Ladies and gentlemen, thank you for standing by, and welcome to the Credo Technology Group second quarter fiscal year 2023 earnings conference call. At this time, all participants are in a listen-only mode. Later, we'll conduct a question-and-answer session. At that time, if you have a question, you will need to press star one one on your telephone. I would now like to turn the call over to Dan O'Neill. Please go ahead, sir.
Good afternoon. Thank you for joining us today on our earnings call for our fiscal 2023 second quarter. 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 in 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 statement.
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, thank you to everybody for joining the call. During this call, I'll review Credo's fiscal Q2 results and share why we remain excited about our future prospects. After I conclude, Dan Fleming, our Chief Financial Officer, will provide a detailed review of our financial results and expectations moving forward. Credo is a pure play, high-speed connectivity company. We built our first solutions for the Ethernet market and are extending into other standards-based markets as the need for higher speed connectivity increases exponentially. Today, our product families include integrated circuits or ICs, active electrical cables or AECs, and SerDes chiplets. Our intellectual property or IP solutions consist primarily of SerDes licensing. Our connectivity solutions address both electrical and optical applications at port speeds currently ranging from 50 Gbps up to 1.6 Tbps.
All our product and IP solutions leverage our unique application-specific SerDes portfolio, enabling us to deliver optimized, secure, and high-speed solutions with better power efficiency and cost. This has led to high growth rates with hyperscale customers and the ecosystem of suppliers that provide the infrastructure for these data centers. Credo continues to be one of the fastest-growing companies in the semiconductor industry, and I'm pleased to report that we achieved record revenue of $51.4 million in the October quarter, an increase of 94% year-over-year and 11% sequentially. I'll now give a brief update on our progress across our various solutions, starting with AECs. For AECs, we continue to deliver strong execution as the pioneer of this product category. In rack, low-speed cabled connections have historically been made with passive copper DACs.
As single lane speeds increase to 100 Gbps , DACs become obsolete due to signal integrity and physical size constraints. The industry assumption has been that when DACs are dead, optical cables or AOCs would take their place. Credo saw an opportunity to develop a broad product family of AECs that deliver half the power, half the cost, and with 10 times better reliability than AOCs. We also saw the opportunity to offer compelling functionality that enables our customers to innovate on server rack and switch rack architectures. Industry analysts are now forecasting AECs to grow to a multi-billion dollar market in the next 4 to 5 years. We continue to ramp volumes with our first customer as they broaden deployments of the new dual TOR architecture enabled by the Credo AEC.
We expect the ramp to continue into calendar 2023. We're developing multiple AEC solutions to solve for their future roadmap of higher speed deployments. I'm happy to report that Credo has completed the stringent qualification with our second hyperscale customer. We expect to begin our revenue ramp near the end of this fiscal year and expect meaningful contribution in fiscal 2024. In addition, we are engaged in developing advanced AEC solutions with this customer for their next generation server rack and switch rack applications as they move to 100G single lane speeds. We have also further broadened our traction in the market. We're currently in qualification with a third hyperscale customer for a 400G port switch rack application. Yet with another hyperscale customer, we're engaged in developing AECs for two future architectural deployments.
Although hyperscalers are clearly our primary focus, we've sold our AECs to dozens of customers, including data centers, 5G carriers, networking OEMs, and ODMs, as well as others in the Ethernet ecosystem. Finally, Credo is very proud to have introduced the industry's first 1.6 terabit per second connectivity solution at OCP, which will be a critical enabler for the 51 terabit per second switch generation. This reinforces Credo as the leader in the AEC market. Now moving to our optical solutions category. As we do across all our product solutions, Credo focuses on delivering disruptive solutions that are optimized for speed, reach, power, and cost.
Our optical solutions include DSPs, laser drivers, and TIAs found in both optical modules and AOCs, and span the breadth of applications with 50 and 100 gigabits per second single lane speeds, including 50G , 64G , 100G , 200G , and 400G modules. In the October quarter, we announced several new 100 Gbps per lane products, including our Dove 800 and 400 optical DSPs with integrated drivers. They've been met with great customer enthusiasm. In addition to engaging the optical module customer base directly, our go-to-market strategy has grown to include the joint development model or JDM. That is, focusing on the end customers of our optical module manufacturing partners. These include data center, 5G, PON, and Fibre Channel end customers as a means to have the end customer pull Credo through to design wins by specifying our solution.
To date, we've been successful with JDM engagements with 2 hyperscalers and a tier 1 OEM. We are now actively engaging all data center customers directly and teaming with optical module partners to jointly pursue our end customers. I'm pleased with our progress as we now have line of sight on new engagements with several data center and 5G customers across a wide range of applications, including 200G , 400G , 800G , and 50G solutions. I'll note that we see the process from initial win to qualification to volume ramp as taking longer in the current environment than we had anticipated a year ago. With that, our ramp to material revenue shifted somewhat, but our opportunity remains the same.
