Credo Technology Group Holding Ltd (CRDO) Earnings Call Transcript & Summary

Ladies and gentlemen, thank you for standing by. [Operator Instructions] I would now like to turn the conference over to Dan O'Neil. Please go ahead, sir.

Good afternoon. Thank you for joining us on our fiscal 2023 third quarter earnings call. Joining me today from Credo are Bill Brennan, our Chief Executive Officer; and Dan Fleming, our Chief Financial Officer. I'd like to remind everyone that certain comments made on this call today may include forward-looking statements regarding expected future financial results, strategies and plans, future operations, the markets in which we operate and other areas of discussion. These forward-looking statements are subject to risks and uncertainties that are discussed in detail in our documents filed with the SEC. It's not possible for the company's management to predict all risks nor can the company assess the impact of all factors on its business or the extent to which any factor or combination of factors may cause actual results to differ materially from those contained in any forward-looking statements. Given these risks, uncertainties and assumptions, the forward-looking events discussed during this call may not occur and actual results could differ materially and adversely from those anticipated or implied. The company undertakes no obligation to publicly update forward-looking statements for any reason after the date of this call to conform these statements to actual results or to changes in the company's expectations except as required by law. Also during this call, we will refer to certain non-GAAP financial measures, which we consider to be important measures of the company's performance. These non-GAAP financial measures are provided in addition to and not as a substitute for or superior to financial performance prepared in accordance with U.S. GAAP. A discussion of why we use non-GAAP financial measures and reconciliations between our GAAP and non-GAAP financial measures is available in the earnings release we issued today, which can be accessed using the Investor Relations portion of our website. With that, I'll now turn the call over to our CEO. Bill?

