Good afternoon, everyone, and welcome to AXT's First Quarter 2024 Earnings Call. Leading the call today is Dr. Morris Young, Chief Executive Officer; and Gary Fischer, Chief Financial Officer. My name is John, and I will be your coordinator for today. [Operator Instructions]. I would now like to turn the call over to Leslie Green, Head of Investor Relations for AXT.

Thank you, John, and good afternoon, everyone. Before we begin, I would like to remind you that during the course of this conference call, including comments made in response to your questions, we will provide projections or make other forward-looking statements regarding, among other things, the future financial performance of the company; market conditions and trends, including expected growth in the markets we serve; emerging applications using chips or devices fabricated on our substrates; our product mix; our ability to increase quarter -- orders in succeeding quarters to control costs and expenses, to improve manufacturing yields and efficiencies, to utilize our manufacturing capacity; the growing environmental, health and safety and chemical industry regulations in China as well as global economic and political conditions, including trade tariffs and restrictions. We wish to caution you that such statements deal with future events are based on management's current expectations and are subject to risks and uncertainties that could cause actual results or events to differ materially. These uncertainties and risks include, but are not limited to, overall conditions in the markets in which the company competes, global financial conditions and uncertainties, and COVID-19 and other outbreaks of contagious disease, potential tariffs and trade restrictions, increased environmental regulations in China, the financial performance of our partially owned supply chain companies and the impact of delays by our customers on the timing of sales and their products. In addition to these factors that may be discussed on this call, we refer you to the company's periodic reports filed with the Securities and Exchange Commission. These are available online by link from our website and contain additional information on risk factors that could cause actual results to differ materially from our current expectations. This conference call will be available on our website at axt.com through May 2, 2025. Also before we begin, I want to note that shortly following the close of market today, we issued a press release reporting financial results for the first quarter of 2024. This information is available on the Investor Relations portion of our website at axt.com. I would now like to turn the call over to Gary Fischer for a review of our first quarter 2024 results. Gary?

Thank you, Leslie, and good afternoon to everyone. Revenue for the first quarter of 2024 was $22.7 million. That's up from $20.4 million in the fourth quarter of 2023 and up from $19.4 million in the first quarter of 2023. To break down our Q1 204 revenue for you by product category, Indium phosphide increased sequentially to $8.1 million. That's reflecting strong growth from data center applications, including AI and continued improvement in passive optical networks, Gallium our side also grew to $7.5 million with broad-based improvement across a number of applications. Germanium substrates were $1.4 million, up from the prior quarter with renewed strength in demand for satellite solar cells. Finally, as expected, revenue from our consolidated raw material joint venture companies in Q1 was $5.8 million, down from Q4 as we consumed a greater portion of their output for our growing substrate demand. In the first quarter of 2024, revenue from Asia Pacific was 79%, Europe was 16%, and North America was 5%. The top 5 customers generated approximately 33% of total revenue and 1 customer was over the 10% level. Non-GAAP gross margin in the first quarter was 27.3% compared with 23.2% in Q4 and 26.9% in Q1 of 2023. For those who prefer to track results on a GAAP basis, gross margin in the first quarter was 26.9% compared with 22.6% in Q4 and 26.3% in Q1. Beyond the near term, we remain confident that we can get back to the mid-30% range as the environment strengthens through higher overall volume, favorable product mix and the benefits of our recycling programs, along with continued efficiency improvements throughout our business. Moving to operating expense. Total non-GAAP operating expense in Q1 was $8.7 million compared with $7.5 million in Q4 of 2023 and $8.7 million in Q1 of…