We continue to play the role of disruptor in the optical market, we will gain share over time given the distinct efficiencies we deliver in the combination of performance, power, and cost. Based on our increasing customer traction, we look forward to announcing meaningful customer wins and growth in our optical business. Regarding our line card PHY solutions , Credo has established leadership for Ethernet line card PHY solutions at 50 gig and 100 Gbps per lane speeds. This includes MACsec PHYs for high security applications needing encryption, as well as retimer and gearbox solutions. Our customers, again, include leading hyperscalers and networking OEMs and ODMs. As single lane speeds increase to 100 gig, the demand for lane card PHYs increases due to the signal integrity challenges that come with higher speed copper connections.
We also see a trend toward greater demand for encryption, driving increased demand for MACsec PHYs. A highlight from the OCP show in October was Meta showing the use of our Osprey 800 MACsec PHY in one of their critical deployments. In October, we announced our Screaming Eagle 100 Gbps per lane solution, a long-reach DSP retimer device with 1.6 terabits per second of bandwidth. This product has received great market reception due to its combination of performance and power efficiency. We have sampled it to many leading OEMs and ODMs and are already kicking off design engagements. Based on our current market position, product positioning, and customer engagements, we expect solid growth and continued share gain in the market. Finally, I'd like to give an update on our SerDes IP licensing and SerDes chiplet business.
We've received very positive feedback from customers on our 5- and 4-nanometer 112G SerDes IP announcement. It confirms that Credo's solution offers a 40%-50% power advantage over our competition, depending on the reach required in the application. This highlights that Credo has extended what we refer to as our N-1 process advantage, which means to compete with the power efficiency of Credo's 5 and 4 nanometer solutions, our competitors will need to move to 3 nanometer. As every industry seeks to lower its carbon footprint, Credo's core SerDes technology is delivering on the need to lower electricity use. We're also an early leader in the chiplet market and are in production with multiple customers. Notably, Tesla selected Credo's SerDes IP and chiplets for their Dojo supercomputer program.
Going forward, we're excited about the prospects for chiplets in light of the UCIe consortium. This Intel-led consortium, where we're a contributing member, is coalescing to standardize the broad use of chiplets inside servers. In summary, we remain highly encouraged about Credo's prospects due to our current solutions and production, near and midterm opportunities we're deeply engaged in, and longer-term opportunities in emerging markets. Today, we remain focused on delivering strong execution in our fiscal 23, and we continue to expect to achieve at least $200 million in revenue, representing more than 88% growth compared to fiscal 22. I'll now turn the call over to our CFO, Dan Fleming, to provide more details on our second quarter and to give guidance on Q3.
Thank you, Bill. Good afternoon. I will first review our Q2 fiscal 23 results and then discuss our outlook for Q3 of fiscal 23. As a reminder, the following financials will be discussed on a non-GAAP basis unless otherwise noted. I'm pleased to share with you that in Q2, we achieved another quarter of record revenue at $51.4 million, up 11% sequentially and up 94% year-over-year. Sequential growth was driven by strong revenue growth of our products, which also reached a record of $48.1 million for the quarter, up 33% sequentially and up 143% year-over-year. This growth in product revenue was led by a continued wave of AEC adoption.
The fundamental driver of our product growth, a strong HSDC expansion outlook at the highest speeds, remains in place in the face of an uncertain economic and geopolitical landscape. Our IP business generated $3.3 million of revenue in Q2. 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 Q2 was 6% IP, well below our long-term expectation for IP, which is 10%-15% of revenue, as the timing of IP deliverables, and therefore IP revenue recognition, shifted during the quarter. We continue to expect IP as a % of revenue to come in above our long-term expectations for the fiscal year.
Due to the revenue mix between product and IP this quarter, our gross margin came in at 54.9%, below our guidance range. More importantly, our product gross margin was 52.6% in the quarter, up 80 basis points sequentially and up 4.8 percentage points year-over-year. This product margin expansion is principally due to leverage from our strong product growth. Total operating expenses in the second quarter were $25 million, within our guidance range, and up 37% year-over-year as we scaled the organization for growth. I think it's important to note that this is considerably below our 94% year-over-year revenue growth. We generally expect that our top line will grow at least twice as fast as our OpEx for the foreseeable future.
