 48% to 50%, which is currently at the lower end of this range. As for the data center, as mentioned before, it increases the operating profit margin but dilutes the gross profit margin, which will lower the gross margin on top of the above framework. This is in line with our statement last quarter, and we are implementing as planned.

Joshua Buchalter (TD Cowen): I'm excited to see this week's HBC demo. Which workloads initially received the most feedback on accelerator products? Any updates on commercialization progress? According to the plan, it should start contributing in the second half of this fiscal year — that is, there are only a few quarters left.

Ammon: Thanks, I'm Cristiano. We are very satisfied with HBC's progress, which is actually better than what investors said: the design target is about 0.5 picojoules per bit (picojoules per bit), laboratory measurements are about 0.3; the yield exceeds expectations, and the technical progress is satisfactory. The workload receiving the most positive feedback is decoding (decoding) — the decoding process is extremely efficient. We're in deep contact with customers: when you go from just PPT to having real silicon in your hands, the nature of the conversation changes. This has not only strengthened our confidence in the future of customized ASICs, but also strengthened our confidence in commercial product solutions aimed at 2029. Another interesting change is that the combination of HBC and custom accelerators is now also entering into customer conversations. Overall, customer engagement was very positive.

Joseph Cardoso (J.P. Morgan Chase Research Department): I'd like you to talk about how data center revenue climbed during the year, and how did customers fit into the guidelines at the pace? Also, you've talked about two customers in the past, and the third one appeared on the slide -- is the third one guiding the beginning of this fiscal year, or is it being scaled at another point in time?

Akash Parshiwara: Joseph, I'm Akash. Let's talk about finance first: The revenue forecast of $5 billion for fiscal year 2027 has already been given. The Q1 guideline exceeds US$650 million. Revenue will rise quarterly during the year. When we exit this fiscal year, we will annualize about 7 billion US dollars and enter fiscal year 2028. Supply-side preparations are sufficient to implement the $5 billion target; the upward portion exceeding $5 billion will be limited by supply rather than demand. In FY2028, we started with $7 billion, and the FY2027 customer relationship will continue until FY2028. The design win list coverage is very strong, enough to support strong year-on-year growth in FY2028. The target of $15 billion for the 2029 fiscal year, and current customer engagements will also give us a high level of coverage.

Anmon: A brief additional customer level: Most of the $5 billion in fiscal year 2027 came from two hyperscale customers, one of which is Amazon. Details of silicon plus interconnect have been made public. We also have another accelerator customer and are in in-depth negotiations on custom silicon for subsequent generations as part of the climb towards FY2028 and 2029. In the manuscript, we also mentioned that in addition to Meta, CPU also has new potential customers and is currently in touch.

Joseph Cardoso (J.P. Morgan Chase Research Department): From the end of this fiscal year to the next fiscal year, will growth mainly be compounded by price increases? How should the magnitude and impact point of price increases be understood? At a larger level, this year's market data shows that the increase in storage prices is shrinking the mobile phone market. What is the initial judgment on the health of the market next year?

Akash Parshiwara: In terms of the overall market, from 2025 to 2026, in line with other industry analysts, we expect a decline from a low double digit to a medium double digit; entering 2027, we expect it to remain flat to a slight decline — this has always been our business planning assumption, and the current evolution is in line with expectations. Specifically, when it comes to QCT Android revenue, as you all know, we called the June quarter the low point of QCT mobile phone revenue and Android in China. Since then, revenue for the September and December quarters has been growing, which is encouraging. Our plans have always been conservative and are now evolving exactly as expected.

Joseph Moore (Morgan Stanley Research): Regarding royalty rates, you said that the QTL revenue operating rate remains the same — can you understand that the rate for each newly contracted device is similar to before? I'd like to confirm this.

Akash Parshiwara: Joe, we won't reveal contract details; they are confidential; but we said the revenue utilization rate hasn't changed—you can understand that we kept our rates and royalties during the renewal.

Joseph Moore (Morgan Stanley Research): Very helpful. Also, with regard to the warrant structure in the Amazon agreement, do you see it as a way to cooperate with other customers? Will there be similar warrant arrangements in the future, or will they be limited to this particular relationship?

Akash Parshiwara: Looking at our overall data center contacts, they are all very traditional semiconductor industry customer cooperation models. Amazon's warrants (warrants) structure is an exception — we issue at-the-money (at-the-money) warrants, which is the only exception we currently have. In the future, we will evaluate other structures on a case-by-case basis for data center cooperation, but the overall framework should be understood to be consistent with the traditional semiconductor model.

