Arm 2027 회계연도 2분기 실적 발표: 매출 27% 증가, AGI CPU 전망 강화

핵심 요약

• Arm은 2027 회계연도 2분기 매출이 전년 동기 대비 27% 증가한 14억 4,000만 달러로 사상 최대를 기록했다고 발표했다. 이는 5개 분기 연속 20%를 상회하는 성장세다.
• 라이선스 및 기타 매출은 50% 증가한 7억 7,300만 달러를 기록했다. 경영진은 라이선스 계약을 맺은 설계가 생산 단계에 진입함에 따라 이러한 모멘텀이 향후 로열티 매출의 선행 지표가 될 것으로 보고 있다.
• 로열티 매출은 8% 증가한 6억 7,000만 달러를 기록했다. 스마트폰 수요 둔화 속에 데이터센터 로열티는 2배로 증가했고 피지컬 AI 로열티는 50% 성장했으며, 에지 AI 로열티는 대체로 보합세를 나타냈다.
• Arm은 2027 회계연도와 2028 회계연도에 걸쳐 최대 20억 달러의 Arm AGI CPU 매출을 달성할 수 있다는 자신감을 높였다. 초기 상업용 납품 목표 시점은 2027 회계연도 4분기로 유지되고 있다.
• 비GAAP 영업이익은 37% 증가한 6억 3,800만 달러, 비GAAP EPS는 33% 증가한 0.52달러를 기록했다. 비GAAP 영업이익률은 전년 동기 대비 300bp 확대된 44.2%를 기록했다.
• 2027 회계연도 3분기 실적과 관련해 경영진은 매출 15억 4,000만 달러(±5,000만 달러), 비GAAP EPS 0.56달러(±0.04달러)의 가이던스를 제시했다.

주요 재무 데이터

사업 및 운영 실적

Arm은 해당 분기 동안 3건의 신규 Arm 토탈 액세스(Arm Total Access) 계약을 체결해 활성 ATA 구독 건수가 60건을 넘어섰다. 또한 데이터센터용 반도체 2건, 스마트폰용 반도체 2건 등 4건의 컴퓨팅 서브시스템 라이선스 계약을 추가로 체결했다. 연간 계약 가치(ACV)는 전년 동기 대비 12% 증가했다.

클라우드 AI는 여전히 최대 로열티 성장 동력으로 작용했다. Arm은 Arm 기반 CPU가 AI 인프라와 범용 클라우드 워크로드 전반으로 확대되면서 데이터센터 로열티가 2배로 증가했다고 밝혔다. 경영진은 구글 악시온, AWS 그래비톤, 마이크로소프트 코발트, 엔비디아 플랫폼에 힘입어 주요 하이퍼스케일러 내 Arm 기반 서버 점유율이 2026년 말까지 60%를 넘어설 것으로 예상하고 있다.

해당 분기 동안 Arm AGI CPU 수요가 강화되었다. 회사는 2027년까지 복수의 고객사로부터 구매 주문을 확보하고 공급망 전반에 걸쳐 공급 계약을 체결했다. 메타가 여전히 주요 고객사이며, 경영진은 다른 고객사들도 참여하고 있다고 밝혔다. 오라클 클라우드 인프라스트럭처는 2027년에 Arm AGI CPU 인스턴스를 제공할 계획이다.

전장 및 로보틱스 도입 확대에 힘입어 피지컬 AI 로열티 매출은 50% 증가했다. 메모리 가격 상승으로 특히 중저가 기기를 중심으로 스마트폰 출하량이 압박을 받으면서 에지 AI 로열티는 거의 보합세를 나타냈다. 다만 더 높은 Armv9 및 CSS 로열티율이 이러한 부진을 상쇄했다.

파트너사들의 Arm 기반 반도체 출하량은 단일 분기 기준 처음으로 100억 개를 돌파해 전년 동기 대비 15억 개 증가했다. 경영진에 따르면 Arm 기반 PC는 전 세계 출하량의 20%에 달했다.

경영진 가이던스

경영진은 2027 회계연도 3분기에도 클라우드 AI와 피지컬 AI가 강세를 이어갈 것으로 예상하는 한편, 스마트폰 수요 둔화와 전년 동기의 높은 기저가 로열티 성장에 부담으로 작용할 것으로 내다봤다.

Arm AGI CPU의 경우 2027 회계연도 4분기에 출하를 시작하는 것을 목표로 하고 있다. 경영진은 추가적인 고객 주문과 공급 가용성에 힘입어 2027 회계연도와 2028 회계연도에 걸쳐 최대 20억 달러의 매출을 달성할 수 있다는 확신을 높였다.

회사는 1세대 Arm AGI CPU의 매출총이익률이 대략 30~40% 범위일 것으로 예상하고 있다. 경영진은 Arm이 자체적인 전문성을 더 구축하고 파트너사에 대한 의존도를 낮춤에 따라 후속 버전에서는 마진이 시장 표준 수준인 50% 이상으로 올라설 수 있다고 밝혔다.