With this, we expect to continue to deliver considerable leverage in the business. Our OpEx increase was driven by a 38% year-over-year increase in R&D as we continue to invest in the resources to deliver innovative solutions. Our SG&A was up 34% year-over-year as we continue to build out public company infrastructure. We delivered operating income of $3.4 million in Q2, an improvement of $5.6 million year-over-year, but down 41% sequentially. Our operating margin was 6.6% in the quarter, an improvement of 15.1 percentage points year-over-year, but down 561 basis points sequentially due to revenue mix that resulted in lower gross profit.
We delivered net income of $2.4 million in Q2, an increase of $5.7 million year-over-year and down 55% sequentially. Cash flow from operations in the second quarter was $1.8 million, an increase of $27.7 million year-over-year, and an increase of $14 million sequentially. CapEx was $5.7 million in the quarter, driven by production mask spending. Free cash flow was -$3.9 million, an increase of $25.7 million year-over-year, and an increase of $13.6 million sequentially. We ended the quarter with cash and equivalents of $240.5 million, a decrease of $3.2 million from the first quarter.
This decrease in cash was a result of continued working capital investments to support our top-line revenue growth. Our accounts receivable balance decreased 5.5% sequentially to $51.8 million, while days sales outstanding decreased to 92 days, down from 107 days in Q1. Our Q2 ending inventory was $47.8 million, up $10.8 million sequentially as we continue our product ramp. Turning to our guidance for the third quarter. We currently expect revenue in Q3 fiscal 2023 to be between $54 million and $56 million, up 7% sequentially at the midpoint and 73% year-over-year. We expect Q3 gross margin to be within a range of 59%-61%.
We expect Q3 operating expenses to be between $25 million and $27 million. Finally, we expect Q3 weighted average diluted share count to be approximately 160 million shares. With that, I will open it up for questions.
Thank you. As a reminder, to ask a question, you will need to press star one one on your telephone. Please stand by while we compile the Q&A roster. Our first question comes from the line of Quinn Bolton with Needham & Company.
Hey, guys. Congratulations on the strong product results. Bill, I guess I wanted to start on the AEC business. I have several questions there. It sounds like you continue to expand your hyperscale relationships, but wondering if you can sort of talk about the growth and I think you mentioned a couple of these guys looking at switch applications rather than server applications and was hoping you could expand on that thought. You know, is this the distributed disaggregated switch chassis applications for the third and the fourth hyperscaler as well as I think you mentioned the switch application for your second hyperscaler?
Yes. Let me say that we really have 2 opportunities for AECs within the switching hierarchy and the server racks as well. The largest opportunity that we see are server racks. The market forecasters show that, you know, that market in comparison to the disaggregated chassis or what we call switch racks market is probably a 5 to 1. Now, in the switching architecture, it's really a big choice between sticking with what has traditionally been, you know, most popular, which is a big chassis, you know, filled with switches connected internally over a back plane, to disaggregation, which means, you know, pulling those switches out of the chassis and basically stacking them vertically in a rack.
Those back plane connections become connections between the switches within the rack 2.5 meters or less. We, you know, fully expect that over time, each data center customer is gonna make their own decision about the architecture that they pursue. It's good. We're encouraged with the fact that, as switching lane speeds increase to 50 and 100 gig, that for those customers that are moving or have been, you know, using switch racks compared to chassis, that, our solution is naturally getting picked up.
Understood. The second question, just on the competitive landscape for the NIC to ToR application. I know the first couple of hyperscalers you're working with, I believe, are using proprietary or non-standard cables. Are you seeing, you know, any evidence that those customers are looking to second source those designs? Or do you feel pretty comfortable that Credo remains the sole source of those cables for the, you know, foreseeable future, which, you know, hopefully would extend that at least a few quarters?
I think that there's no question that the data center customers, generally speaking, want multiple sources. You know, that's absolutely part of the world that we live in. My feeling is that more and more as the data center customers recognize that, they can add functionality to an AEC that they haven't even, you know, been able to think about with DACs or optical solutions. You know, the idea of having this fabric become more intelligent, you know, it's natural that the team that we've built, which is, you know, a large team of engineers and different functions to be able to entertain, you know, specific requests for added functionality to the AEC. We're open for business.
If the engineers within the data centers that we're working with have ideas, you know, we'll entertain those ideas as a way to make, you know, their jobs easier, make the innovations better from their end, from a server rack or switch rack perspective. Competitively, you know, we've talked about the competitive moat that we've established and the fact that I've got the team that's doing the full system integration. It's not as if selling a chip to a copper cable manufacturer really opens that door for this kind of innovation. I think that yes, we naturally see competitors. We naturally see our customers wanting to second source. You know, I think that, you know, we still haven't seen competitors in the wild in the labs at our customers. We've heard about competitors' intention to enter the market.