Thank you, Gary. We believe our business is on a firm path to recovery, as evidenced by the continued growth in our business and solid execution, which allowed us to exceed our Q1 revenue and profitability pace. In the first quarter, we achieved 11% sequential growth in our revenues and 29% sequential improvement in our non-GAAP net income. While certain parts of the demand environment remains somewhat soft, all 3 of our substrate product line showed improvement, including a 48% growth in time phosphide revenue from Q4. Looking individually at these product lines. Indium phosphide once again became our biggest selling material in Q1. Sales was driven by continued recovery in the power market and a meaningful increase in demand related to AI. We review AI as an exciting catalyst for Indium phosphide in the years to come. As AI requires high-speed lasers and detectors for rapid data transfer with increased bandwidth, low attenuation and low distortion. Today, AI applications are primarily using gallium as [indiscernible] which require a relatively small amount of substrate material. But as the industry moves to 800 gig and then 1.6 terabyte speed, we expect that there will be a necessary transition to indium phosphide. We are already seeing development work happening today with next-generation silicon photonics devices and electro-absorption modulated laser or EML for high-speed data center transceivers. Revenue from these applications contributed to our strong indium phosphide growth in Q1 and will help drive our expected growth in 2024. Our gallium phosphide revenue grew 24% sequentially in Q1, reflecting increased demand across a broad base of applications, including power amplifiers, HPT applications for wireless our switches, high-power lasers and LEDs. We believe much of the excess inventory in the supply chain has been decided, and we are benefiting from a desire among our…

Thank you, Morris. In keeping with our comments today, we expect Q2 revenue to be between $25.5 million and $27.5 million. We expect our non-GAAP net loss will be in the range of $0.03 to $0.05, and GAAP net loss will be in the range of $0.05 to $0.07. Share count will be approximately 43.0 million shares. Okay. Well, this concludes our prepared comments. Morris and I will be glad to answer your questions now. John, can you take it back?

[Operator Instructions]. The first question comes from the line of Charles Shi from Needham.

Morris, Gary, congrats on the good Q1 results and very happy about Q2 guidance. I want to start with the AI really related question. The indium phosphide, I think that last time when we speak -- when we spoke about this, you were talking about semi-insulating indium phosphide wafers potentially being used for some of the high-speed detectors application. But in your prepared remarks, it sounds like you are getting a little bit more upbeat about the high-speed laser type of applications for 800 gig, 1.6 tera those kind of applications. For the next-generation [indiscernible] et cetera. So can you provide us a little bit of update your engagement on these AI applications so far? Do you -- when do you see the laser application will start to contribute some meaningful revenues for the indium phosphide business?

Charles, first of all, the order we received one customer specifically told us is for us for AI, and that was from selling germanium phosphide. And the fact that we're guessing if it was for detectors, it was a guess, okay? Because usually, indium phosphide good semi-insulating either for electronic applications or for high-speed detectors. However, recently in the industry, we also heard some of the EMLs require semi-insulating substrates. So this -- whether it's for lasers or for detectors, we all know from what we are making, the customer demand for our product is sort of a slight change from silicon do -- I mean sulfate semiconducting substrates, usually good for lasers. And now it's for semi-insulating in of the indium phosphide, but whether it's detectors or lasers, it's a an insulate estimate. We don't really know. I'm sorry. But we do hear from our customer, it's for a pretty large AI customer.

Got it. Well, Morris, thanks for the color you provide. I guess it's something we will continue to chat. And obviously, we don't want to dig more if there's something about your customer you don't want to share. So maybe a second question, maybe for Gary, I think in your prepared remarks, you talked a little bit more about the SAR listing. It sounds like you're making some progress on that front. Can you kind of provide a little bit more color in terms of what kind of progress exactly? And give us a little bit better sense about why you think it's moving forward a little bit more at this time.

So the process of communication with Shanghai Stock Exchange a CSRC is such that once we send in our applications, they will continue to ask questions. They want us to clarify a few things. And there are things which that they express, they have concerns with and we provide them with answers. And we think we made some good progress on the questions they have. And so now we think we are ready to go through the next step, which is go through their review process again, and they will then look at our whole application and either continue to ask questions or Improv... Approve...

Congrats on the results and guide.

The next question comes from the line of Richard Shannon from Craig-Hallum.