리스크 및 주요 점검 요인

• 메모리 가격 상승은 특히 저가형 부문을 중심으로 스마트폰 출하량 수요를 계속 압박하고 있다.
• Arm AGI CPU의 성장은 웨이퍼 생산능력, 기판, 테스터, 메모리 및 기타 공급망 부품 확보에 달려 있다.
• 메모리 및 시스템 비용 상승이 Arm AGI CPU 마진을 압박할 수 있으나, 경영진은 가격 책정에서 어느 정도 유연성이 있다고 밝혔다.
• 데이터센터 로열티는 아마존, 구글, 마이크로소프트, 엔비디아에 집중되어 있으며, 경영진은 이들을 최대 기여 기업으로 지목했다.
• 라이선스 매출은 대형 계약의 체결 시점과 회계 처리에 따라 분기별로 크게 달라질 수 있다.
• 1세대 Arm AGI CPU의 예상 매출총이익률은 후속 세대에 대한 경영진의 장기 목표치보다 낮다.

애널리스트 Q&A 주요 내용

경영진은 Arm AGI CPU 매출에 대한 확신이 높아진 배경으로 견조한 수요, 복수의 고객사로부터 확보한 구매 주문, 개선된 공급 가시성을 꼽았다. Arm은 수 개 분기 동안 웨이퍼, 메모리, 기판 및 테스트 생산능력 전반에 걸쳐 공급업체들과 협력해 왔다.

x86과의 경쟁에 대해 르네 하스 CEO는 와트당 성능, 비용당 성능, 그리고 고객이 맞춤형 Arm 기반 시스템을 설계할 때 누릴 수 있는 유연성을 강조했다. Arm이 인용한 독립 기관 Signal65의 테스트 결과에 따르면 Arm AGI CPU는 AMD 에픽(EPYC) 9755 대비 랙당 약 2배의 성능을 발휘하는 것으로 나타났다.

경영진은 데이터센터 로열티 성장이 시장 점유율 확대와 함께 가속기에 연결되는 헤드 노드로 사용되는 CPU가 x86에서 Arm으로 전환되는 흐름에 힘입은 것이라고 설명했다. Arm은 향후 수년간 데이터센터 로열티가 세 자릿수 성장률을 기록할 수 있다고 자신감을 나타냈다.

향후 AGI CPU 제품군은 약 18~24개월의 생산 주기를 따를 것으로 예상된다. 아울러 Arm은 랙 수준의 시스템 성능에 집중하면서 네트워킹, 커넥티비티, 고성능 I/O 기능도 검토하고 있다.

스마트폰의 경우, 경영진은 Armv9 채택률이 50%를 넘어서고 최신 CSS 제품이 더 높은 로열티율과 반도체 가격을 적용받음에 따라 Arm의 로열티 사업이 기저 시장보다 빠르게 성장할 것으로 내다보고 있다.

실적 발표 컨퍼런스콜 전문

전체 실적 발표 컨퍼런스 콜 녹취록

경영진 발표

Operator

Good day, and thank you for standing by. Welcome to the ARM's Second Quarter Fiscal Year 2027 Webcast and Conference Call. [Operator Instructions] Please be advised that today's conference is being recorded.

I would now like to hand the conference over to your first speaker today, Ian Thornton, Vice President of Investor Relations. Please go ahead.

Ian Thornton

Thank you very much, and welcome to our second quarter fiscal '27 earnings call. On the call are Rene Haas, Arm's Chief Executive Officer; and Jason Child, Arm's Chief Financial Officer.

Today's call contains forward-looking information about the company and its financial results. While these statements represent our best current judgment, our business is subjected to many risks and uncertainties that could cause actual results to differ materially. Important risk factors that may affect our business and future financial results are described in our annual report on Form 20-F filed with the SEC. Arm assumes no obligation to update any forward-looking statements.

We will also refer to non-GAAP financial measures. Reconciliations of these non-GAAP financial measures to the most directly comparable GAAP measures can be found in our shareholder letter as can the discussion of certain projected non-GAAP financial measures that we're not able to reconcile without unreasonable efforts and supplemental financial information. Our earnings materials are available at investors.arm.com.

And with that, I'll turn the call over to Rene.

Rene Haas

Thank you, Ian, and good afternoon, everyone. Arm delivered a record second quarter. Revenue reached $1.44 billion, up 27% year-over-year, marking our fifth straight quarter of growth above 20%. License and other revenue grew 50% to $773 million, driven by our partners increasing their investment in the Arm compute platform. This strong licensing momentum is an important leading indicator of future royalties, which we expect to accelerate as these new designs enter production.