We've seen static demos where they'll have a cable that's not hooked up to anything, and they will describe what the cable does. We've seen demos with eval boards that are connected to passive copper cables. These are far from actual qualification at a hyperscaler. You know, I think that we look at the OCP conference in October. It was a really great measure of our leadership, especially in contrast to the other competitors. I think that anybody that attended the conference could not avoid seeing purple, which is the color of our cables. We were, you know, the purple AECs were ubiquitous from end to end on the show floor.
And to note the progress that we've made year-over-year, last year, we showed the world's largest router to date, a 3-rack deployment, that made up a 350 terabits per second router. This is built with 400 gig ports, each with 8 lanes of 56 gig or 50 gig. This year, we showed the clear benefit of going faster on the lane speed and wider, you know, with the number of lanes. We introduced our 1.6 terabits per second AEC at the show, 16 lanes of 100 gig. We demonstrated that same capacity or that same bandwidth, the 350 terabits per second in a 1 rack deployment.
This is one of many demonstrations, you know, that we gave, and it was just clear to anybody at the show that, you know, just physically you can see it, we're far ahead of our competition.
Great. Thank you for all that color, Bill.
Thank you. Our next question comes from the line of Toshiya Hari with Goldman Sachs.
Hi, good afternoon. Thanks so much for taking my questions. First one for you, Bill. You know, based on, you know, the fact that you're reiterating the full year guide, you know, you guys are clearly doing really well. Curious if you've sensed any change in customer behavior, whether it be, you know, the large hyperscalers or some of the enterprise OEMs across your business. Have you seen any, you know, projects get pushed out or downsized? Is the adoption of AEC, for instance, too strategic and too important for your customers to really tweak projects even going into fiscal year 2024? Thanks.
Yeah, I think that what we see in our customer base is really, you know, two different threads. We have seen a reduction in CapEx and a delay within our Chinese hyperscale customers. There's no question about that. Over the past quarter, we've seen kind of a shift in ramp to high volume on the next generation optical. That's, you know, I think everybody's well aware of, you know, the macro situation in China. You know, these hyperscalers don't necessarily serve a wide customer base globally. The second thread is really the US hyperscalers. Although, you know, we've heard that there may be a slight bending of the curve, we haven't heard that anyone is gonna decrease spending year-over-year.
These are with the customers that we're engaged with that we've ramped to production and we will ramp production with. It might be a slowing of the growth, but the growth still is very significant. I will reiterate that for us it looks a little bit different because we haven't reached the point of saturation. Every new product ramp that we see is next generation, you know, better than the current generation technology. If anything, there's still a fierce competition amongst data centers to deliver better services to their customer base in order to win market share. We see that there is a very consistent pull for next generation better than technology to get better productivity, to get better performance.
You know, even in an environment like this at a macro level, that becomes critical to be able to differentiate. We still remain, you know, quite bullish on the customers that we're ramping. We haven't seen any major shifts.
That's great. Thank you. As my follow-up, on the second customer, in your AEC business, you talked about revenue recognition in toward the end of this fiscal year and then a meaningful ramp in fiscal year 2024. I understand you can't give too specific with these customers, but I was hoping you could compare and contrast the ramp that you're expecting in fiscal year 2024 with the second customer vis-a-vis what you've experienced so far with the first customer. I think you've talked about, you know, a potential uplift in pricing just given the complexity. If you can kinda level set us on your thoughts there into fiscal year 2024, that would be super helpful. Thank you.
As a contrast, I think we've been relatively pleased with what we've been seeing from the second customer in a sense that, you know, over a year ago we engaged with them. We delivered samples, full samples of this unique cable that we've built in December of last year. We've now, you know, finished the very stringent qualification. It's a full green light on ramping as soon as they're ready. Their schedule has been really consistent over the past several quarters. The contrast there is that, you know, with our first customer, our solution was enabling a new architecture, but it wasn't necessarily that the servers were changing.
They, you know, had a high volume stream of deployments and what they were trying to do is shift to an architecture, you know, that gave them, you know, much, much better utilization of floor space and much better utilization of equipment. It wasn't like it was the next generation server. We kind of naturally saw the customer kind of dual pathing it and trying to, you know, trying to cut it over, you know, in an orderly fashion. It was a little bit delayed compared to what their first objectives were. That's kind of the big contrast is that the second customer, this is a brand new generation, and there's extremely strong pull to deploy and deploy on time. That's, you know, something that we're pretty encouraged about.
Our next question comes from the line of Vivek Arya with Bank of America.
Thanks for taking my questions. I actually had 2. For the first one, Bill, you mentioned the engagements for the third and fourth hyperscaler also. That seems like a positive. I'm curious, you know, the adoption of AEC seems like a no-brainer on surface. What is the main pushback that you get? Is it just a matter of time? Is it the ramp of a certain, you know, speed? Is it, you know, 400 gig or 800 gig? What do you think drives that sharp inflection upwards with multiple customers?