Congratulations on a really nice guide here. Great to see. To that point here on the guidance here, I wonder if you could elaborate on a couple of items here, first of all, just trying to understand the relative growth dynamics of each of your 4 revenue buckets here, I would assume you deposits probably your best growth driver, but wondering if you can kind of rank those and whether you expect them all to be growing sequentially. And then second of all, trying to get a sense of what's implied here for gross margins, given that you said that OpEx could generally be the same in the second quarter as the first. I'm kind of getting a number that's slightly higher than the first quarter, maybe as high as $30 or 30%, but just wondering if I'm doing the right math on that one.

Yes. So let me first try on the business conditions. Actually, the strongest growth we think next quarter is going to come from Gallium phosphide and also germanium. Germanium seems to be very strong and raw material is very strong as well. Indium phosphide actually at this point of the timing that we are projecting flat. And one other thing which is interesting about this business environment right now is that wherever the customer wants something, it's the short -- the lead time is extremely short. I mean, we usually ask the customer to give us at least 4 weeks lead time because we need to process the wafers and sometimes we even have to grow crystal. So the lead time issue is long. We don't have a standard product for offering. But nowadays, I mean, customers just want 2 weeks lead time, which is great. However, it's all of prevented from having better visibility. But we can see that gallium phosphide is extremely busy. And indium phosphide right now, as we see now is a flat quarter, but things can change because we still got almost 2 months to go, given 2 weeks lead time. We have 6 weeks to take orders and deliver. And on the margin, Gary?

Maybe we can do this on a follow-up call, Richard, because it gets kind of complicated and I didn't understand exactly what your model is telling you, but we can work on it with you, so.

Okay. Fair enough. Morris, I did want to follow up on your comment regarding gallium phosphide being your highest growth driver here. To what degree is this maybe some inventory fill in some product areas that were driven down so hard in the last 1.5 years? And then I think you also mentioned some pickup in the wireless/HPT market. So I'm wondering to what degree that's contributing in this quarter.

Well, first of all, I would say gallium phosphide was the first product coming down in terms of revenue, right? I mean, I remember it was the second quarter of 2022, gallium phosphide start to come down. So it takes usually 6 weeks before all the inventory got depleted. So now it's coming back. So because indium phosphide takes, I think, 3 longer weeks before it start to come down. 3, yes, 3 quarters, too. So I expect indium phosphide perhaps not to recover as quickly as gallium phosphide. Although the pickup of indium phosphide last quarter, I attribute that to the new business on AI and because I see PAM business is doing okay but not really robust. The telecom business is not great. And data center, from what I see, they are still inventory out there. So I think the AI part of the data center, I think, is pulling the demand for us and increase our revenue by, what, 40-some-odd percent 48% from Q4 to Q1. So the... Yes.

I didn't mean to interrupt Morris, so please continue.

No, I actually have forgotten [indiscernible]. So what was another part of your question? Yes.

Yes, I think so. But there is an interesting follow-up to those comments here, which is like how much of your Indium phosphide businesses is AI today? It seems like given these kind of growth rates, it's got to be a pretty substantial part now. Can you quantify that in any way?

Well, I think last quarter was significant. I would say, could be as much as 20%, although it's a little bit difficult for us to really nail it because some customers would tell us and some customers don't, okay? But as I said, indium phosphide business is very interesting. I want to encourage analysts to help us to do that analysis. There are 2 parts of indium phosphide one is semi-insulating and the other semi-conducting. The semi-insulating is usually made for detectors and electronic applications. And the software do semiconducting are usually low EPD and good for lasers, and they are the dominant demand for indium phosphide for many, many years. It's almost like 8-1 in favor of semiconducting wafers, okay? But the last 2 or 3 quarters, the trend is reversed. There's so much more demand on semi-insulating iron doled wafers. So it's as if the laser just didn't grow or maybe they still got too much inventory. And this new demand of iron doled making either some kind of laser or more likely high-speed detector is growing very strong. So what I'm hoping for is that the sulfide material will recover as the inventory get consumed and will come back again. But this iron actually is a new application will continue its own trajectory of growth.

Okay. Very interesting dynamics there. I'm going to ponder that while I ask a couple of other questions here. I guess just a quick one here. You had a 10% customer in the quarter. Can you identify that or at least the sector and whether that they've been a 10% customer in the past.