Second quarter royalty revenue grew 8% to $670 million, with data center royalties doubling year-over-year. And non-GAAP earnings per share increased 33% to $0.52, while we continue to increase our investment in R&D. AI is driving unprecedented demand for compute across the cloud, devices and autonomous systems, with Arm being the compute platform of choice.

Today, our business is now driven by 2 rapidly growing areas: the IP business, where partners are increasing investment in the Arm compute platform and deploying record numbers of Arm-based chips and our new silicon business, where demand for the Arm AGI CPU is extremely strong, both leverage the same Arm compute platform and software ecosystem. We are seeing these 2 growth vectors play out clearly in the data center, where our partners are expanding their deployment of Arm-based infrastructure alongside strong demand for Arm-designed silicon.

Oracle Cloud Infrastructure, OCI, is a key example of how these 2 businesses reinforce one another. Today, OCI offers NVIDIA Vera CPUs built on Arm, and in addition, plans to offer Arm AGI CPU instances in 2027. Since launching the Arm AGI CPU in March, customer demand has continued to strengthen. We have secured purchase orders from multiple customers extending through calendar 2027 and first commercial deliveries remain on track for fiscal Q4 '27. Our confidence in achieving up to $2 billion of revenue in Arm AGI CPU across fiscal 2027 and 2028 has increased over the last 90 days, supported by additional supply and customer mix.

We remain focused on expanding capacity to meet growing customer demand. Power availability makes performance and efficiency even more important. Morgan Stanley projects a 38 gigawatt global data center power shortfall through 2028. Arm is helping customers address this constraint by delivering more useful compute from the power already available. At OCP, independent testing by Signal65 showed the Arm AGI CPU delivering approximately twice the performance per rack versus the AMD EPYC 9755. This type of improvement could reduce the projected global power shortfall by half.

Our data center royalties doubled year-over-year, and Arm-based server market share at the top hyperscalers has more than tripled from 15% in calendar 2024 to more than 50% in calendar 2025. Based upon expected ramps across Google Axion, AWS Graviton, Microsoft Cobalt and NVIDIA platforms, we now expect that share to exceed 60% by the end of calendar 2026. Beyond the data center, Arm's role in edge AI increases as AI agents become more capable and persistent.

Arm-based PCs have now reached 20% of global shipments. Google recently announced Googlebook, a new category of AI-native laptops built for Gemini Intelligence with Arm serving as the primary compute platform for many models. NVIDIA's RTX Spark systems, which are now available, further expand Arm into local AI, content creation and gaming. In smartphones, Arm recently announced CSS for Mobile 2, which improves the performance of agentic AI workloads by up to 70% through SME2.

Our first AI-native Mali GPU delivers up to 4x the frame rate at the same sustained power. These capabilities allow more AI and richer experiences to run directly on the device without compromising battery life. Arm is the platform of choice for physical AI. Royalty revenue grew 50% year-over-year, driven by advanced driver assistance and in-vehicle infotainment. We also launched Arm Total Design for physical AI, bringing together more than 80 companies to accelerate the development and deployment of complete physical AI systems.

For the first time, our partners reported more than 10 billion Arm-based chips shipped in a single quarter, an increase of 1.5 billion year-on-year, supported by a software ecosystem of more than 30 million developers. Demand for Arm technology has never been stronger as AI is driving more complex designs and greater demand for power-efficient compute. Our IP business generates licensing revenue as those designs begin and royalties as they enter production. And now Arm AGI CPU as a new direct silicon business built on the same compute platform and software ecosystem. That gives Arm 2 powerful growth engines, which we expect to accelerate revenue at a faster pace in the future.

And with that, I'll turn it over to Jason.

Jason Child

Thank you, Rene. We delivered another strong quarter with total revenue growing 27% year-on-year to $1.44 billion. This was a record for our second quarter and our second highest quarterly revenue ever, reflecting continued strength in our business. License and other revenue was $773 million, up 50% year-on-year, also a record for the second quarter and our second highest quarter ever. License revenue was around $100 million higher than we had anticipated due to an acceleration of interest from customers wanting to sign long-term commitments to access Arm technology.

During the quarter, we signed 3 new Arm Total Access agreements, bringing the number -- total number of active ATA subscriptions to more than 60. This included one of our major customers signing their first ATA license. We also signed 4 additional compute subsystem licenses, 2 for data center chips and 2 for smartphones. For established customers, these agreements broaden and deepen access to our future technology road map and provide an important foundation for Arm's future royalty growth.

In addition, we are seeing ATA and CSS as a way for newer customers who have previously made only limited use of Arm technology to accelerate their adoption with several signing ATA or CSS licenses during the quarter. Of the $773 million in license revenue, our agreement with SoftBank for technology licensing and design services contributed $209 million, up 17% year-on-year.