I think there's really, you know, 2 catalysts to cause people to think about the AEC solution. 1 catalyst that we've seen first is added functionality. Yeah, that's a real differentiation, and that's why you've seen our first customer ramp with 25 gig lanes, which you know, clearly, if you weren't doing something special, you could get the job done with DACs. Our second customer is a combination of functionality as well as speed. As the world goes to, you know, 50 gig lanes for a bulk of, you know, for a large number of customers that we're talking to, they don't wanna fight the signal integrity and form factor challenges of staying with a DAC. Speed is the other catalyst.
Where we see the 50 gig per lane market being in kind of a crossover generation, where some will battle, you know, deployment with DACs. Others, you know, it becomes a very, almost a default decision at this point because they see a solution that's half the power, half the cost, way more reliable, way more rugged than an AOC. For short in-rack connections, we don't really see, you know, a big decision-making that has to occur. If they're not gonna fight, you know, the challenges of DACs, they're gonna use AECs, and we've seen that across the board. As we progress towards 100 gig, you know, which is, it's just a function of time, for sure there is no DACs, and now it becomes just a question about, you know, AEC versus AOCs.
I think we've established that, you know, that game is over. You know, people will choose AECs just because of the, you know, the huge CapEx and OpEx advantage, right? If you look at the total cost of ownership, it's hands down a better idea. With the fact that you throw, you know, the fact that, you know, the installers aren't gonna break the cables routing you know, huge number of them in a very tight space, so it's a much more rugged design.
Got it. For my follow-up, maybe one for you, Dan. You know, in your Q2, seems like IP sales were, you know, $5 million-$7 million, kind of below expectations, but you more than made up for it, because of the upside on the product side. I'm curious, what is the expectation for this IP in Q3? Do you expect to make up for that $5 million-$7 million shortfall that you had in Q2? Because I think you mentioned something about a shift. Is the Q3 just that, you know, the missing part of the IP revenue from Q2 that comes into Q3? Just how should we look at Q3 and what happened to that missing IP revenue from Q2? Thank you.
Yeah, thanks. The important thing to note here is that there's no change in our expectation for IP revenue for the full year. In other words, you know, our revenue mix for fiscal year 23 is exactly what we have expected it to be, for the, you know, for the year plus that we've been talking. Also bear in mind that for the full year, we expect IP revenue to be above our long-term target, which is 10%-15% of the overall revenue mix. We've talked a lot about historically, about the quarter-to-quarter variability in revenue, when it comes to IP.
This is largely driven by ASC 606 and the way the revenue recognition rules work around license revenue, where we recognize the lion's share of the contract value on most of our contracts at the point of delivery of that IP database. The last two quarters, Q4 of fiscal 2022 and Q1 of fiscal 2023, we happen to be on the higher end of revenue contribution from IP. That of course swings both ways. One of the things that we track critically from a gross margin perspective, of course, is our product gross margin. That has continued to expand as we've increased our product shipment volumes. You know, in fact, it was up 80 basis points, the product gross margin.
We're quite pleased with our margins for the quarter, and we're exactly where we expect to be for the full year. Hopefully, that helps.
Thank you, guys.
Our next question comes from the line of Richard Shannon with Craig-Hallum.
Well, hi guys. Thanks for taking my questions. Dan, I guess I wanted to follow up on the topic of product gross margins. If I'm running my numbers right here, and if I exclude the product NRE from the calculation, it looks like product margins were actually down very slightly. Am I calculating this right? If so, can you help us understand the dynamic that took that down slightly?
Yeah. I wouldn't read too much into that. You know, we look at product gross margin a little bit more holistically. If you look at the elements that come into the other cost of goods sold bucket, that, similar to our IP revenue, can vary quite a bit quarter over quarter. Our view is that, you know, from an overall product gross margin perspective, as our volume continues to increase, at this stage where we are as a company, we should have that, a slight uptick in our product gross margin. But you're right that, you know, excluding NRE in this particular quarter, it did tick down a little bit, if you just look at the product margin.
As we go forward, especially if, and I think most people are assuming that your AEC mix is going to increase here. Should we expect those product growth margins, excluding NRE again, to grow if very slightly? Is that fair?
Yeah, that's fair. You know, our long-term model remains the same. Just to reiterate what that is, you know, 10%-15% of our overall revenue mix will be IP, the remainder being product. From a gross margin perspective, 63%-65% gross margin in that long-term model. So, what is long-term? We really view that as a 3-plus year model. We've stated in the past that this fiscal year, we expect the gross margins to expand purely out of increasing scale. There are subsequent factors in FY 2024 and 2025 that will increase the product margin as well. The one notable difference, if you kind of read into some of Bill's comments for the year, you know, AEC has been ramping faster than we initially expected.