Timothy Arcuri (UBS Investment Banking Research): Akash, can you give me an outlook for the March quarter? I know you don't want to guide March, but in the last call, you said that March revenue will increase — Apple's comparison base is easy, and Android should be better. Can you give it a feel?

Akash Parshiwara: Tim, you're right, we're not giving guidance for the March quarter right now, but there are a few things that might help. First, in the past, December was usually our strongest quarter of the year; as Apple product revenue declined, we no longer expected December to be a full year high. Second, data center revenue in Q1 exceeded 650 million US dollars and climbed quarterly during the year, which determined the quarterly pace to a certain extent. Third, from 2026 to 2027, non-mobile phone revenue such as automobiles plus IoT plus data centers is expected to increase by more than 60%, which is enough to cover the decline in revenue from Apple products. Finally, mobile phone revenue for the 2027 fiscal year is expected to be less than half of the company's total revenue — a key milestone in the diversification plan, and we're excited to achieve this fiscal year.

Timothy Arcuri (UBS Investment Banking Research): Understood, the last time you said there would be growth in March, but now you don't want to promise.

Akash Parshiwara: I think the market's current framework for this business is reasonable.

Stacy Rasgon (Bernstein Research): I want to dig deeper into the mobile phone market outlook. The flat expectations surprised me a bit - Q3 market revenue is likely to drop by more than 20% year over year, and Q4 doesn't look very good; your own Chinese Android revenue has been growing since June or September, and you are still raising prices. How to rule out the possibility that Android will be stocked in advance before the price increases? I'm a little confused about the phone outlook, any additions would be helpful.

Akash Parshiwara: Stacy, talk in two parts: first talk about the market, then talk about QCT revenue. The overall market declined from a low double digit to a medium double digit in 2025-2026, and is expected to remain flat to a slight decline in 2026-2027 — a prudent planning assumption given the uncertainty of the overall environment. Specifically, when it comes to QCT revenue, the impact in FY2026 came not only from market changes, but also from inventory removal from OEM channels; we don't expect this impact in FY2027 because we have been tracking channel inventory — we also have QTL data that can be cross-verified, and channel inventory is still low. Therefore, we expect QCT revenue to be consistent with overall market trends.

Ammon: Stacy, I'll add one more point. Due to long-term partnerships with customers, the reason why many price increases are higher than cost increases is because we mainly price new platforms and models; price increases are only fully reflected in the figures for the second half of the year, so they are less affected by prebuy (prebuy). As Akash said, according to the market data we saw, the June quarter was still the low point for Android in China.

Stacy Rasgon (Bernstein Research): You mentioned that Samsung's market share framework is 75%. When the next generation platform is launched in 2027, will the share framework still be the same?

Ammon: Yes, absolutely. The way to think about Samsung's business is simple: in the past, our baseline was 50%; now this baseline is no longer 50%, but more than 70% — this is how we plan our Galaxy business for future generations, and I also recommend that you plan the same way. Additionally, we are expanding our scope of business with Samsung: traditionally we are not the main supplier of all wearables and other devices, and you'll see this change with high-end Galaxy designs.

Benjamin Reitzes (Melius Research): Another question (because Stacy just asked Samsung). The automobile business surpassed expectations this quarter, and you also mentioned that prices are rising - does the automobile business also prepare goods in advance? Is this performance sustainable throughout the year?

Anmon: No, great question. We talked about the transition to fifth-generation platforms last quarter. To give you an intuitive number: Currently, the models on the road are equipped with our third-generation platform. The fifth-generation platform contains eight times the silicon content of the third-generation platform — the computational content of the car has increased dramatically. Most of our automotive growth comes from increased share, a process where the design-win pipeline is converted into revenue; the stepwise growth in SoC revenue will continue.

Benjamin Reitzes (Melius Research): So Akash, how did the QCT profit margin before tax (EBT) start at 20% to 22% of the guideline, evolve throughout the year against the backdrop of an increase in the share of data centers? How should I model it?

Akash Parshiwara: Ben, take this opportunity to clarify: We've made changes to the division, and QCT now includes the data center — in FY2026 it's outside of QCT. This change alone reduced QCT's operating margin by about 3 percentage points. Of course, we are investing in several product lines before revenue is increased. This doesn't change Qualcomm's overall numbers; it's just a change within the division. We don't guide operating margins according to quarterly or annual guidelines, but QCT's operating margin target remains the same and is still 30%; as revenue is scaled up over the next two to three years, we are confident that it will be achieved.