We expect the quarterly run rate for the rest of the year to remain around this level. Annualized contract value grew 12% year-on-year. As we have discussed before, license revenue can vary significantly from quarter-to-quarter based on the timing and accounting treatment of large agreements. So ACV remains a useful measure of the underlying licensing trend.

Turning now to royalty. Royalty revenue was $670 million, up 8% year-on-year, also our highest ever figure for Q2. Cloud AI remained the largest growth driver. Once again, data center royalty doubled year-on-year as Arm-based CPUs ramp across AI infrastructure and general-purpose cloud workloads. Physical AI royalty revenue grew 50% year-on-year, driven by increasing adoption of Arm technology in automotive and robotics.

In edge AI, royalty revenue was roughly flat year-on-year. Higher memory prices continue to pressure smartphone unit sales, particularly in the mid- and lower-tier segments. That headwind was offset by the benefit we continue to see from higher royalty rates as our latest Armv9 and CSS technologies are adopted. As smartphone market conditions normalize, we expect edge AI royalty revenue to reaccelerate. As Rene discussed, customer demand for the Arm AGI CPU remains very strong.

Last quarter, we said that we increased confidence -- we had increased confidence in achieving more than $1 billion across fiscal year '27 and '28. In the past 90 days, we have signed purchase orders from multiple customers and signed supply agreements with companies in our supply chain, giving us increased confidence in achieving up to $2 billion across fiscal '27 and '28. We will provide another update at our Q3 results, which is when we'll have better visibility.

Turning to operating expenses and profits. Non-GAAP operating expense was $774 million, up 19% year-on-year as we continue to expand our engineering capacity and invest in next-generation platforms, compute subsystems and our Arm AGI CPU product family. Non-GAAP operating income was $638 million, up 37% year-on-year. Non-GAAP operating margin was 44.2%, up 300 basis points from a year ago as revenue outpaced operating expense growth. Non-GAAP EPS was $0.52, up 33% year-on-year. Importantly, even at these elevated levels of investment, we generate free cash flow of $1.1 billion over the trailing 12 months, giving us the flexibility to continue to invest for long-term growth.

Turning now to guidance. For Q3, we expect revenue of $1.54 billion, plus or minus $50 million. At the midpoint, this represents growth of approximately 24% year-on-year. We expect continued strong growth in license and other revenue, up a little over 50% year-on-year, supported by strong customer demand and our visibility into the licensing pipeline. In royalties, we expect continued strength in cloud AI and physical AI, offset by ongoing pressure in edge AI from weaker smartphone demand, and we have a tough year-on-year comp.

As a result, we expect royalty revenue to grow low teens sequentially and low single digits year-on-year. We expect non-GAAP operating expenses of approximately $825 million and non-GAAP EPS of $0.56, plus or minus $0.04. Looking further ahead, we continue to see strong customer demand across our business. Combined with the expansion of our technology portfolio and deepening strategic customer relationships, this gives us confidence in our long-term growth outlook.

With that, I'll turn the call back to the operator for Q&A.

Operator

[Operator Instructions] And your first question today comes from the line of Gary Mobley from StoneX Group.

질의응답

Gary Mobley

I think on the top of the mind for everybody is Arm AGI CPU, and you're obviously signaling something closer to $2 billion in revenue over the 5-quarter period up to the end of fiscal year '28. So I'm wondering what's driving your confidence in that with respect to securing more supply with your different partners? And I would assume most of the revenue initially goes to supporting Meta, but maybe if you can talk about whether or not you've signed and committed contractually with additional customers?

Rene Haas

Yes. Thanks, Gary. I'll take that question, and Jason can add if I've missed anything. Each one of the quarters, as we talk about the ramp of the product, we've been careful to make sure that we manage the expectations of supply and demand. I think there's no secret that in general, demand is incredibly strong for the Arm AGI CPU, really driven by everything we see with agentic workloads and the CPU being central to that.

Demand is stronger today than we talked about when we did the Arm Everywhere event in March. And subsequently, demand for multiple customers, Meta being the chief customer has also been beyond what we've expected. That being said, we've been working very hard over the last few quarters on securing what's necessary in the supply chain to address that demand. There's multiple components to it. There is wafer supply, there are substrates, there are testers, there's memory, et cetera. We have wanted to be confident when we signal our ability to achieve this extra demand with supply in the ongoing quarters.

So at Arm Everywhere Day, we talked about visibility of $1 billion. And quickly, we have seen that to about a $2 billion number. I can tell you today that the confidence is the strongest it's been to date, and it's definitely stronger than it was 90 days ago in terms of securing that supply. And as I said, it's across all of those independent areas. But we didn't start working on it yesterday or the previous quarter. We've been working on it actually for the last number of quarters. We've got great partners who are supporting us here. And as you mentioned, Meta is a chief customer, but there will be multiple customers beyond Meta.

Jason, I don't know if there's anything you want to add. So if not, thanks, Gary, for the question.

Gary Mobley

If I could ask...