If you look out a year from now in our FY 2024, that has an implicit margin impact. With optical taking a little bit longer to ramp than we initially expected, again, somewhat of an impact in FY 2024. Overall, you know, our corporate gross margin in FY 2024 is probably going to be similar to what it is for the full year of fiscal 2023.
Okay. Perfect. Thanks for that detail, Dan. Bill, maybe a big picture question for you. Obviously, Ethernet is your dominant protocol standard you're supporting and obviously a lot of growth opportunities there. You've talked about USB and PCI Express in the past, and I think you even alluded here recently to, cable opportunities, AEC opportunities that exist with one or both of those. Maybe you can just kind of give us a big picture on your thought process on when those technologies and products start to contribute more meaningfully to your outlook.
For PCIe, our intent is to really enter the market in a big way when the market moves to Gen 6. Of course, we'll build a product that is, you know, compatible with earlier generations, and we expect to get traction, you know, earlier, to get cycles prior to Gen 6. That's really, you know, what we expect as the point when we're gonna enter the market in a big way. We see that being in the 2025, 2026 timeframe. You know, it's really dependent on the schedules, you know, the server schedules and hopefully things get back on track and the world, you know, goes faster.
I'll say that, you know, our view of the overall market opportunity for PCIe, and it can be measured, you know, from a PCIe retimer within the server and also within the UCIe chiplet. We view this as a very large market opportunity. On the USB front, you know, it's probably the same kind of timeframe that we see the USB4 80Gbps or two lanes of 40 gig PAM4. It's probably the same timeframe that we see that opportunity. As it relates to the AEC opportunity, we definitely see opportunity within both segments or both standards. And it's a very natural extension for us, you know, to look at that opportunity, the same way that we look at AECs for Ethernet.
For USB, I'll tell you that, you know, the consumer market will probably not be the one, you know, the company, you know, building cables. We'll probably go straight to a reference design model.
Thank you. Our next question comes from the line of Roth Capital Partners.
Hi, Bill. Hi, Dan. Question specifically on the revenue breakout, perhaps for Dan. The product engineering services would, I know it's a smaller part of the revenue, but would growth in that revenue be a lead indicator of activity you have with hyperscalers for AEC? Are those kind of indicators, you know, as you go from 1 to 2 to 3 to 4 hyperscalers, would that grow? Is that the way to read that line?
I would not read it that way. It's. You know, we've from day one as a company, we've been able to capture some NRE dollars from customers as we've developed chips and solutions for them. It really speaks to the innovative nature of our solutions. You know, longer term, we don't expect that to grow necessarily in absolute dollars. Just like our IP revenue, it can vary quite a bit quarter-over-quarter. It's not. I wouldn't really read it as a lead indicator for anything such as that.
Okay. Thanks, Dan. The second question, perhaps for Bill. You talked about the TAM for AEC being $4 billion-$5 billion. I think I heard the number correct, if I heard it correctly. Is that the vast majority of that hyperscaler or is there a meaningful non-hyperscale part that could kick in as you kind of evolve your offering beyond these initial customers?
Yeah. Just to clarify, I referenced a multi-billion dollar market that the market forecasters are forecasting, and it's really, you know, in the 4-5 year timeframe.
Oh, 4 or 5 year. Okay.
Yeah. I definitely see the hyperscale market as the market that's gonna drive the near term growth. I can say that, you know, as, you know, as the enterprise moves to higher speed, there's gonna be an opportunity there. I would say that, you know, even markets that are outside of, you know, what we consider hyperscalers, I think there can be, you know, significant contribution from a revenue standpoint. We're already engaged with the first 5G carriers. It's not the same, you know, order of magnitude as hyperscalers, but if we look at, you know, the different engagements that we've kind of quickly converted into customers, I think collectively, you know, they can, they can look like one of the major hyperscalers in the total size of revenue for us.
I don't think there's... You know, in the near term, I guess, you know, our very, very primary focus is on the hyperscalers to drive the revenue quickly. We are engaging across the board with many others that, you know, again, collectively can add significant revenue for us.
Okay. Thanks, Bill.
Thank you. Our next question comes from the line of Vijay Rakesh with Mizuho.
Yeah. Hi, Bill. Then just a quick question on the quarter. I know, the IP came in light, but looks like you made up well with the product side. Just wondering where the strength was. Was it in cable or optical? If you can give some color. Was it specific to some customers or markets?
it was strength in AEC, as you would expect.
Got it.
That has been with our lead customer that we've discussed in the past.
