Q1 2025 Earnings Call Transcript

Ladies and gentlemen, welcome to the Synopsys Earnings Conference Call for the First Quarter Fiscal Year 2025. [Operator Instructions]. As a reminder, today's call is being recorded. At this time, I would like to turn the conference over to Trey Campbell, Senior Vice President, Investor Relations. Please go ahead.

Good afternoon, everyone. With us today are Sassine Ghazi, President and CEO of Synopsys; and Shelagh Glaser, CFO. Before we begin, I'd like to remind everyone that during the course of this conference call, Synopsys will discuss forecasts, targets and other forward-looking statements regarding the company and its financial results. While these statements represent our best current judgment about future results and performance as of today, our actual results are subject to many risks and uncertainties that could cause actual results to differ materially from what we expect.

In addition to any risks that we highlight during this call, important factors that may affect our future results are described in our most recent SEC reports and today's earnings press release. In addition, we will refer to certain non-GAAP financial measures during the discussion. Reconciliations to their most directly comparable GAAP financial measures and supplemental financial information can be found in the earnings press release financial supplement and 8-K that we released earlier today. All of these items plus the recent investor presentation are available on our website at www.synopsys.com.

In addition, the prepared remarks will be posted on our website at the conclusion of the call. With that, I'll turn the call over to Sassine.

Good afternoon. We had a solid start to 2025, exceeding the midpoint of our Q1 revenue guidance and delivering non-GAAP EPS above our guidance range. As we outlined in December, Q1 revenue was down 4% year-over-year, and non-GAAP EPS was down 10% as we had one less work week in Q1 '25 versus Q1 '24. . Let me take a few minutes to share some business highlights, and then Sheila will discuss the financials in more detail.

From an end market perspective, AI and HPC remained robust in the first quarter, while industrial, automotive and consumer electronics remained challenged. Despite the sale of two markets, along with headwinds in China as we anticipated, Synopsys opportunity is tied to R&D and underpinned by the mega trends of AI and silicon proliferation and software-defined systems. These trends are increasing design complexity and cost while driving greater compute and energy demands. New design paradigms are essential to address these challenges and Synopsys is racing to deliver.

I had the privilege to meet with semi and automotive customers at in January, who all expressed their strong belief in the strategy we are driving. They underscored the pressing need for solutions to design, validate and optimize intelligent products virtually from silicon to systems. Our pending acquisition of ANSYS will pave the way for new AI-powered design solutions that use electronics and physics giving R&D teams the tools they need to ignite their future innovation. In January, the European Commission approved our pending procompetitive acquisition of ANSYS and the U.K.

CMA provisionally accepted our remedies toward the Phase 1 approval. As previously communicated, the U.S. HSR Act waiting period has expired. And we're making strong progress with other regulatory agencies, including China.

Customers overwhelmingly support this transaction and we continue to anticipate closing in the first half of 2025. Moving to business highlights. In Q1, Design Automation revenue was up 4% year-over-year with 1 less week of revenue versus the prior Q1, while design activity remained strong. Synopsys is the leader in hardware-assisted verification or HAV solutions. And this month, we strengthened our position expanding our industry-leading HAV portfolio to include new HAPS 200 prototyping systems and new ZeBu 200 emulation systems with up to 2x better performance versus our prior generation.

AMD, ARM, NVIDIA and SI5 are among a number of customers who are deploying our new prototyping and emulation technologies, and we were honored by their participation in our recent launch. Last year, we had our best year ever in hardware, and we expect another year of strong performance based on the enthusiasm for our newly expanded Synopsys HAV portfolio. which provides the unmatched performance and flexibility our customers require to prototype emulate and verify ever more integrated complex and software-defined systems. Turning to EDA software, where we are seeing strong design activity at advanced nodes with 2-nanometer projects accelerating rapidly.