Operator

Please continue, Gary.

Gary Mobley

My apologies. I was just hoping to ask a follow-up question briefly. You've given us enough information in the shareholder letter to calculate your royalty rate per unit on a year-over-year basis, and it looks like it's trending up at least mid-single-digit percent. So I'm wondering if that's just sort of an anomaly with respect to mix or if that should be how we view the royalty rate per unit growth over the long term?

Jason Child

Yes. It's -- I mean, it's mix, but it's also, I think we said, what, 10 billion chips shipped in the quarter. And so it's going to bounce around based on any particular quarter. I would say, certainly, mobile on average is probably a little lower than certainly server and physical AI given the size and value of the chips. And so that probably had some impact. So I don't know that I would take the overall average across 10 billion chips and use that as the best way to forecast going forward.

I do think if you break it out by the 3 different business units, as we said, pretty flattish growth in the mobile business, very strong double-digit growth in server, where we're also seeing increasing core counts and more CSS revenue flowing through this quarter, but we're going to continue seeing that going forward. That's why we have high confidence that we're going to have triple-digit growth for the next few years.

And then on physical AI, that one is also -- we said growing 50% year-on-year. That's also going to, I'd say, be a bit of a tailwind to overall royalty rates and overall royalty growth. But again, I would probably break it into those 3 categories as opposed to just looking at it all combined together.

Rene Haas

Yes. Maybe, Gary, just a finer point on the data center growth. I think there's 2 vectors independent of the Arm AGI CPU on the royalty. One is Arm taking market share, which I mentioned before. Now we think it's going to be 60% by the end of the year. That market share gain is driven both in terms of units going up at agentic AI driving increased design starts for Graviton, Cobalt, Axion, et cetera.

But in addition, we are seeing a transition away from x86 to Arm for the head node CPU that connect to accelerators. We already know about Vera Rubin, but Google has already announced the conversion of x86 to Axion connecting to TPU version 8. So to Jason's point, this is why we have a high degree of confidence on data center royalties growing triple digits over the next number of years.

Operator

[Operator Instructions] And our next question comes from the line of Sebastien Naji from William Blair.

Sebastien Cyrus Naji

I want to ask about how Arm wins versus x86 in the AI data center. At OCP, you introduced agentic AI benchmarks that showed your AGI CPU can deliver twice the performance per rack versus a competing x86 solution. But then last night on its earnings call, AMD said that its Venice CPU offers the best performance per watt. So maybe, Rene, could you talk about where you see the most meaningful opportunity for the Arm CPU to win in the data center versus x86?

Rene Haas

Yes. Sure. I'll take the question kind of from 2 angles. One is looking at it in the eyes of a Amazon or Google or Microsoft who are building custom CPUs for their own implementations and their own racks and their own systems. In those designs, particularly when an end customer is designing an SoC, the degree and flexibility that they have to innovate around their own system is extremely high. Trade-offs that they can make in terms of what goes inside the SoC for data movement versus what takes place inside of DPU, which is what takes place inside of storage.

At that level, you almost have to look at it in terms of performance per rack versus the chip basis. But at a chip level, Arm not only gives them the most efficient core on the planet, but allows them to design the most efficient SoC and subsequently, the ultimate system. So -- and we've seen that just in terms of how quickly these customers have been converting designs away from standard x86 to Arm. Now the Arm AGI CPU benchmark that we ran, that was done through Signal65 at OCP, was looking at the previous generation to Venice. And we are very confident in those numbers, and we're very, very confident in terms of the choices that customers make vis-a-vis head-to-head against that design.

Venice is a great product, but we're very confident in terms of the demand we have for our product, Arm AGI CPU, and the road map going forward. But I would say 2 vectors to look at it. One is just efficiency, cost per -- performance per dollar and performance per watt with AGI CPU. And then on the custom level, there are just simply things that end customers can do that a standard product is never going to be able to touch. And that's kind of the beauty of our business model.

Operator

Our next question today comes from the line of Krish Sankar from TD Cowen.

Sreekrishnan Sankarnarayanan

Rene or Jason, at OCP, you introduced your AGI CPU reference server design. I'm just wondering, are you working with ODMs or server manufacturers? Are you targeting the enterprise private cloud? And along the same path on FY '31, $15 billion in AGI CPU revenues, which is the biggest factor? Is it core count increase? Or is it ASP increasing from the current version 1?

Rene Haas

Yes. No, thank you for the question. When I think about the partners that we're working with, it's Supermicro, Lenovo and ASRock. They are building the racks for us, and then are basically standing those up for deployment at data centers such as OCI. So they are the partners that we build with. As far as the revenue outlook, and Jason can speak to the specifics of where we think and when we think those numbers are going to land. There is clearly going to be an acceleration of number of cores inside these SoCs.