It was an [indiscernible] house, Richard. And...

So It's actually related to the AI customer, right?

Yes.

Yes.

Okay. Yes. It's a historically long-term customer for us that does epi... And we don't know for sure where that was going. But actually, we do know, but we can't say I'm not going to lie to you.

Okay. Fair enough, Gary. Maybe just a couple of last questions here, and I'll ask both of you to put on your long-term lenses here, or I guess we'll call it a medium-term lens here. But just kind of wanted to conjecture on the opportunity for getting back to breakeven level by the end of this year or early next year? And maybe just as a reminder what that model looks like in terms of revenues and gross margins.

Well, we certainly think that it's a reasonable target and goal to do that maybe this year. So we're not giving up on it.

I would say it's this year. Yes. So

Yes. It's probably somewhere between $28 million and $30 million a quarter, and we would need to get the gross margins to go up maybe close to 28% and keep the OpEx somewhere around 9.3%, 9.4%. That's GAAP and OpEx, by the way. So...

And I think we do need some help from indium phosphide. I mean right now indium phosphide is only 50% of what we used to do at a peak.

Yes. I mean that's probably the wildcard. And it's -- I would say that the indium phosphide surge, especially since it's centered in the artificial intelligence area, it's happening sooner than we expected. We thought it would -- we believe early on it was going to happen, but it's happening sooner. So the question is what's going to happen in the next 2 or 3 quarters. But it's a tremendous opportunity, and there's billions and billions of dollars that are being invested in the hardware side of AI and the software side. So -- and of course, we play in the hardware side. So it's pretty amazing what's on the horizon. So...

Okay. One last question, I'll jump on the line here, guys. The short report that came out on your stock early last month, I think, is -- I think we've established pretty well, there's a bunch of Malarky for the most part. But it does bring up an interesting topic that I think it would be very interesting for you to address. And I think it's risk that a lot of investors bring up with me. And that is related to the STAR listing and the investment by private equity investors. And if there's any rejection of the application for whatever reason here, what's the risk? And how do you get around the risk of having to repay that money and then find other ways to finance the company.

Well, as long as we're in the process they have no right for redemption. And equally important, they don't want to be redeemed. I mean the last thing they want is their money back. And so they're very willing to be patient and that work with us. And it's everyone was disappointed that it's taken longer than we all expected. But...

So let me try to help out a little bit here, Gary. Look, I think -- I mean, hard assets and doing manufacturing semiconductor or materials job industry is hot in demand in China. And that is showing up in the fact that we have other investors interested in acquiring a minority share of our other joint ventures, okay. So although the IPO takes longer than we thought it is. But I think our assets in China are highly valuable. I think the psychology or the thinking from a Chinese investor point of view, the money no longer can go into real estate to invest and doing materials and doing manufacturing and the real fundamental business manufacturing is highly desirable. So if we don't, for whatever reason, we don't go IPO, I would say one possibility is inviting other customers -- I mean, investors investing into our joint ventures. And we have several of them, which are highly valuable. I mean, of course, the most valuable is Toma manufacturing substrates, and we have the world ranking leadership in those fields, although they are smaller, but nevertheless, it's highly desirable. But because we're going public in China, so we cannot separate them and then invite investor into the main body of the investment to me. But for our joint ventures, we can certainly get other investors investing in those joint ventures.

Yes, let me give you -- we have talked about this internally. So, we're not worried about it. But as I said, the PE companies want us to go public. They don't want -- they don't want to pull out. So the -- these these equity -- private equity investors are all large and premier institutions, each with an investment of just under $5 million, which represents a relatively small part of their respective portfolios. And so far, since they have to be patient no matter what, they're being patient. So -- and another thing that's probably interesting to note is that they don't have any recourse or their investment is not collateralized. So that's why they needed to succeed. So -- and we made some significant progress and some developments that we really can't give any details, but some very good steps were taken in the recent quarter, which continues to sustain our hope regarding the IPO.

The next question comes from the line of Tim Savageaux from Northland Securities.