Fusion Compiler is the industry-leading platform for advanced node digital design implementation. And this quarter, we saw a U.S. hyperscaler tape out a 2-nanometer test chip exclusively using Synopsys design flow. Additionally, at 2-nanometer, Fusion Compiler was the platform of choice for a U.S. HPC, CPU tape-out and an Asian mobile customers 2-nanometer at SOC.

Moving to Synopsys technologies where we offer the industries essential trusted solutions to close out timing, signal integrity, power and variation aware analysis. Headlining our portfolio is prime time. which is used by virtually all key advanced node customers. Customers are reporting significant productivity improvements with the most recent prime time release with one customer achieving 30% faster turnaround time with multi-core scaling. Our IC Validator product family is delivering tremendous value and physical verification sign-off.

And recent product improvements have unleashed even greater turnaround time improvements for customers. Leading edge customers are achieving greater than 2x turnaround time for full chip physical verification sign-off at 3-nanometer and below, enabling design teams to finish more sign-off runs within the budget cycle time to improve the quality of results. Before moving out of sign-off, a few points on Star RC, which is the industry-leading tool for extraction on advanced process nodes. To date, we've seen all of our major CPU and GPU customers on TSMC and [indiscernible] and Intel 18A using STAR RC for sign-off extraction with the key differentiation being the accuracy of results and tool performance relative to competition.

The massive AI infrastructure build-out that's currently underway paves the way for a transformation across all industries, including our own. AI is fueling chip innovation and the AI-driven EDA capabilities we pioneered from reinforcement learning to generative AI capabilities are delivering significant productivity gains and cementing our leadership position, but we're only at the beginning. While customers are realizing compelling value from our initial AI-driven optimization engines, these enhanced capabilities have not yet dramatically altered the underlying design flow for a chip. We see a paradigm shift coming with agent AI where engineers can task autonomous agents with executing complex workflows.

We believe this will be massive value and productivity unlock for our industry. Which we'll talk more about at our Synopsys User Group Conference in March. But first, AI business highlights from this quarter. In Q1, we continue to drive Synopis.ai adoption across our tools in design implementation, verification, test and analog.

In verification or VSO.ai, we saw a large U.S. memory company begin deployment of VSO.ai to find corner case bugs realize a 2x improvement in hardware utilization, while an Asian hyperscale customer achieved a 4x turnaround time improvement with VSO.ai on its HPC design, significantly improving hardware utilization and outperforming the competition. Our analog migration tool, ASO.ai continues to build a strong pipeline of customer opportunities and in Q1 delivered a significant competitive displacement at a leading aerospace company. We also continue to expand our generative AI offerings for customers.

We recently added script generation capabilities to the core pilots for Fusion Compiler and prime time, and early customer results are demonstrating 30% average productivity improvements for designers. Additionally, synopsys.ai generative formal verification capability in Verde is delivering up to a 35% productivity boost in early engagements with key partners. On to design IP, in line with our expectations, revenue was down 17% year-over-year versus a record setting prior year compare. While IP revenue can fluctuate quarter-to-quarter, the opportunity set for IP continues to expand, particularly as AI customers accelerate protocol transitions and look for creative ways to drive enhanced performance per watt.

This quarter, we launched the industry's first Ultra Accelerator Link, or UAL and Ultra Ethernet IP solutions to connect massive AI accelerator clusters addressing the industry need for open standard solutions to scale AI accelerator infrastructure. We also continue to optimize our foundation IP libraries to deliver unparalleled AI performance. One, high-performance AI customer used our memory and logic libraries to deliver breakthrough LLM performance at 5-nanometer. Across our interface IP portfolio, AI continues to push protocols forward at breakneck pace as customers drive for additional performance per watt.

This quarter, we captured several key design wins, including a Kind Edge PCIe 7.0 design with an AI infrastructure chip provider, and we secured a 224 gig ethernet win with a major ecosystem player. We also secured 112-gig SerDes and PCIe 6.0 agreement with a leading European telecommunications equipment provider and an interface IP development deal for a leading auto OEMs advanced 2-nanometer design. Our IP development for the foundry ecosystem is a mission-critical ingredient for the industry. And in Q1, we announced silicon success for PCIe 4.05 IP on Samsung's SS-8 process used in auto, mobile, networking and storage applications.