And that is largely driven by agentic workloads and specifically the fact that each one of these agents, when you think of Muse agent that's running a specific job, typically runs as a virtual machine on one CPU core. The software is rather complex to manage that across a single core or multiple cores, if you will. So if you can run that on a single core, it's a better thing. So do I see CPUs going from 128 cores to 256 cores to 500-plus cores? Absolutely. You can imagine that the average selling price of the chips with more cores will be higher, certainly. But I think what we will see is probably a more rapid acceleration to that, which will help grow the top line.

And Jason, anything you want to add on to that numbers-wise?

Jason Child

Yes. No, I mean, we said back at the Arm Everywhere event last March that we had the line of sight to $1 billion for this kind of end of '27 and fiscal '28. And of course, now we're increasing our confidence to deliver $2 billion. And then the following year, we haven't made any updates, but I think the original plan was that we were going to get to about $2 billion a year after that, and then we're going to go up to about $5 billion a year after that. And so to the extent that we're successful in delivering over $2 billion, you should expect that we'll probably have some upsides after that.

That's something that we'll plan on providing in terms of updated possible targets when we have the next Arm Everywhere event. We're still working on timing, but likely probably in, say, spring or first half of next year is probably when we'll provide that update.

Rene Haas

Yes. I think what we can say is when we did the Arm Everywhere Day in March, the conventional wisdom was the market was a $60 billion market and wasn't going to be growing all that much. When we did our calculations, we felt differently, and we talked about a number that day of $100 billion by 2030. I think that number has been increased by various folks. I think Lisa may have even said $200 billion or more. I think that's about right and maybe even could be undercalling it, who knows?

What is obvious is that the agentic demands on AI jobs are CPU bound. And the more agentic workloads that run, that's going to need more CPUs. So I think that's what's going to end up driving these higher core count devices, which I think is very -- will be very good for us because what's going to matter there is performance per watt at the core level, having the most efficient core on the planet, which is what we're very good at.

Operator

Your next question today comes from the line of Mehdi Hosseini from Susquehanna.

Mehdi Hosseini

Yes. Two follow-ups. Going back to the new updated AGI silicon revenue target of $2 billion, should we assume any revenue contribution in September or December quarter of this year? FY '27?

Jason Child

Assume that our shipments will start in Q4. If we have something in Q3, we'll let you know. It wouldn't be a lot. We're really targeting Q4.

Mehdi Hosseini

Okay. And then just a quick follow-up. Would it be fair to say that data center mix of the royalty revenue is tracking to 1/3 or 30% to 35%? And for next fiscal year, FY '28, it could go up towards mid-40%?

Rene Haas

I'm sorry, the -- are you asking...

Jason Child

Mix of total royalties?

Rene Haas

Percentage of total royalties that are data center. Is that the question?

Mehdi Hosseini

Yes. Would it be fair to assume that it's tracking to like 1/3 in FY '27 and it would step up again in '28, moving towards 45%?

Jason Child

Yes. I don't -- yes, I mean, we look at them independently. So I would assume that royalties for edge are going to be pretty flattish this quarter. And I think we said we'll start to see some recovery in Q4. But if you have that in the kind of high single digits and then you have data center growing at double digits. I think we said last year, we were kind of approaching 20-ish percent of total royalties.

And then you have physical AI growing somewhere probably in the -- somewhere between 20%, 30-ish percent range, then that's how you should build your model. I haven't done the math to look at the percentage. I don't think data center will get quite to 1/3, but it will sometime soon. I just don't know if it's going to get there by the end of this year.

Operator

Your next question comes from the line of Simon Leopold from Raymond James.

Simon Leopold

I wanted to ask about the -- basically the cloud and AI group of products and how you might frame the customer concentration there to help us give us a better appreciation? For example, could we guess that maybe 60% of these royalties are coming from 3 customers? Any kind of quantification would be helpful to understand that.

Rene Haas

I think we can say, and Jason, chime in on this thing, Amazon, Google and Microsoft are the largest contributors, and then you've now got NVIDIA ramping up quickly. They are the 4 guys inside the data center that command the majority of the royalties.

Operator

Our next question today comes from the line of Harlan Sur from JPMorgan.

Harlan Sur

With the strong success of your first-generation AGI CPU, I'm sure your customers are already looking towards your next-generation platform, right? Can you just give us an update on second-gen AGI? I think the plan was for the team to tape out either end of this year, beginning of next year for ramp in calendar '28. And then maybe just as importantly, as you continue to add more capabilities to your compute silicon portfolio, does it actually make sense to add more networking and high-performance I/O capabilities as well?

More and more, we're seeing that compute and networking are very closely coupled to determine very good systems performance. So does the team -- is the team going to add more networking and capabilities to the portfolio?