Got it. On the JDM side, the joint development program, do you expect that to become a bigger mix of your distribution as you look at calendar 2023 or fiscal 2024? Would those have similar margins to your direct sales?
Oh, definitely. When we talk about a joint development model, what we're really referring to is that the hyperscaler would be involved in the selection of the DSP or other components. Typically, even under a JDM model, we would be selling to the optical module manufacturer. I do expect that, you know, this JDM model or if we kind of back up and we say the model where the hyperscaler gets involved in the decision making, I expect that to be more and more popular as we go forward for Credo. You know, if we kind of look at it, you know, from a Credo development perspective in the market, how have we been progressing?
Really, our view is that, you know, first we succeeded in engaging, you know, 3 JDM customers where the end customer really selected the DSP component from Credo. The ramp to high volume looks delayed due to, you know, the first 2 hyperscalers being in China. Kind of in the second, you know, phase here, we've gained traction among tier ones. We're talking with all of them directly on high volume deployments. We see multiple programs in sight on 200 and 400. You know, the major benefit that we give is a refresh that's got a better combination of performance, power and cost. It's becoming more important to the hyperscalers and the optical companies running high volumes.
CapEx and OpEx are more and more in focus recently. If there's, you know, low-hanging fruit, it seems like it makes sense. You know, I might have mentioned on the last call, but I feel even more confident that we're going to officially engage with a tier 1 hyperscaler in the U.S. on a 400 gig optical solution, and it will include the DSP with an integrated driver as well as the TIA. That's more than kind of line of sight. That's, you know, basically, a few phases left to entering a contract with them. I would say the, kinda the third phase of this is for the next generation or 800 gig.
Now that we've opened up conversations about their existing high volume deployments, the same kind of compelling performance, power and cost benefit that we offer, added to that will be the fact that we're, you know, on time. Our time to market is good for the 800 gig devices and for the 800 gig market. The, you know, the first testing that's been done by the customer base has been very well received. They see, you know, very clearly the performance is, is clearly good enough, the power is clearly good enough, and the cost is compelling in a sense because we're building in 12 nanometer versus a more advanced process like 7 or 5.
Given the fact that we've established, you know, market credibility, you know, through our first three engagements and the engagements that we're pursuing on existing, you know, programs, I expect the wins for 800 gig to come as the market takes off.
t question. I know China is again going back into COVID restrictions and all that, and it looks like you have successfully resolved many of your supply constraints, it feels like, because it didn't really come up on the commentary. Can you talk to what you are doing in terms of maybe diversifying your supply chain? I think you had talked about maybe Vietnam or Philippines, or what you are doing there in terms of getting around this whole restriction
I feel good about where we are today, even if we face disruptions in China. We've signaled that we're gonna build inventory. That's a surefire way of making sure that we've got the product that our customers need as they ramp. We're gonna continue to, you know, be in that mode until we can land ourselves in, you know, in a location that's not dependent on China. We've made progress in the last 90 days, and my expectation is that, you know, we'll be in production in less than 1 year in alternative locations for our current supply chain. I think that kind of matches with what the customer base is looking for as well.
Got it. Great. Thanks a lot.
Thank you. Our next question comes from the line of Tore Svanberg with Stifel.
Yes, thank you, Bill. Thank you, Dan. Congrats on the record results. I have a non-AEC related question. You talked about the UCI Express opportunity, and I do recognize that this is kind of further out, but as we think about, you know, chiplets and licensing, how should we think about this playing out for you over the next few years?
I think that, first of all, we kind of classify chiplets as a product because, you know, we're building and selling those. As it relates to, you know, the IP, I think our long-term guidance kinda, you know, fits in the 10%-15% range that Dan has articulated. For the UCIe chiplets, we see this. I mean, you know, we were very early on in chiplets. We were, you know, by far ahead of the rest of the competition, given that we kinda saw this move towards disaggregation early. It didn't really play out. Now that we see UCIe on the horizon, we're very big believers.
You know, if you look at the, you know, the consortium that Intel has really driven, you look at what they're doing today, chiplets are, you know, gonna be popular and high volume in, I think, the relatively, you know, near term. As speeds move to, you know, to 32 gig to 64 gig, this is gonna become more and more popular. I think that if you look within a server, you can see, you know, 8 to 16 chiplets per server in the future. With that kind of volume, it becomes, you know, you can do the math, it's gonna be a very large revenue opportunity.
That's very exciting. To follow up on AEC and the sort of the second hyperscaler that's expected to launch, sounds like you have a lot of confidence in the timing there. I was just hoping you could give us a little bit more, you know, background information. I mean, obviously, you know, when we think about the hyperscalers cutting CapEx, you know, I'm sure there's some priority CapEx and some not so priority CapEx. I mean, would this fall into like, you know, the really highest end priority for that particular customer? Is that how we should think about it?