Also in Q1, we demonstrated silicon success for our onetime programmable nonvolatile memory IP. This technology enables secure storage for encryption keys, product configuration and SRAM repair information and is now available in TSMC N4 and N5 and N6 and N7 processes. Moving to mobile and consumer markets, where end market demand is challenging, but design activity continues as customers ready a next wave of innovative products. A leading Asian automotive supplier adopted Synopsys interface processor and foundation IP due to our long track record of delivering high-quality IP.

Also in Q1, we closed the design win, including PCIe 4.0, MIPI and USB with a leading mobile provider for an ARM-based application processor. UFS or universal flash storage is a key technology in these verticals. And we closed a UFS design with a key company driving a AIPCs this quarter. We also continue to see strong demand for the advanced UFS protocol in mobile to support LLM storage for Gen AI use cases.

A few closing comments before we transition to Shelagh's remarks, we have a very resilient business model, and our solutions are mission-critical to our customers' innovation. We have strong momentum across the business, bolstered by secular growth tailwinds, including AI. The application of AI for EDA and engineering more broadly is just beginning, which we'll discuss in more detail at Snug in March. Finally, thank you to our employees, customers and partners for a strong start to 2025. We are excited to continue our partnership journey with you through the year.

With that, I'll turn it over to Shelagh.

Thank you, Sassine. We delivered a solid start to the year, with revenue in the upper end of our guided range, non-GAAP operating margin of 36.5% and and non-GAAP earnings above the high end of our guidance range. Our Q1 results are driven by our strong execution across the business leading technology that is mission-critical to our customers and a resilient and stable business model. As a result, we are reaffirming our full year 2025 targets for revenue non-GAAP operating margin and non-GAAP EPS.

I'll now review our first quarter results. All comparisons are year-over-year unless otherwise stated. We generated total revenue of $1.46 billion. Total GAAP costs and expenses were $1.2 billion, and total non-GAAP cost and expenses were $924 million resulting in non-GAAP operating margin of 36.5%.

GAAP earnings per share were $1.89, and non-GAAP earnings per share were $3.03.

Now to guidance. For fiscal year 2025, the full year targets are revenue of $6.745 billion to $6.05 billion, total GAAP costs and expenses between $4.97 billion and $5.03 billion, total non-GAAP costs and expenses between $4.05 billion and $4.09 billion, resulting in a non-GAAP operating margin of 40% at the midpoint. Non-tax rate of 16%; GAAP earnings of $10.09 to $10.31 per share. Non-GAAP earnings of $14.88 to $14.96 per share.

Cash flow from operations of approximately $1.8 billion and free cash flow of approximately $1.6 billion. Now to targets for the second quarter. Revenue between $1.585 billion and $1.615 billion total GAAP costs and expenses between $1.19 billion and $1.21 billion; total non-GAAP costs and expenses between $985 million and $995 million. GAAP earnings of $2.21 to $2.33 per share and non-GAAP earnings of $3.37 and to $3.42 per share.

Our press release and financial supplement include additional targets and GAAP to non-GAAP reconciliations. In conclusion, we delivered a solid start to the year. We continue to execute and for 2025 are reiterating 10.1% to 11.1% revenue growth. Non-GAAP operating margin of 3% and approximately 13% non-GAAP EPS growth.

Our confidence reflects our relentless execution and leadership position across our segments, mission-critical products to enable our customers' innovation and a stable and resilient business model. With that, I'll turn it over to the operator for questions.

[Operator Instructions]. Your first question comes from Sitikantha Panigrahi with Mizuho.

Congrats on a nice quarter, and Shelagh that 200 bps margin date is impressive. Sassine, I want to ask about growth trends that you expect to unfold over this short to long term. And when you segment your growth into AI and non-AI side, you talked about the strong AI-driven design activity. So how do you see DeepSeek going to impact EDA in general and Synopsys in particular?