Thank you, Dan. Good afternoon, everyone, and thank you for joining the call. During this call, I'll discuss Credo's FY '23 third quarter results and our updated outlook, along with the industry trends and customer engagements that have us optimistic about our future growth. Credo is a pure-play high-speed connectivity company. We built our first solutions for the Ethernet market. Our current portfolio of high-performance and low-power connectivity solutions include, active electrical cables or AECs, integrated circuits, or ICs, SerDes chiplets and intellectual property or IP. All of our product and IP solutions leverage our unique application-specific SerDes portfolio, enabling us to deliver optimized, secure, high-speed solutions with better power efficiency and cost. Our connectivity solutions address both electrical and optical applications at port speeds currently ranging from 50 gigabits per second up to 1.6 terabits per second. Our customers include hyperscale data center operators, networking equipment OEMs and ODMs, optical module manufacturers and others in the Ethernet ecosystem. While we primarily serve the Ethernet market today, we continue to expand into other standards-based markets as the need for higher speed and more power-efficient connectivity increases exponentially. For Q3, Credo reported revenue of $54.3 million, representing a 6% sequential growth and year-over-year growth of more than 71%. Additionally, we reported non-GAAP gross margin of 59.3%. Both revenue and gross margin were within our guidance for Q3. Prior to giving an update on our progress with our business, I would like to briefly discuss the 8-K we filed 2 weeks ago. We filed the 8-K because our largest customer reduced their near-term demand forecast for our products. We continue to believe the reduction was due to reasons unrelated to our performance and that we've maintained our sole source position at this customer. This reduction in forecast, along with the overall macroeconomic headwinds in the market will impact our revenue in Q4 and fiscal '24. We strongly believe the fundamental connectivity technology trends remain consistent across the data infrastructure market. The need for speed, power efficiency and collaborative problem-solving is ever increasing. Our long-term financial model remains unchanged, and we remain optimistic about our growth opportunities, given our ongoing engagements in several advanced programs with our largest customer and other hyperscalers as well as 5G carriers. Now moving on to an update on our overall business, beginning with AEC. During Q3, our growth was driven by our AEC solutions, a product category we pioneered and which continues to gain broad industry acceptance. Our growing list of [ AEC ] customers are migrating to AECs for short in-rack cable solutions for various reasons, including the limitation of alternative solutions and a unique tailored feature set of our products. As data center and other network operators migrate to higher speed lane rates of 50-gig or greater, AECs offer better signal integrity and physical attributes compared to DACs and our AECs offer compelling benefits over AOCs across power, cost and reliability. Additionally, our AECs offered functionality our customers can use to innovate on rack architectures for servers in both compute and AI/ML applications as well as for disaggregated switch racks. Industry analysts expect AECs to become a multibillion-dollar market within 4 or 5 years as AECs become the most prevalent solution for in-rack connectivity. While we'll be shipping reduced volumes to our largest AEC customer in the near term, we continue to engage with them on future programs on the roadmap for both compute and AI server racks, and we remain confident in the strength of this relationship. I'm happy to report that Credo has begun the early stages of the production ramp with our second hyperscale customer. And although the detailed time line of their volume ramp is still developing, we expect meaningful contribution in fiscal '24. We're also deeply engaged in developing advanced AEC solutions with this customer for their next-generation server rack and switch rack applications as they move to 100-gig single lane speeds. Additionally, we continue to make progress with other customers as well. We're in flight with 2 additional hyperscalers for switch rack and AI server rack applications using both 400-gig and 800-gig AECs. While hyperscalers remain our primary focus, we're engaged in meaningful opportunities with other data center operators and 5G carriers. Regarding our progress on optical solutions, Credo also leverages our differentiated SerDes technology to deliver disruptive solutions to the optical market. In this category, we deliver a breadth of products across DSPs, laser drivers and TIAs for 50-gig through 800-gig port applications. We've seen success thus far securing wins with both optical module manufacturers as well as with hyperscalers directly through joint development models. While the time to achieve our volume production revenue ramp has shifted in the current environment, we continue to play the role of disruptor in the optical market. Our initial success has come with optical DSPs for the 200-gig and 400-gig port markets, and we're confident our recently introduced Dove 800 optical DSP provides considerable differentiation. Enabled by our superior SerDes technology, we remain confident we will gain share over time due to our compelling combination of performance, power and cost. Credo will have a strong presence at the OFC conference in San Diego next week, where we'll be showing several demonstrations with our Dove 800 family products. I look forward to seeing many of you there. Moving to our Line Card PHY solutions, Credo continues to extend our leadership in the Ethernet Line Card PHY market across hyperscalers and networking OEMs and ODMs. Our solutions include MACsec PHYs for high security applications needing encryption, as well as retimer and gearbox solutions at 50-gig and 100-gig lane speeds. Our recently announced 1.6T screaming Eagle Line Card PHY, which is a long-reach DSP retimer and gearbox, highlights our performance and power leadership in this product category. I will note that we've delivered the most power-efficient Line Card PHY solutions in the market using the 12-nanometer process. As we move to 5-nanometer, we expect to achieve a 40% decrease in power, forcing our competition to move to 3-nanometer to compete and as a result, creating significant time to market and cost advantages for Credo. I would now like to give an update on our SerDes IP licensing and SerDes chiplet business. Our Q3 results were bolstered by revenue attributable to our lead consumer electronics customer and 56-gig Ethernet licensing. While our IP business is variable due to recognition rules and the IP sales cycle, our customer funnel is robust. Here, too, we have confidence in our growth prospects due to customer feedback. Customers have shared that our 5- and 4-nanometer 112-gig SerDes IP offers a 40% to 50% power advantage compared to the competition, depending on the reach of their application. This again highlights our N-1 process advantage, which means to achieve the power efficiency of Credo's 5- and 4-nanometer IP solutions, customers would need to move to 3-nanometer if considering our competition. Finally, regarding the SerDes chiplet opportunity, market trends, including the advancement of the UCIE consortium, suggests that SerDes chiplet market will become a meaningful long-term opportunity for us. Our leadership position in the SerDes chiplet market is evidenced by our multiple production revenue wins in this category, including with Tesla and Intel. In summary, Credo remains laser focused on the large opportunity afforded us by our differentiated solutions in a market that's demanding higher speeds and better power efficiency. Our N-1 process advantage gives us a distinct advantage across multiple competitive axes, but maybe most importantly, in power. Power efficiency is increasingly important, and credo can help our customers lower electricity consumption through use of all of our products. Clearly, we're disappointed about the adjustment in the near-term outlook given the disruption with our largest customer. However, the feedback from that customer as well as from other customers and the additional developments I just discussed, fuel our optimism about our growth prospects in the future. We remain committed to close customer collaboration, continued innovation and the expansion of our solution portfolio to address the ever-increasing needs for higher bandwidth and more power-efficient connectivity solutions. At this time, I'll turn the call over to our CFO, Dan Fleming, who will provide additional details. We'll then open the line for questions. Thank you.