Rene Haas

Yes. Thank you, Harlan. Those are great questions. Without disclosing the precise schedule on the follow-on chip, you can think about this market goes at about an 18- to 24-month cadence in terms of production ramps. So as we are ramping into production now on the first one, you can kind of do the math and when the second one would be available. Regarding other things that you can add to the toolbox to improve performance, whether that's around connectivity, broadly speaking, whether that's around specific I/O and/or networking and/or other aspects of the system, we are absolutely looking at all of that.

When we think about developing these products, we're thinking about rack system performance. The chip, the CPU is a big part of it, but how you move data in and out of the system, to your question, is as large an area as any. And we have a lot of activity underway in all these areas. Nothing that we can announce here on this call, but completely cognizant of that being an area that we could have even larger differentiation.

Operator

Your next question comes from the line of Lee Simpson from Morgan Stanley.

Lee Simpson

Maybe just 2 quick ones, if I could. Maybe first one for Jason. $773 million for licensing really stands out. Congratulations to everyone there. You did talk about Total Access. You did talk about CSS deals, particularly smartphones and I think data center, you mentioned. But do we have any sort of architecture license agreements in there? And could we look at 770 plus as a new level, a structural level?

And maybe for follow-on for Rene, maybe in line with some of the earlier questions, it's interesting that as we look to higher ASP CPUs in the future with higher core count, could it be that we see a higher per core dollar equivalence in that pricing? In other words, playing off this capability that are associated capabilities, perhaps networking and I/O as mentioned in the last question, that drives that per core dollar value up?

Jason Child

Okay. Thanks, Lee. On your first question around license, yes, we didn't -- we talk about ATA and CSS, which is the vast majority of everything that we're selling going forward. Architecture licenses are very, very few and far between. And so I would think of it as really being ATA and CSS. That is the bulk of most everything we're selling going forward. In terms of absolute magnitude of dollars, yes, $773 million is -- was a great quarter. But if you -- I think kind of work out the details, I said it was going to be increasing and going up from there in this next quarter. And even in Q4, it's going to step up again.

So we're not -- we haven't talked about next year, whether or not that's the new the new number every quarter. I don't know if I would say that. But I will say the pipeline has been surprisingly strong, stronger than we expected when we started the year. And a little bit like the last couple of years where we've had maybe deals that weren't initially -- they were in the pipeline, but they weren't -- they were not renewals. They were -- these were incremental and new deals that are kind of hard to forecast. We're in that same situation again.

And so what's interesting is when we started this year, we thought it was going to be much more around royalties and that a lot of the license growth related to all the AI investments, we thought maybe we kind of had passed a lot of that. Well, what's interesting is it seems to be that with AI and the consistent variability and now everyone's trying to address all the agentic workloads we're seeing that the need for additional license and additional product variations and different technologies is not slowing down at all.

And so that's particularly exciting. As you know, any license beats, especially with ATAs and CSSs, that will -- is very highly likely to turn into design wins and very likely to turn into even more royalties probably 12, 18 months, 24 months down the road. So that's a really good setup for next year and after. And so -- but we'll have to, I guess, get through the next couple of quarters before I can talk about next year. But the pipelines do look strong.

And even with the memory issues that we've seen really kind of making for some tough comps in the mobile business, which really that does end mostly in this quarter that we're now in Q3. We start lapping those headwinds that really started in January, February of next year. So the comps also get easier. But overall, we think it's a really good setup for the balance of this year and likely going into next year as well.

Rene Haas

Yes. On pricing, it's a tricky one from the perspective of memory pricing has obviously increased a lot since we announced the Arm AGI CPU. Data centers are getting more and more complicated to build, whether it's through just, again, memory pricing/finding laborers, et cetera, et cetera. So while we want to be cognizant of not pushing the pricing piece too hard given all these other factors, I think it is fair to say that there's probably some pricing elasticity that we could potentially exercise on these larger chips and the chips that have higher core counts. We'll look at all of it, obviously. But I think it's fair to say that there probably is some headroom there.

Operator

Your next question today comes from the line of Vivek Arya from Bank of America Securities.

Vivek Arya

One for each. Jason, let's say, smartphone units are down again double digit next year. Do you think royalties will kind of be similarly profiled to what we saw this year? Or are there other dynamics at play? And then, Rene, on the AGI platform, let's assume we believe in the technical merits. My question is, does Arm have the scale in the ecosystem to secure supply? Because, for example, yesterday, when AMD reported they raised their server CPU business, right? By several billions of dollars.

So that tells me that they have the scale to go and get incremental supply, right? So how are you dealing with that question about does Arm have enough scale to go and get the supply to be a meaningful player in the CPU markets, right? Given your belief in the technical merits of the product?

Rene Haas

Yes. Thank you, Vivek. I'll start, take your second part of your question first, and then Jason will talk about the smartphone scenario that you had described. Your question is a valid one. The way I think about it is that the Arm AGI CPU is our first product. AMD has been at it for a long time. They've got a lot of SKUs. They buy a lot of capacity from TSMC. They have a lot of relationships that they've built over decades. This is our first product. That said, it's not our first relationship with folks like TSMC.