And then I have a follow-up.

Yes. Thank you for the question. We started talking about the tale of two markets maybe about a year ago. And the reason for that is if you look at the semiconductor market, in particular, you have the grouping of customers that they're developing semiconductor chips for AI, HPC and they've had a very strong demand and and a very healthy road map that we are engaged with and supporting and selling to.

Then you have the other grouping of customers, which is primarily the consumer electronics, automotive, industrial, that their opportunity to leverage AI has been shifting to the right. Now you look at the consumer electronics, in particular, the PC and mobile, it has picked up recently due to applications like AI on PC, AI on phone. And this is where DeepSeek will provide an opportunity to expand the adoption of AI on devices, given the affordability, the effectiveness, you don't have to go back to cloud in order to retrieve information, et cetera, et cetera. Now you look at the whole semiconductor R&D, it's expected to grow from a 6% of sales per year to about 9%.

For us, that's fantastic. Because remember, we sell to R&D inside our customers. Now you zoom out completely to the Synopsys opportunity. we don't only sell to semiconductor, we sell to system companies, which are developing chips not to be sold, but developing chips for them to consume. And that opportunity as well has been great for Synopsys given they are consuming IT hardware, software to design some of the more sophisticated chips.

So that's how we see the landscape. I hope that gave you a sense of how we see the monetization.

Okay. and do you see that now the non -- in the traditional semi non-AI side, do you think that we already hit the trough and -- do you expect -- what kind of trends you are seeing on those outside of making AI chips?

On the consumer electronics side, PC, mobile, yes, we have seen over the last 2 to 3 quarters, an acceleration in our customers' road map in building these chips. In automotive, industrial, I want to say it's still about the same.

The next question comes from Lee Simpson of Morgan Stanley.

Great. Great quarter, guys. Maybe just a first question just on China. It does seem like Chinese growth is flattening off, I think, 12% of sales in this last quarter.

And I assume here in, given the discussion that you made around hardware, that hardwood at least is shipping quite nicely to China. But probably not much else. So I wonder if you can maybe give us a sense for the moving parts. If we leave aside export controls and the changes there, what is driving that flattening of sales in China?

And how should we think of that as we go through the year? And maybe as a follow-on, Sassine, I couldn't help but you did tease us with the Agentic AI and the productivity gains. And so I'm just trying to understand the viewpoint that Synopsys would have here utilizing, I assume deep research and like, are you looking at this as an operating margin driver? Or are you looking at this as a product development acceleration?

Regarding China, if you recall, when we reported Q4 '23 and guided FY '24, we said that we're going to be pragmatic and regarding China. And we saw absolutely a trend of deceleration driven by two factors. The first one is the cumulative effect of restrictions. And the second one that cannot be ignored, which is the slowing local economy and the money that's flowing into start-ups and the overall economy in China.

As we wrapped up FY '24, China finished roughly at corporate average. As we started looking at FY '25 and all these, I want to call them stress points in China, we have assumed in our guidance that China will continue on decelerating and below the corporate average due to the two factors I just mentioned. Now you pointed out hardware in particular, and that really has nothing to do with one part of the portfolio or the other. Because customers, when they're buying from us, they're still buying the software, the IP and the hardware but it has all this impact of road map shifting and customers we cannot sell to due to the restrictions of either technology or entity list.

So that's China. In terms of genic AI, it's such an exciting opportunity. And the reason is so exciting is truly the evolution of AI to get to a point of maturity that will truly change the workflow for -- engineering workflow for our customers. And we're well on our way to partner with key leading AI partners to bring in that transformation from generative AI to genetic AI to the engineering world.

In terms of us inside our company, absolutely. We are taking advantage of every opportunity for every function, not only the engineering functions, the various functions inside the company to modernize and capture the productivity of AI.

The next question comes from Joe Quatrochi with Wells Fargo.