Thank you, Bill, and good afternoon. I'll first review our Q3 fiscal 2023 results and then discuss our outlook for Q4 of fiscal '23. As a reminder, the following financials will be discussed on a non-GAAP basis, unless otherwise noted. In Q3, we reported revenue of $54.3 million within our guidance range, up 6% sequentially and up 71% year-over-year. Sequential growth was driven by continued traction in our IP business, which reached $12.6 million in Q3, up 285% sequentially and up 145% year-over-year. IP remains a strategic part of our business, but as a reminder, our IP results may vary from quarter-to-quarter, driven largely by specific deliverables to pre-existing contracts. While the mix of IP and product revenue will vary in any given quarter over time, our revenue mix in Q3 was 23% IP, above our long-term expectation for IP, which is 10% to 15% of revenue. We continue to expect IP as a percentage of revenue to come in above our long-term expectation for the current fiscal year. Our product business generated $41.7 million of revenue in Q3, down 13% sequentially and up 56% year-over-year. Year-over-year growth in product revenue was driven principally by AEC growth. With a strong IP result this quarter, our team delivered gross margin of 59.3% within our guidance range and up 438 basis points sequentially. Our IP gross margin generally hovers near 100% and was 98.2% in Q3. Our product gross margin was 47.6% in the quarter, down 496 basis points sequentially and down 585 basis points year-over-year, due principally to revenue mix. Total operating expenses in the third quarter were $25.7 million, below the midpoint of our guidance and up 41% year-over-year, considerably below our 71% year-over-year revenue growth. Our OpEx increase was driven by a 53% year-over-year increase in R&D as we continue to invest in the resources needed to deliver innovative solutions. Our SG&A was up 26% year-over-year as we built out public company infrastructure. We delivered non-GAAP operating income of $6.7 million in Q3, an improvement of $5.3 million year-over-year and up 97% sequentially. Our operating margin was 12.3% in the quarter, an improvement of 8 percentage points year-over-year and up 569 basis points sequentially due to top line growth that resulted in higher gross profit. We delivered net income of $7.5 million in Q3, an increase of $5.1 million year-over-year and up 208% sequentially. Cash flow used by operations in the third quarter was $2.3 million, a decrease of $4 million year-over-year and a decrease of $4.1 million sequentially due largely to an increase in inventory. CapEx was $6.9 million in the quarter, driven by production mask spending. And free cash flow was negative $9.2 million, a decrease of $8 million year-over-year and a decrease of $5.3 million sequentially. We ended the quarter with cash and equivalents of $233 million, a decrease of $7.5 million from the second quarter. This decrease in cash was a result of capital expenditures and working capital investments. Our accounts receivable balance decreased 16.6% sequentially to $43.2 million, while days sales outstanding decreased to 72 days, down from 92 days in Q2. Our Q3 ending inventory was $50.3 million, up $2.5 million sequentially. Now turning to our guidance. We currently expect revenue in Q4 of fiscal '23 to be between $30 million and $32 million, down 43% sequentially at the midpoint. We expect Q4 gross margin to be within a range of 56% to 58%. We expect Q4 operating expenses to be between $26 million and $28 million. We expect Q4 weighted average diluted share count to be approximately 163 million shares. And finally, we expect fiscal year 2024 revenue to be flat compared to fiscal year 2023. And with that, I will open it up for questions.

[Operator Instructions] And our first question will come from Toshiya Hari of Goldman Sachs.

I had 2, if I may. The first one is on your largest customer. Bill, I just want to better understand what's driving the change in the near-term outlook here? Is it simply a push in the customers' time line and projects? Or are you seeing cancellations from them? Is it an inventory dynamic? I think in your prepared remarks, you mentioned that you're confident that you're still sole source there. But any risk of a fundamental change of their infrastructure going forward?