That's the way that I think about it, for sure. I view this program as, you know, something that is a strategic imperative as they, you know, talk about it, you know, within the supply chain as well as within our customer base. I feel pretty confident that this is, you know, gonna be mainstream. Although they have been pretty consistent with their schedule, you know, of course, things can happen that, you know, might cause, you know, might cause a delay. We're not seeing any signs of that. The closer we get, obviously the more confidence we have in the ramp. It's not as if I can make a, you know, firm commitment on exactly when they're gonna go to production. It feels like no change from our end.
That's very fair. Last question for Dan. Dan, you said in the last quarter you would build inventory. You did that. The inventory days now are just over 100. You know, based on Bill's comment about sort of, you know, keep building inventory until you feel better about, you know, the whole China lockdown situation, how high should we expect inventory days to potentially get to, you know, before you feel like you got the situation under control?
Well, I don't expect that the days will increase from where they are right now, or if they do, it would be a very small increase. We were essentially flat quarter-over-quarter. Even though it looked like, you know, a significant increase in inventory, it was actually kind of in line with our product growth, quarter-over-quarter. Bearing that in mind, we're comfortable with where we are. We continue to build, you know, an excess amount of days of inventory, of cable inventory to ensure that we don't run into any situation that we had back in Q4, when that COVID lockdown kind of interrupted our, you know, fulfillment of demand.
We're comfortable with where we are right now. I wouldn't expect any major deviation from a days of inventory perspective. Longer term, of course, that'll settle down to maybe 100 days of inventory. That's, you know, off in the future.
Okay. Sounds good. Congrats again. Thank you.
Thanks.
Thank you. Our next question comes from the line of Matt Ramsay with TD Cowen.
Hi, this is Lanny on for Matt Ramsay. I wanted to extend my congratulations for the quarter as well. Going back to the optical solutions, could you confirm that the push-out was mostly due to your first two customers being Chinese hyperscalers? Is there any line of sight as to how long that program is and where you are in terms of the qualification to product volume ramp?
Yeah, I will confirm that the first two hyperscaler customers are Chinese data centers. During the last quarter, you know, we, you know, we were told that there's going to be a shift in the timing of the deployment. I haven't got clarity on specifics on when that's gonna ramp and, I mean, from my perspective, it's really when it goes to high volume. We might be doing small volume, but, you know, our focus is really on, you know, how long is the shift to high volume. I will say in the, you know, in the efforts that we're making with the U.S. hyperscalers, we continue to make progress. If anything, I think we've got more clarity on that, even though it might be, you know, a bit further.
I think that's gonna drive higher volume as well.
Understood. That's helpful. Thank you. I know in past earnings calls you've mentioned a consumer customer for USB-C, I believe, for licensing. Any updates there that we should know about in terms of progress?
Yes, we're hip deep in executing on that IP license. We were selected by this large consumer company as the partner for this really important next generation USB standard, which is 80 gigabits total bandwidth, 2 lanes of 40 gig PAM4 modulation . I think it's reflective of the fact that our architecture is unique. We deliver lower power than anybody in the industry, so it's really a great confirmation that they would select us as their partner. This goes back to, you know, even 4 years ago that they were doing due diligence, and I think we signed the contract finally a little more than 1 year ago.
Yeah, we're hip deep on execution, and we expect to, you know, be wrapping up the, you know, the technical part of the work, really within the next six months. Then we'll absolutely be there as they move from their own samples to production.
Got it. Thank you so much.
Thank you. Our next question comes from the line of Quinn Bolton with Needham & Company.
Hi. Thanks for taking it. A quick couple of follow-ups first. Dan, could you give us a sense of your lead customer, what percent of revenue, it was in the quarter? I know they've been above 10% for the past several quarters. Then a follow-up for Bill. As you look to the PCIe market, wondering if that also includes opportunities, in CXL, since CXL does run on the PCIe electricals. Thanks.
Yeah. Quinn, so, what you'll see in our Q as we file it in a day or two, is that we had 3 10% customers in the quarter, the largest of which was 44%. There was a 19% and a 16% customer. You can kind of fill in the blank from there. You know, we don't disclose the specifics of who those customers are, but in the 44% case, it's pretty obvious.
Yeah. Okay.
And to answer-
Thanks.
To answer your question on PCIe, we definitely include CXL. We kind of talk about that collectively.
Perfect. Okay. Thank you.
Thank you. There are no further questions at this time. Mr. Brennan, I'll turn the call back over to you.
I'd like to thank everybody for joining the call. I appreciate all the thoughtful questions. With that, we will end the call. Thank you very much.
This concludes today's conference call. You may now disconnect.