I was wondering if I could ask again on just kind of the design activity for non-AI versus AI. For the non-AI customers, have you seen like the rate of change over the past few quarters in terms of design activity. Has that stabilized and improved maybe outside of the PC smartphone AI dynamic?

Thanks, Joe, for the question. For the way Synopsys engage with semiconductor companies, we engage on pretty much every single chip they have on their road map. So our observation and insights is based on the following: when a customer is designing a chip, say, within a window of about 14 to 16 months tape-out meaning before they hand it over to the foundry to manufacture, are we seeing a shift to the right of that road map. And that's what we classified as the tale of two markets where if you are in AI HPC, get going from 16, 18 months development down to 12 months.

So it's a significant acceleration. For the consumer electronic market as well as auto industrial, for a period of time, the road map was not being fueled by the opportunity to accelerate. So that had -- we observed it through our IP pull down and the pace in which they move to the next project on their road map. We are seeing absolutely the pickup in the mobile and PC.

And when it comes to auto and industrial, I want to say there is not much change compared to what we've seen over the last 3-ish 4 quarters or so. But it does not mean they are not investing in their road map and they're not building chips, but the pace in which they're developing and releasing, it's not with the same pace and acceleration.

That's helpful. I appreciate the detail. As a follow-up, you talked about the new hardware solutions that you launched this month. How should we think about just the growth trajectory there, the pipeline. And then in the context of I think your inventory increased a decent amount this quarter. I think it's at a record level.

Yes. I'll let me make a comment in general on the market, then I would like to turn it to Sheila. In terms of the new hardware system, we're so excited about launch because we focused on the use case of a hybrid emulation prototyping use case where customers are looking for every bit of performance they can achieve in order to validate their software before the chip is available. And this has been the use case that Synopsys has led for many years.

And with this new launch, we continue on expanding that leadership. Shelagh?

Yes. And we've been investing in hardware. As you know, we're up about 15% quarter-on-quarter, about 9% year-on-year. That's not all finished goods.

So obviously, as Sassine mentioned, we just launched a few weeks ago. the exciting new platforms. We're racing to build those products, finish those products out demand exceeds supply right now. We see a much more back half loaded. In fact, Q4 loaded hardware year just because of availability of those new units.

So we're racing as quickly as we can to fulfill that customer demand, but that's why we're putting so much more investment in hardware to be able to service those important workloads for our customers.

The next question comes from Charles Shi of Needham & Company.

Maybe the first one, a housekeeping item. Shelagh Glaser, what's the backlog exiting fiscal Q1 '25?

The backlog exiting Q1 is $7.7 billion.

Okay. Got it. So maybe more of a longer-term question. But the question is rooted in last year's number.

Last year, fiscal '24, the EDA revenue for Synopsys if I back out the extra week impact in fiscal Q1 '24, the growth rate was around 9% year-on-year. But when I look at your peers, they are core EDA revenue also kind of show some sort of a deceleration from double digits to like high single digit in the last fiscal year. So it feels like you guys have probably seen the same thing here, but I wonder you think about the long term, you guys were thinking about 12% over the long term, but was last year kind of hitting a cyclical bottom in terms of EDA revenue growth? Or we still need to weigh a little bit of a recovery, let's say, the non-AI side, you keep referring to a payoff two markets, right, for that to recover to get back to the low double-digit EDA growth.

Charles, thanks for the question. For design automation, it has both EDA software and hardware. So you can expect lumpiness quarter-over-quarter. If you look at trailing 12 months and what we have communicated in our Investor Day, which is a 12% growth for design automation.

That was based on the next 5-year CAGR. And we do believe that, that will continue, and we have no signals at this stage to believe that, that will be any different. The Remember, we sell to grouping of customers. They are the system companies, and there are the semiconductor companies.