It's not our first relationship with memory vendors such as Micron, Samsung and SK. They've known us for years and decades, and we've been working pretty diligently to share with them our long-term plans. I think supply for the industry is going to be tight. And as Jason mentioned, we'll talk a bit more about our future outlook in upcoming quarters. But we are going to be working very, very hard to secure the supply that we need for the expanded demand because we think it's very, very, very large.

And our partners know that. And we've been talking to them not for weeks, not for months, but probably a year plus as we've talked about doing this platform. So we've looked at this as a long-term endeavor. It's not a 1-year single project. It's a 5- to 10-year endeavor that's going to require relationships across all the supply chain, and we're very, very serious about making sure we get all that done.

Jason Child

Thanks, Vivek. So regarding the mobile business, when we look at some of the external forecasts, I think IDC has units in mobile growing basically flat. I think it's minus 1% next year. As we talk to partners, that's -- I'd say, flat is generally what folks are expecting. We -- there's already been kind of the hit and the reduction due to the memory constraints. And I think the expectations are those constraints stay relatively consistent, which basically means flat growth, not further deterioration. And then because we have -- we continue to have more v9 adoption, it's -- we're now getting over 50%. So it's a little less of a tailwind.

But CSS adoption continues to grow. We also have multiple partners on next generations of CSS, which are at slightly higher royalty rates with -- and higher chip prices. So all that means that we will grow the overall business much faster than the market. So think of it at least being in the kind of double digits plus in growth rate for next year. That's our best guess right now. If we see anything that changes, of course, we'll let you guys know. But that's what industry as well as our partners all seem to think at this point.

Operator

We will now take our final question for today. And the final question comes from the line of Joe Quatrochi from Wells Fargo.

Joseph Quatrochi

You talked about securing more supply on AGI, and I assume that comes at a higher cost. So can you just kind of talk us through the ability to offset that and how we should think about just the gross margin puts and takes as you start to ship?

Rene Haas

Yes, I'm sorry. Could you repeat that one more time?

Joseph Quatrochi

Yes. Just -- I assume that the increased supply that you're getting for AGI is coming at a higher cost. So can you just talk about your ability to offset those higher costs, pass that through to customers in terms of pricing and then the puts and takes on gross margin?

Rene Haas

Yes. No, thank you. As I mentioned earlier, the overall cost of the systems, memory being probably the single largest contributor to all that. And then you've got kind of downstream other impacts to your point, on different parts of the supply chain, definitely put margin pressure. That being said, we do have some elasticity on pricing and the ability to increase it, which I think we've already demonstrated an ability to do that as we've gotten started with the program.

So I'm actually not as worried on that point relative to offsetting any constraints on supply and cost and managing it downstream. You've seen public comments from everybody else who's in this business that that's what we're doing. We'll be following where the market goes and making sure that we're not trailing common sense, but at the same time, not trying to cause too much pain on the end customers. It's our first product. We want to make sure everybody wins with it, but I'm not as worried about that piece impacting gross margins.

Jason Child

Yes. Maybe just to reiterate, what we said was for this first version, we're expecting to be somewhere in the high 30s to low 40% gross margin. And for the first version, we expect it to be probably a little lower. And then as we develop more expertise, probably rely on doing a little more work ourselves and maybe less from partners over time, which others in the market do over time, we'll get ourselves to more of the market standard of 50% plus.

But I don't expect that to happen on the first version. That's probably going to happen more on the next version. So for this version, I would stick to the kind of 30% to 40% range. And then we'll obviously give you updates on a quarterly basis as we start shipping in Q4.

Operator

That was our final question for today. I will now hand the call back to Rene for closing remarks.

Rene Haas

Thank you, and thank you for all the questions. They were very, very timely and excellent questions. Just to summarize, this was a great quarter for the company. 50% licensing growth means a lot of growth coming in the future in royalties, which we feel great about. We also feel very, very good about the fact that our data center business continues to flourish. We had talked for many, many years about Arm getting to 20% in data centers. The fact that we are 50% in the last year and will be 60% by the end of this year is also something we're very, very proud of.

But most importantly, I think what we're now starting to see is what was one great business is now turning into 2 great businesses. The IP business continues to do very, very well. We're very confident in this growth going forward. But now with our new AGI CPU ramping into production, increased confidence to achieve the $2 billion number that we've been talking about, it just thrills me to be able to talk about 2 businesses now rather than 1 that are both going to be on amazing growth paths and could not be prouder of what the company has achieved, thanking our employees, our partners and customers who helped make it happen. So thank you again for all the questions and comments, and congratulations to the Arm team for a great quarter.

Operator

Thank you. This concludes today's conference call. Thank you for participating. You may now disconnect.