Sure. So clearly, this is a complex customer that we're working with. We fundamentally believe it's a combination of 2 things. One is inventory that they're working through. I think there's a change in behavior coming off of the supply constraints that we saw over the past few years. And I think they're bringing their inventory levels down to more of a normal level. This is combined with other pockets of inventory that were built up within our supply chain. So I think that's the first factor. The second factor is really that they've seen pushouts in key programs, additional programs that use our AECs, there's been delays of qualification and pushouts due to various factors. I will reiterate that we have nothing to suggest that we're no longer a sole source position with this customer. We feel pretty confident with that. And we feel pretty confident that there's no technical issues with our products. So I think it really boils down to inventory as well as new SKU pushouts and delays.

Got it. And then as my second question on the fiscal '24 outlook maybe for Dan or Bill, for that matter. So you're guiding to a flattish revenue outlook. Prior to the 8-K, I think the Street was modeling somewhere in the low $300 million range for revenue. So clearly, there's been a pretty significant change there. Can you sort of break out the individual sub-components, if you will, the drivers of that change, the largest customer, the dynamics that you just talked about, if it's a push, I guess, I would think and I would hope some of that revenue materializes in the second half of fiscal '24, but how are you thinking about that? Bill, you talked about your second AEC customer ramping in fiscal '24. And then obviously, you have a robust Line Card business, IP business as well. So can you talk through the individual components as you think about fiscal '24?

Yes, Toshiya, let me address that initially. So starting with our largest customer, what you'll see in our Q as we file it in a day or so, our largest customer was 52% of our revenue in Q3, which was up from 44% in the prior quarter. So I mentioned that because the trajectory of that customer was still on the upswing, right? So with this new forecast reduction, that trajectory obviously changes, we previously -- 2 weeks ago, we had described this FY '24 reduction is about 2/3 related to this largest customer and then about 1/3 of it due to just increased conservatism that we're applying across the board based on just macroeconomic factors that we see. So if you do the math, our largest customer remains still a very large customer in FY '24. The second hyperscaler, we expect to see ramping throughout the course of the year. But we're quite comfortable right now where -- with analyst estimates for the full fiscal year '24 if that's helpful.

And sorry, one quick follow-up to that, Dan. The 1/3 of the fiscal year '24 cut, are you baking in conservatism in your IP business as well? Or is that primarily the product side outside of your largest customer?

It's -- you should consider it just kind of across the board.

[Operator Instructions] And our next question will come from Quinn Bolton of Needham & Company.

I wanted to follow up on questions around your largest customer. I guess, at the end of the day, do you think the opportunity set for AECs has changed at that customer meaning that the steady state revenue you thought you would get to, say, 6 months ago, do you still think you can get there? And I guess a related question, it seems like there may be some confusion in the application of your cables. What percent of your largest customers' servers today are connected with InfiniBand for AI applications versus Ethernet for just standard compute applications or instances?

Let me answer the second part first. The -- so the applications as we understand it, where our AECs are being used is in the compute portion of their server deployments. And so that would line up with Intel and AMD processors as well as the nicks that connect those servers. We're -- from the standpoint of InfiniBand, there's not really an opportunity for us there. InfiniBand is a slightly different protocol. Our AECs are not built to support InfiniBand. From a pure speed standpoint, of course, the speeds are something we would be able to negotiate, but there are specifics related to the protocol that -- are just not something that we've qualified. If you wanted to speculate in the future about could we build InfiniBand cable? Yes, we could. But right now, the way that we view it, the market is a pretty closed market and it's not something that we're spending cycles on. And Quinn, will you repeat the first part of your question?

Just -- I've got a follow-up on InfiniBand, but let me come back to the first part of the question, which just 6 months ago, if you had an internal estimate of what you think AECs could reach at a steady state, an annualized revenue opportunity at that customer, has that changed given the customer forecast or the pushout and SKUs that you mentioned?