The semiconductor R&D investment has been going up. which is very good for Synopsys because as I said earlier, from 6% to 9%, we benefit from that growth because they are investing more in hardware and IP and EDA, et cetera. And then you have the system companies, which is growing at a higher rate than the 9%, and that's why we believe the 12% is appropriate for us to commit to and deliver towards the 12%.

And maybe a quick follow-up on China. Sassine I think last quarter, when you provided the first look at fiscal '25 on China, you adopted like the prior fiscal year, you want to be a little bit more cautious. But you actually did not provide a quantitative guidance on where you think China growth rate or percentage contribution to our total revenues can be in fiscal '25. But I look at the reported fiscal Q1 number, it's a pretty meaningful step down from last year's average China revenue run rate, let's say, $250 million per quarter, but just for Q1 was like $174 million. even if I think China is going to be flat year-on-year for you, that means a pretty meaningful catch-up for the next 3 quarters.

So are you at a place where maybe you can start to call the direction of China, maybe even the absolute dollar base or percentage-wise where the China revenue can go this year in fiscal '25. Are you there yet to make a call?

So Charles, what I can say at this point, last year, we finished at corporate average the deceleration and the headwinds in China are getting stronger, that we don't believe it will be at corporate average. We will finish below corporate average in terms of China growth and that has been accounted for in our guide.

The next question comes from Joe Vruwink with Baird.

Great. Hi, everyone. I see in the 10-Q that backlog count position actually swung a bit more to current balances this quarter. where I think over recent history, you've had strong backlog developments, but it's been particularly notable in long-term RPO and those multiyear engagements. I guess I'm just wondering how you kind of think about the trade-off between annual contract value and total contract value, is true that there's maybe shifting in duration that you generally don't mind as it provides an opportunity to reengage with customers over that contract?

Or are some of the changes we've seen in backlog composition? Is that at all reflecting how customers are preferring to engage with you?

There is definitely no change in terms of customer behavior. And as you know, Joe, the backlog you build and burn the backlog, the the EDA type of contracts, they are still on the same average duration. So we have not seen a big customer behavior change at all. on IP and hardware, as I mentioned in the tale of two markets remarks, that even though the contract may be committed by when the customer pulls it down varies based on the pace in which they are building these chips. So -- but there is no behavior change from customer or market point of view or anything differently we're driving in terms of extending durations with customers. .

Yes. And Joe, I would just add sort of a nuance. When we do our backlog, we also report our FSA that next 12 months of backlog is x the FSA. So if you look at it that way, there's not really as much variation.

Anyway, just -- and again, that's customer preference when we go to sign a contract with them, what's in the base contract and then what's in the FSA. And as you know, the FSA is a committed noncancelable, but it's going to be pulled down when the customer needs it.

Yes. Okay. That makes sense, Shelagh. I wanted to go back to your comments, sent the current generation of AI, you have available the optimization products. that's not necessarily changing design methodology.

Would you say that impacts your view at all that you expressed year ago about how AI could lift industry growth rates for EDA by about 200 basis points? Or I guess, a different way to ask, have you actually seen that type of benefit and it's either because of your product demand or the markets you serve, but that 200 basis points of benefit is maybe being a bit masked by the more tepid results from the analog markets you've been discussing.

Yes. Excellent question, Joe. When we talked a year ago, we talked about two things, if you remember, we talked about Synopsys.ai having two offerings from Synopsys, the AI optimization and the generative we were not talking at the time about a Gentech AI because we thought it was further down the road. And we kind of teased autonomous design at the time but it did not -- from a technology point of view, feel that it's coming as fast as it's coming right now.

From a monetization point of view, the AI optimization, we are monetizing it. But is that monetization in aggregate will result on the 200 basis points, not by itself, no. But when you add generative and Igentec, are we still seeing an opportunity when we change the workflow for our customers and give them a different approach to design a chip cheaper, faster for us to monetize, given the business model for Agentic will be very different. The answer is yes.