Good afternoon, and congrats on both the results and the outlook and especially the growth in indium phosphide. And I think there was a mention of kind of peak levels that you had achieved in indium phosphide. I think that's getting up towards $20 million a quarter, maybe $1 million -- but at that point, you also had some additional consumer business. And I guess my question is as you look at it now, the market opportunity you're seeing, do you think this AI kind of optical data center opportunity can move you back toward peak indium phosphide levels by itself? And I have a follow-up.

That's a good question. I -- the customer who use our product for consumer products at the time -- I mean, they still have one product using it for consumer product, by the way. They didn't go all completely, and we are believe a big dominant supplier for that product. And we are also in negotiation and in qualification for yet other product although I don't know what's the launch time and whatnot. And the other thing is that talking to their engineers, at one time, they told us they have 11 projects centered around using indium phosphide in their consumer product devices. So I don't think I'm giving that up, but indium phosphide just has so many different characteristics such as fire infrared and using as a detector, et cetera, et cetera. So it's unique in its place to be used as a technology device. Whereas AI application, I think it's an extension of what people use it for data center. If you want high bandwidth, high-speed data transfer, low attenuation, I mean that's the perfect choice. And if you have AI, that means you've got to access data much in 100, 1,000 times, even 10,000 times, and you have so many more data center, you want to exchange information. What's the best way to transfer that information, it's no banner. It has to be some kind of an optical device to transfer that data. So I think it's difficult for me to say when it's going to be so strong and so how big but whether it's going to rival to the consumer product. But I think both of them do need indium phosphide...

Yes, when Morris said optical, that means it's got to have the indium phosphide because the wavelengths can be read by indium phosphide -- so -- but yes, we believe we will get back to those levels, Tim. As Morris says, we're not sure on the timing. And it very well could get back to that level sooner than we thought because of AI. 6 months ago, we didn't have this kind of expectation. I don't know for the people listening to it. I think last week, there was really an amazing article in the New York Times about the amount of money invested in AI in Q1 of this year, and it was, I think, $32 billion. And that's not just software, that's hardware, too. So it's data center stuff. So I mean, we've all been around a long time and so we've seen stuff, but I was very amazed by that. And I sent it to Morris and Leslie, I said it to our Board of Directors because that's going to -- we know we're at the bottom of the food chain or the end of the train, but that's going to benefit any phosphide substrates. And so the dream or the hope is that the consumer comes back and it competes against AI for who's going to be the biggest contributor to indium phosphide revenue. And we don't need a miracle for that to happen. I think it's a very reasonable goal.

Got it. And I want to ask a little bit more about the improvement or the pretty dramatic growth that you saw in Q1 in indium phosphide. And you've already talked about -- you have a unique perspective kind of in the ecosystem. And maybe I'm going to ask you to guess a little bit more here. But I guess my question is around the breadth and how you could characterize that strength. Obviously, you can sell to a lot of different folks in this arena, whether it's Epiwafer suppliers or vertically integrated laser manufacturers or module manufacturers. And I think you've done business with all of those types of guys. And so whether it's looking at the type of customers or looking at wafer sizes, what does that tell you about the breadth of demand that you see as the -- historically, you've had a couple of major customers on the data center side? Or can you see to the extent new customers are showing up? Or are there a couple of customers really moving the needle for you when you see this extraordinary growth in Q1?

The Q1 customer actually is new. I mean they are a customer on other products. But for that application for AI was new. And they're telling us their customer, but I can't repeat it. It's a big, big customer.

Well, there's a -- we know the players, right? There's Meta, there's Google, there's Microsoft, all of those data centers are going to -- where indium phosphide is going to have a play is rack to rack, rack to the aggregation point within the data center and then the aggregation point to a different data center. All of that needs high speed. If you think of the data that's computed in AI then you've got to transfer it and move it around in order to get an end result. And so within the rack, we don't think they'll go to AI to indium phosphide. But rack to rack and beyond, that's where indium-phosphide has a definite application. So...

Of course, Gary, you're not saying anything different. Why is it AI, data set are already doing this. But it's just that data center has to get so busy talking to each other and exchange information a thousand times or 10,000 times that requires higher speed highway. That is where the indium phosphate come in.