Well, I think the way that I look at that is 2 ways. I look short-term and then I look long-term. And I would say short-term, look, I mean, we were -- our confidence was bolstered by the backlog I had at hand, the forecast that I had at hand. Clearly, we're learning that things can change. And I don't think this is something fundamental with the change in strategy on deployments. And so the way that I felt about it maybe 3 to 6 months ago was that our opportunity would be pretty aligned with the server deployments for the near-term compute portion of their spend. And so from the standpoint of how that will change over the kind of the short to mid-term, look, that's really something that they're planning right now, but I don't think there's a fundamental shift in strategy. And then the other way that I answer the question is really long term. And I would say that I'll go back to the fundamentals and say that there's 2 main reasons why our AECs are being adopted at any given customer. And that's, first of all, functionality, we're able to build innovative functionality into our solutions, which is why our first customer ramped, quite frankly, the white cable, the switch cable that we're selling plays a role that no other cable solution in the market could accomplish. The other reason that we're seeing conversion is [ speed ] and we've talked about this as lane rates go faster, DACs become obsolete, DACs being passive copper cables. And there's a fundamental advantage with AECs when we're talking about just a pure head-to-head competitive situation with AOCs, we're at a much lower cost, much lower power solution. And maybe reliability is the most important thing. We're 10x at a minimum, more reliable than an AOC. And so the way I see it playing out at our largest customer is that over time, our opportunity will really expand as they move from just compute to factoring in more of an AI spend. I see all of the server racks, whether they'd be compute or AI long term, will be a really robust opportunity for us as speeds move from 25-gig to 50-gig and then ultimately 100-gig per lane where I think we'll see a very strong correlation between deployments and the AEC TAM.

I guess, Bill, just to follow up on that. What percent of the AI racks today are connected with Ethernet versus InfiniBand because it sounds like you're not playing in InfiniBand, so AI today is connected with InfiniBand, it sounds like an architecture change may you have to occur. And then one for just quickly for Dan. Maybe my numbers are up, but Dan, it looks to me that despite -- I know you're mix shifting to AECs, but the product gross margins look like even on the AEC, they've probably come down over the past couple of quarters, even though AECs have ramped pretty nicely, and I would have thought there would have been some volume benefits to that AEC ramp. So can you give us some sense specifically what's going on with AEC margins? Because it looks like even with the mix shift to AECs, that the AEC margins themselves have come down over the past couple of quarters?

So let me start just by addressing your product margin question. So what I mentioned earlier in my script is that really it was revenue mix related in Q3. But there are, as you'd imagine, a number of factors. Revenue mix being the key one because a higher percentage of our overall revenue was AEC. So simply the fact that AEC has a lower gross margin than our corporate average, that mitigates some of the scaling benefit we would get. But the second item, over the prior 6 quarters, like you alluded to, product margin has expanded quite nicely due to that increasing scale, but our overall product revenue was actually slightly down sequentially in Q3. So that mitigated some of that gain you might have expected. And then finally, the last thing just to mention with product margin in general for Q3 is that our forecast recently declined for fiscal year '24. So when you calculate things such as inventory reserves, the math changes because you're comparing against your forecast. So you would expect us to take a greater reserve position than in prior quarters, just fundamentally. So those -- a combination of those factors really drove that decline in product margin sequentially.

Got it. So it sounds like there might have been some inventory reserves -- higher inventory reserves in the January quarter given the forecast change that we might have missed. Okay, great. And then Bill, I don't know if you have a sense of what part of the AI infrastructure today is InfiniBand versus Ethernet.

I can't be specific about that, but I can tell you my thoughts are that I think first mover has been InfiniBand from a protocol perspective. But I can tell you there's -- across the industry, there's an extremely robust pipeline of Ethernet applications for the AI space.

[Operator Instructions] And our next question comes from Tore Svanberg of Stifel.

Bill, this is a bit of a longer-term question, but it's kind of like due to recent events. So you're vertically integrated, you sell the whole system in AEC. Could that have led to your larger customer perhaps building a little bit more inventory? And as you think longer term, is there a strategy to selling semiconductors and perhaps then outsourcing the manufacturing of the cable or the sale of the cable that is?