And that's why if you recall during our Investor Day, when we talked about AI optimization and generative, we did not put necessarily a time line for that 200 basis point for that specific reason. But adoption from customers excitement about the technology, penetration of the optimization across DSO, VSO, et cetera, is something actually where we're very positive about and exactly what -- where we thought we will be in the cycle.

The next question comes from Jason Celino with KeyBanc Capital Markets.

I just wanted to follow up on some of Charles' questions on China. So compared to 90 days ago when you initially set guidance, have your assumptions changed at all? I understand below corporate average, it makes sense. But if this is a change, the implication, if it's true, since you're holding the full year guidance the same, is that something maybe in your EDA or IP business might have improved since 90 days ago.

So I wanted to clarify that and see if it's the case on those segments.

Yes, Jason, the assumptions in our planning and therefore, guiding the year has not changed. We've always believed that there will be headwind and ongoing deceleration in China. What we -- I want to say, in our communication have our clarifying and changing is Will it end similar to last year, the deceleration towards corporate average? Or do we believe it will go further below corporate average.

And that's really the nuance in clarifying where do we believe China will be. But the assumptions we made in our guidance and our internal forecasting based on regions the portfolio, EDA software, IP hardware, none of that changed.

Okay. And then the HAPS 200, the ZeBu 200 announcements, you upgrade your products more frequently than your two competitors. How should we view these versus the prior iterations. I'm just trying to wonder if we could see an air pocket in demand as customers wait for their orders in the second half?

That's the advantage we have with an FPGA versus a custom chip that you have a much faster refresh cycle and ability. So with the HAPS 200 and ZeBu 200, actually, no air pocket. And the reason for that is the same customers that they were buying from us the EP. Remember, we talked about the EP before.

Now with the next generation, they have already are purchasing the prior generation EP plus the new capacity that we can provide with the 200. And as Shelagh mentioned, right now, the demand is absolutely there. It's our ability to deliver to that demand, which we're building the capacity and our ability to expand it.

Yes. And Jason, my comment was more about availability, not about air pocket in demand. So that is much more Q4 weighted.

The next question comes from Jay Vleeschhouwer with Griffin Securities.

So Sassine, I'd like to ask your tale of two markets point from a different perspective, and that is strictly with respect to semi R&D. And that is we do seem to have seen over the last number of quarters, more concentration of those semiconductor companies that are continuing to show good or even very good growth as compared to what we might have seen previously where the growth was broader among the population of semi companies. Is this something that you've seen or might be concerned about that in an already concentrated market in the semis, the sources of really good growth are becoming fewer and fewer.

Then I have my follow-up.

Jay, as you can imagine, we are very intimate with every customers, chips, road map, investments, et cetera, and we track it very, very closely -- the -- if you look at it in aggregate, the increased investment in order to support their road map, be it very advanced chips and two and three multi-die advanced package or even actually earlier this week, discussions with automotive and industrial customers looking for the next opportunity to deal with more integration on their chip and move to the next node, albeit it is [indiscernible] MC or M3, but going to FinFET, for example, in order to provide a cheaper, more competitive chip.

That's a great opportunity for us because they are investing more in R&D and they're doing a refresh per se on the IP that they want to pull down, using more hardware using the latest EDA software, et cetera. So in aggregate, when we see the growth from 6% to 9% in the semiconductor R&D we are seeing how it's matching with our growth, not like they're growing from 6% to 9%, and we're staying flat. We are growing with their growth of R&D investment. In aggregate that's as a whole.

Yes.

Right. Understood. So your point about changes in productivity and workflow and the like is historically, I think, really interesting. And over the last for decades that commercial EDA has existed.

It has probably always been the case that there are differences in how customers employ EDA and get a return on investment has never been equally distributed productivity, let's say. And you look at a current example, we're the largest spender on EDA spend a higher percentage of R&D on commercial EDA probably more than double the next largest company and yet we can look at differences of return on the employment of the EDA. So the question is, as you infuse you and your peers infuse more AI capabilities across the design process, do you think that, that historical differentiation of capability changes and customers become perhaps more and more alike in their capabilities because of the availability of this whole new class of technology.