The other thing that one of our marketing guys taught me is that some of the large data centers are now having a shortage of power of electricity to run the thing. So they're downsizing in the future architecture to smaller -- more data centers that are smaller but are spread around, and that's going to benefit us. So...

Got it. Well, okay, it sounds like there was sort of a big customer helping to drive that growth in Q1. And sort of the rest of the color I was looking for. I know there's only so much you can say. But clearly, you're at the substrate level, you've got an epiwafer supplier who maybe selling into a module manufacturer. I mean I imagine that you're maybe not all the way up to the cloud provider level, but you have customer touch points all along that ecosystem. And I was just looking for as you see the pieces move around what you're seeing in terms of opportunity at any of those levels, right, whether it's going directly to a chip supplier, obviously, you used to do a lot of business with Intel and their module assets are over at Jabil now. Are you seeing sort of different customers kind of show up at different parts of the food chain, I guess, in driving this AI growth.

Not yet. In fact, I think if I -- I mean we have some China customer, which is showing up -- taking a lot of wafer as well, but they're not telling us anything. I mean I think the customer base seems to be shifting somewhat before these things all happen to consumer product that sort of win way. And then there was a data center that you were talking about, which was taking a lot of wafer, they are not prime customers at the table right now. There is a new Chinese customer coming in, but we don't know where that is coming from. And this new customer is taking this as leading in the phosphide and we know they're telling us it's going to AI. But I wish I can see that's the second one, the third one coming in, that would be great.

We think we understand the architectural needs well enough that there's little doubt in our mind, there's no doubt, frankly, that we're going to end up benefiting and getting into multiple data centers along the big names that I just ran off. So...

The next question comes from the line of Danny Cheng from B. Riley Securities.

My first question is actually on gallium arsenide. So you're expecting that to be the main driver from first quarter to second quarter. Just wondering if there any applications or customers kind of driving this strong growth?

LED is strong for automobile. Lasers is strong, but not as high as at the peak. I would say it's about 50% of the peak level. Automobile is probably 70%, 80% of the peak and HPT is a new, I believe, opportunity for us, and we're trying very hard to get -- get ourselves back into it. And if we are successful, we can expect more revenue growth for HPT as well. And the China market seems to be fairly strong in gallium arsenide.

Got it. And just wondering how sustainable your expectation is on gallium arsenide, I mean are we talking just a few quarters or kind of a multiyear cycle.

I can guarantee multiyear. I would say we have probably good visibility, at least to Q3. I mean demand is strong. But as you know that I do worry about the world economy. I mean I think -- but then people are saying there's a recession, but it never come, right? It's... I mean --

Most of our cycles are more than 1 quarter. So Gallium arsenide is more robust than we expected it to be, but we're not thinking like, oh, then it's going to drop back down in July. So...

Got it. And then just on indium phosphide for AI applications. Is there any data on market share? How big that is and the market share between you versus competitors?

No, we don't know. There is some public projections on markets that coherent shared publicly in one of their presentations. If you haven't seen that, you might want to take a look because there's some stuff in there that might give you some information. So...

But just on the market share. I mean, you think -- should we expect maybe 1/3 each for you and your competitors or somebody has a dominant market share -- just for -- on the AI. Any color on that?

The customer we have, I think, they are giving us all the order, but I don't know whether our competitor is taking order from a different customer okay? In other words, the order we got, we know we got 100% of that order from that customer. And they told us it's for AI. But I don't know whether our competitors are serving yet down the channel.

The next question comes from the line of Matt Bryson from Wedbush Securities.

On the HPT side of things, if you're successful in getting traction, any idea of what the size of that opportunity might be?

Say it again, I didn't hear. The expected value...

How big is the HBT market... What could that look like on a quarterly or annual revenue run rate, assuming you're successful in getting traction in that market?

Okay. I would say close to $20 million a year.