Liam Pharr (Bank of America Securities Research Division, asked on behalf of Vivek Arya): Regarding network interconnection, you mentioned AWS optical interconnect capacity and the second optical interconnect customer. Can it help quantify the size of the interconnect business today and the 2027 and 2028 fiscal year growth as it evolves to 224G and 1.6T? Does the integration of the optical interconnect industry have an impact on you? From an equity perspective, is there a natural synergy between interconnect and lasers?

Ammon: Very good question. We're not splitting, but we should assume that interconnect revenue is included in $15 billion (FY2029) and $5 billion (FY2027). We are very optimistic and continue to invest in it. This asset comes from the acquisition of Alphawave — I think Qualcomm is the natural destination for communications technology. At ECOC 2026 this quarter, we and Lumentum demonstrated the progress of optics and co-packaged optics (co-packaged optics). You mentioned a good point: the structure of the industry will change, and my analogy is what happened in the RF front end (RF front end), which will create opportunities. We are having very interesting conversations with laser companies, and in particular, various forms of partnerships and strategic alliances are being formed. This is an exciting opportunity for Qualcomm. In the future, you will see Qualcomm launch more products in the field of optics; we are continuing to advance the 400G, 800G, and 1.6T roadmaps.

Liam Pharr (Bank of America Securities Research Division): Regarding Dragonfly CPUs, how do new potential customers other than Meta view the upside of CPUs compared to customized accelerators? The Arm relief phase isn't over yet, will the results affect the CPU roadmap or cost structure?

Ammon: Answer in two parts. First, there's no impact on our CPU roadmap; we're already very clear about Arm's position on the lawsuit, and we're very satisfied with our ability to continue developing and shipping Oryon products. Second, I think new customers should be viewed as an upward space outside of the $5 billion and $15 billion target — currently we only have Meta CPU customer, and any new customer is an upward trend that exceeds the revenue target.

James Schneider (Goldman Sachs Research): Please clarify: compared to the three hyperscale customers that are climbing in FY2027, did all three reach revenue importance when they left this fiscal year, or were only two and the other outside of the fiscal year?

Akash Parshiwara: Understandably, two of these customers will reach a very impressive size, form a $5 billion entity, and will continue until the 2028 and 2029 fiscal year — this is a long-term, intergenerational, cross-product partnership, and both have taken shape.

James Schneider (Goldman Sachs Research): When modeling the 2027 fiscal year, apart from the overall review of operating profit margins, should there be a major step-up in OpEx?

Akash Parshiwara: Our OpEx approach is consistent with the past year: the September quarter guidance was $2.8 billion. OpEx's focus is on introducing the new capabilities needed to diversify the company, particularly in the data center sector, and this will continue to be an investment direction. There is one point in time: Every year in January, employee-related expenses for the calendar year are reset, which is consistent with previous years.

Christopher Caso (Wolfe Research): First question about Starlink (Starlink) announced this morning: Can you provide more details on the timing, scale, and what Qualcomm is in this project?

Anmon: In terms of time, we're talking about 2027 — from a financial perspective, revenue will start showing at the end of fiscal year 2027, and more in fiscal year 2028. In terms of scale, it can currently be said that this is a terrestrial network covering the entire United States, and is probably one of the largest network constructions in the US. We provide silicon content for the RF portion, including massive MIMO, accelerators for digital units, and Layer 1 (Layer 1) software. This is a fantastic partnership with SpaceX on next-generation infrastructure. The industry has been running on dedicated equipment for a long time, and we have seen many innovative trends from Open RAN to native AI. This is a multi-year agreement and includes our shared plans for the next generation of wireless technology. We are very proud, and this is also important from the perspective of America's technological leadership, and we will do our best to implement it.

Christopher Caso (Wolfe Research): Question: About the data center's revenue composition of $5 billion in fiscal 2027 — is this year mostly interconnect revenue? Is custom accelerator revenue not among them? When will other products be added in?

Akash Parshiwara: Chris, out of the $5 billion in fiscal year 2027, most of the revenue actually comes from two hyperscale customers' main AI accelerator custom silicon projects; the next significant category is interconnection revenue. These two categories drive the 2027 fiscal year's climb. As for $15 billion in fiscal year 2029, it will include the expansion of custom silicon projects, expansion of interconnection revenue, and the addition of CPUs and (commercial) AI accelerators — given our technical performance and the interest of various hyperscale vendors, this is where our confidence lies in reaching and surpassing $15 billion.

Anmon (concluding remarks): I just want to thank our employees and partners. As stated in the preparation statement, this is the end of an important fiscal year for Qualcomm, allowing us to enter the 2027 fiscal year and focus on making a more diverse company continue to grow. We're very excited about the future. Thank you all.