So I think that -- I think our model -- I guess I would say that it really wouldn't impact the amount of inventory somebody would hold, I would say. The bottom line is, we were a one component of many that were being sourced for the racks that were being built. I don't think there was any kind of special scenarios for Credo. And I don't think the model generally would lead to that in a general sense. The second part of your question, I think we've talked about before, it was a necessity for us to take ownership of the complete system. And we go back 4 or 5 years ago, and it was really necessary for us to build a vertical organization, a group that was fully responsible for the AEC definition development, engineering, qualification and ultimately fulfillment and support following production. I think that was critical for the success of Credo with our customers and that includes our -- the first multiple customers that we've got. And I think it's going to be important long term that we maintain ownership of the complete system. The idea of us selling a semiconductor to a company that makes passive copper cables today and thinking that they would have the wherewithal to bring a highly complex solution to market. I think that was the first idea that we had, and we quickly shifted our strategy. And so when we think about, is there a different model than us selling the complete AEC solution, I think long term, there is. And we've talked about the idea in certain cases where it would make sense going to a reference design model. And that would be on a case-by-case basis where volumes would justify it. But our role in the development and support of the solution will not change. This is one where we would stand behind the design, we would bring it up, certify it, and we would support it in production in the same way that we do today. So we've really had a chance to think through this. And the scenario of us selling a chip to a copper cable company is not a scenario that exists, but we will be able to achieve a shift in business model that eliminates full margin stacking and that will play itself out in the future.

And my second question is, could you just give us a time line update on going from 12- to 5-nanometer. You mentioned it a little bit when you talked about the Line Card PHYs, but if you look at the different business units, obviously, IP is different, right? But when should we expect products in 5-nanometer across the various product lines?

So this is something that we're actively in flight on today. And we've got 5-nanometer developments really across the board for all of our product segments. The tape-outs will occur in the upcoming quarters, and we'll have products probably within the next 12 to 18 months.

Just one last question. I think there's a lot of focus on [ ASC ] but you did highlight the chiplets IP, especially with some non-communications customers. How should we think about that business for fiscal '24? Will that actually grow with the others declining?

Well, I would say that our IP business remains something that's strategic to us. We don't see it growing beyond our long-term model, which has been stated at 10% to 15% of our revenue. There's no change in that. Short term, we'll probably outperform that based on the reduction in the product forecasts that we're giving. But I don't think there's any fundamental shift in the way we're looking at that business. I think that we see in the future that we will have success in advanced processes, high speeds. I think we will have success with other protocols as well because we offer really unique solutions for those markets.

I guess I was talking more specifically about chiplets. So it sounds like that's probably more of a fiscal '25 driver, but will you have any growth at all in fiscal '24?

No, I think it really is a fiscal '25 driver. And I think the inflection point will be the UCIe chiplet market as that takes off. And so the original -- I guess the idea is that as PCIe lane rates move from 32 to 64, that's where we're going to enter the market and it maps directly to UCIe chiplets as well.

[Operator Instructions] And our next question comes from Suji Desilva of ROTH Capital.

So first question really to understand the AEC at the lead customer. Can you give us a rough idea how many programs you're in kind of ballpark to understand if the impact was uniform across those programs? Or whether certain programs that are impacting others, weren't?

So I think as we look at the number of different rack SKUs that we're involved in, it's probably on the order of 4 to 5. And the different SKUs, there's different components that change within the SKUs from the servers and NICs that they're using to the ToRs that they're using, but our AECs are consistent. I don't think there's anything that I could point to that would suggest there's something fundamental that's changing. The first SKUs ramped and I think the other high-volume SKUs that were planned have been delayed.

And then the other question as you guided fiscal '24 along with the fourth quarter, wondering if you could kind of give us some sense of the linearity or shape of the quarters and the recovery? Would it be more back-end loaded or perhaps a linear recovery from the current inventory digestion?

Yes. So from a revenue perspective, it should be relatively linear throughout the year in terms of a steady progression. Recall that we stated our -- we expect our bottom to be either Q4 or Q1. Q4, we guided $30 million to $32 million. So if you kind of view something similar to that in Q1 to get to flat year-over-year, it shows pretty strong sequential growth throughout the year. So we do expect to exit the year at a very high rate of revenue.

[Operator Instructions] And our next question will come from Vivek Arya of Bank of America.

So I think in fiscal '23, AECs are roughly, I imagine about 40%, 45% of the sales mix. I don't know if I'm -- it's perfectly right, but it's probably in that ballpark. How much do AECs represent of the mix in fiscal '24, roughly?

Yes. So we haven't given specific guidance on that. But -- and a lot of that really depends on the shape of the curve from the second hyperscaler. But right now, to get to flat year-over-year, I would expect it to be down a little bit, but that could change as we kind of get closer to entering FY '24 and give more specific guidance.