Got it. So that's a nice big round number. Similarly, or slightly different question. I think a lot of the focus on indium phosphide has been around the AI opportunity. But do you have any sense of how close to to the point where inventory is normalized, we are? Like is that 2, 3 quarters out? Is it a year out? And then any idea in terms of how much revenue you think you're losing because there's inventory out there right now? Or what might your revenue look like if that inventory didn't exist, any thoughts?

Yes. I think right now, it's very difficult for me to estimate because I don't even know whether they are in full production or not, although the amount of substrate are buying doesn't look like it's pilot. I think they're making something. But we haven't seen -- I mean, the first thing I'd like to see is, I mean, the last order was at 3 months. If they give us other 6 months order, if they're increasing, then I can estimate and better yet is if there's a second customer sort of coming and wants the same thing, and that's even better. At this point...

And sorry to interrupt Morris. So I'm actually -- I'm asking on kind of that traditional indium phosphide business, like where you know you have inventory kicking around in the supply chain. If you have any idea, so not so much the AI side, but the traditional business, if you have any idea like how -- what the impact on your revenue is today, how much it's holding it off? And then any thoughts on when that inventory might get worked out, and you might resume normal revenue run rate on that older business.

I think that's -- because the business is right now just beginning, it's hard for me to tell. But I mean, I think we have the capacity. We can definitely make 3 to 5x what they are ordering now or even 10x if giving us a little bit of time to increase our facility. So I think the volume is no problem. And I think our product quality really fits well with what they wanted. So at this point, it's I think -- I don't know whether I'm answering a question or not. I think I'm excited about it, and we're trying to get as much information as we can. And we know the customer and customer which is large, I don't think they're pulling around. So hopefully be coming back with increased order or somebody else is going to come in following their lead, please. So I think that's diversion.

Yes. I've had some of these conversations with our marketing guys. And I think they expect that we're going to continue to work through the inventory into the second half of this year.

But that's not AI?

No, no. It's not AI. It's just in general.

Data center...

Data center Inventory Yes. So that's what I was referring to. So -- there's still some out there. But...

You're getting cleared out and then restoring normal levels, that's probably a -- an early 2025 type phenomenon as you're thinking right now?

Yes, maybe Q4 of this year, but definitely in 2025. So it's -- it's going to happen. So we can't wait. We're so excited about it.

Yes. I guess last one for me, Gary. I completely understand that customers don't want to hold inventory. And so they're putting in rush orders, which makes it hard on your end to clear out your inventory because you don't want to turn down business. But I guess given that environment, how confident are you can take down inventory by $10 million? Or what are the dynamics involved in that where you're not effectively having a turn down business because you can't meet those rush orders?

Well, I wanted to take it down $10 million last year in -- it came down last year, but not as much as our target. However, I have a couple of reasons that I think it's a realistic target. One is, if you look back at our historical inventory levels compared to our revenue run rate levels, it was the inventory was in the $60 million level range, $65 million level range. So the difference is we have more inventory in the consolidated joint ventures now because they have different added product lines and things like that. And our recycling program, which is good, it helps us on gross margin and it helps us with ESG, but you're converting what I would call scraps of materials or slurry, which has a little or no book value, and then you bring it in at standard cost. So your inventory goes up. So it's counterintuitive. But even so, I'm absolutely convinced we can take money out of the inventory. Is it going to be $10 million? That's my target. And if revenue grows for us, then it makes it a little easier to take the inventory down. So yes, I don't need a miracle to have that happen. I just need some good business decisions, and if it was easier, frankly, when I could go to China because I would hold inventory review meetings and I haven't been there for a while because of the COVID thing, but I'm going to go this year. So we're getting back in that cycle. But it's a good question. So thanks.

As there are no further questions at the queue this time, this concludes our Q&A session. I would like to turn the call over back to Dr. Morris Young for closing remarks.

Thank you for participating in our conference call. This quarter we will participate in the Northland Security Growth Conference until June 25, and I hope to see you there. As always, please feel free to contact me, Gary Fischer or Leslie Green, if you would like to set up a call with us. We look forward to speaking with you in the near future.

Ladies and gentlemen, this concludes today's conference call. Thank you for your participation.