The leading customers, they -- I want to say, they worry about as they invest a bigger amount of their budget towards R&D. Will AI emocratize across and will make the net catch up faster to the #1 because of the AI and improving productivity. The answer is the following. If you look at the workflow, so far, AI has not changed the workflow.

The steps that you go through from phase -- the first phase until tape-out of a chip has been the same from a workflow point of view, even though there has been tremendous amount of automation and innovation to deal with the increased complexity, where we believe the genic AI will have an opportunity is to change the workflow. And the moment -- and by the way, this is no different than other leading enterprise software companies, the way they are talking about Agentech. The moment it changes the workflow, then you have an opportunity to monetize. So far, EDA for the last x number of years or a couple of decades has provided amazing amount of technology and innovation, but the workflow has remained the same.

So that's the nuance that we're highlighting in terms of opportunity with Agentec.

The next question comes from Nay Soe Naing with Berenberg.

Just got one, please, on your cost control in the quarter and then how you guidance next quarter as well. If I got my math right, I think you're guiding towards at the midpoint about 5% year-on-year total expense growth. I think the quarter came in around 2%. So really nice outperformance there.

But for Q2, you were guiding for cost growth about 8.5% at the midpoint. So in light of how you performed in Q1, how should we think about the Q2 guide? And also, if you could maybe just on thoughts on the cost development in H2 as well, please?

Sure. So we were a little bit lighter on costs in Q1 than we had originally anticipated, and that really came down to just timing of hiring and timing of some big ticket expense items. We expect no change for the year. Structurally, Q2 always steps up because that's when our merit, our annual performance budget kicks in.

And so that's the structural change that you see between Q1 and Q2, but no change for the full year, even though Q1, we were a little bit lighter just due to some timing element.

We have time for one more question. It will come from the line of Joshua Tilton of Wolfe Research.

Guys. I really appreciate you sneaking me in here. Sassine, maybe this one is for you. It's more of a first, a clarification and then trying maybe a point of understanding.

Like if I remember correctly, the top track around the guidance for this year coming out of Q4 was that you guys expected China to grow in line with the corporate average. And I feel -- correct me if I'm wrong, but I feel like now the message is that the expectation is that China is actually going to grow below the corporate average for the year. So my first part of my question is just to clarify if that's correct. And I guess my follow-up part to that question is, I guess what gives you guys the confidence to reiterate the guide given this now that you feel like China will be growing below the corporate average for the year.

Yes. Joshua, you got it correctly. When we guided FY '25, given we just came out of FY '24, where China decelerated to corporate average, we did say that based on our guide of FY '25, it will be around the corporate average growth. Now given the headwinds that we are seeing and I want to say they're getting stronger, even after we announced or guided number of customers got added to the entity list, et cetera, we believe that China by itself will be decelerating below the corporate average.

However, the guide in itself is not changing. We see a number of strengths in technology and other regions, customers that we feel very good about the guide that we provided for the year.

Super helpful. Maybe is there any way you could just take that one step further and just maybe unpack one layer deeper into some of the strengths that you're seeing that is offsetting the China weakness? .

Yes. If for example, the one that we just talked about, the HAPS 200 and ZeBu 200 family. That's a great opportunity for us to monetize IP and the strong demand and requirement from system companies and AI, HPC grouping of semiconductor given our scale for that's another opportunity. Today with IP, I want to say it's really a challenge and an opportunity around scaling.

It's not the lack of opportunities. We have plenty of opportunities and given our position and leadership position in the market, that's a great opportunity. These chips require very advanced EDA and to reemphasize what I said earlier about AI optimization, the latest Fusion design platform, the ICV, the IC Validator competitive displacement, ASO, I mean, those are all things that are giving us great confidence that both growing share and being able to monetize that share and therefore sticking to the guide that we provided.

Thank you. This concludes today's conference call. We thank you for joining. You may now disconnect. Thank you.