All right. Because then when I look at -- you mentioned a very steep ramp in the back half. So when you look at your second half of fiscal '24 growth versus the first half, is that mainly coming from the AEC ramp at the second hyperscaler? Is that kind of the big ramp? Or is there anything else in the non-AEC part of the business that can also grow significantly, right, year-on-year in that time frame?

Yes. So the short answer to that is, yes. It's largely AEC related. We'll be at a lower point in Q4 and Q1, and that recovers throughout the year.

Got it. And my quick follow-up, Dan, is if, let's say, your revenues next year are sort of flattish, what happens to gross margins and OpEx that you're roughly thinking about for fiscal '24?

Yes. So the way we've looked at gross margin, we expect it to relatively be flat year-over-year. And the reason we say that or how we get there is we -- if you look at all the different things that are changing, we lose scale versus our prior expectations, right, for FY '24, which obviously is negative to margin in terms of expansion by scale. However, AEC has come down versus what we had originally thought. Again, that's favorable to overall margin. And even though the remaining, let's say, 1/3 of the reduction is kind of across the board, things like IP to become a larger percentage of the overall revenue mix. So those favorable revenue mix things really offset the loss of scale is our expectation. So FY '24, we would expect to be similar to FY '23 gross margin.

And OpEx [ raise ] Dan?

So OpEx, we guided the midpoint of our guidance for Q4 was $27 million. So a couple of things to bear in mind there is that we have an annual focal process that's calendar year based. So that's kind of -- our Q4 number should be kind of our all-in OpEx from a headcount-related perspective as we're kind of maintaining our overall headcount and not adding incremental heads through the year. So there will be some variability quarter-over-quarter we would expect. But our Q4 guide should be pretty consistent on average throughout the remainder of fiscal year '24.

[Operator Instructions] And our next question comes from Yang Pu of BNP Paribas.

This is Karl Ackerman at BNP Paribas. Two, if I may. Bill, and organic, I guess, could you discuss whether the qualification ramps of your AEC offering at the previously announced 3 other hyperscalers has been impacted at all from the design considerations or the pause at your largest customer?

I think they're largely unrelated. I don't see any kind of correlation between the 2. What's happening in our largest customer is a function of their own planning. The second customer that we're in the early stages of ramping, the situation there is that I think the overall plan is coming together. The sequence that they go through, it seems similar with these large hyperscalers is there's a group of engineers that are responsible for developing and qualifying it given the SKU and that -- when that qualification is complete, that SKU becomes available for different businesses to pick up and deploy. We've made it through the qualification portion with our second customer. And now I think the detailed forecast is going to be coming together over the next several months. And so as we look at that program, generally, we think that it can drive volume, but we don't want to be overly aggressive in how we're planning. And that's why we've kind of stepped back a bit and put a more conservative look on it until we've got really good visibility.

I guess, Dan, then a question for you, if I may. I did notice a small impairment charge in the quarter. I was curious if you could discuss what that is. And then second, how fungible are your AEC products across the design engagements, you have at some of these hyperscalers? I guess, is your -- is your current inventory reusable? Or is there an obsolescence risk that we should worry about or think about [ given the sort of ] with revised outlook?

Yes. So we don't expect there to be generally speaking, any obsolescence risk in our inventory right now. I did mention we've increased our reserve. That's kind of a -- just a mathematical equation and how we -- based on our internal procedures and how we reserve inventory. Speaking of the impairment, though, you saw a $2.4 million impairment and that was associated with the program that we terminated during the quarter. We've been -- if we look back in time, we've been very extremely fortunate with our success across a lot of different investments that we've made over the years. But in this case, we made the decision to move forward with a different -- a higher ROI product, which can happen from time to time. So you can expect us to continue following just a very strict fiscal discipline as we look at our various ROIs on products that we tape out. And eventually, this will lead us toward our long-term financial target model.

And there are no further questions at this time. I'd like to turn the call back to Mr. Brennan.

All right. Well, I'd like to thank everybody for joining the call. I appreciate all the questions. And with that, we'll end the call. Thank you very much.

This concludes today's conference call. You may now disconnect.