超微電腦 (SMCI) 2023 Q4 法說會逐字稿

Thank you for standing by. My name is Anna, and I will be your conference operator today. At this time, I would like to welcome everyone to the Super Micro Computer Fiscal Fourth Quarter 2023 Results Conference. With us today, Charles Liang, Founder, President and Chief Executive Officer; David Weigand, CFO; and Michael Staiger, Vice President of Corporate Development. (Operator Instructions) Michael Staiger, you may begin your conference.

Good afternoon, and thank you for attending Super Micro's call to discuss financial results for the fourth quarter and full fiscal year, which ended June 30, 2023. With me today are Charles Liang, Founder, Chairman and Chief Executive Officer; and David Weigand, Chief Financial Officer. By now, you should have received a copy of the news release from the company that was distributed at the close of regular trading and is available on the company's website. As a reminder, during today's call, the company will refer to a presentation that is available to participants in the Investor Relations section of the company's website under the Events Presentations tab. We have also published management's scripting commentary on our website.

Please note that some of the information you'll hear during our discussion today will consist of forward-looking statements, including, without limitation, those regarding revenue, gross margin, operating expenses and other income and expenses, taxes, capital allocation and future business outlook, including guidance for the first quarter fiscal year 2024 and the full fiscal year 2024. There are a number of risk factors that could cause Super Micro's future results to differ materially from our expectations. You can learn more about these risks in the press release we issued earlier this afternoon, our most recent 10-K filing for fiscal 2022 and other SEC filings. All these documents are available on the Investor Relations page of Super Micro's website. We assume no obligation to update any forward-looking statements.

Most of today's presentation will refer to non-GAAP financial results and business outlook. For an explanation of our non-GAAP financial measures, please refer to the accompanying presentation or to our press release published earlier today. In addition, a reconciliation of GAAP to non-GAAP results is contained in today's press release and in the supplemental information attached to today's presentation. At the end of today's prepared remarks, we will have a Q&A session for sell-side analysts to ask questions.

I'll now turn the call over to Charles.

Thank you, Michael, and good afternoon, everyone.

Today, I am pleased to announce a new record fiscal result of $7.12 billion annual revenue, a 37% year-over-year growth rate. We have also achieved our first landmark $2.18 billion quarter, which is 34% growth year-over-year and 70% quarter-on-quarter. We are continuing to exceed both top and bottom-line results and performing well beyond our previously guided ranges. It's no secret that our strong growth has been driven by the demand for our leading AI platforms in plug-and-play rack-scale, especially for the Large Language Model-optimized NVIDIA HGX-based Delta Next solution. Our engineering capability enabled us to deliver optimized, first-to-market AI products and solutions to our customers, distinguishing us from the competition and empowering us to take market share.

I am also proud of our logistics and production teams' execution to maintain our time to market advantage and deliver our total solution to our key partners much sooner than competition. Our investment of 4,000 racks-per-month state-of-the-art validation and production facility in Silicon Valley is one of the major factors in delivering high-performance AI racks quickly with both air-cooled and liquid-cooled options.

Let's go over some key financial highlights. First, we enter FY'24 with record-high backorders, many more new design wins and new customers. Again, our fiscal year 2023 net revenue totaled $7.12 billion, up 37% year-on-year, above our mid-point guidance range of $6.7 billion. Fiscal year 2023 non-GAAP earnings per share grew 109% year-over-year to $11.81, compared to $5.65 a year ago, exceeding the higher end of our revised guidance range of USD 10.50 to USD 11.00.

Fiscal Q4 net revenue totaled $2.18 billion, up 34% year-on-year and up 70% quarter-on-quarter, above our guidance range of USD 1.7 billion to USD 1.9 billion. Fiscal Q4 non-GAAP earnings of $3.51 per share grew 34% year-over-year compared to $2.62 a year ago, and it was well above the high end of our guidance range of USD 2.21 to USD 2.71, demonstrating strong operating leverage and continued customer preference towards our rack-scale Total IT Solutions.

For 30 years, we have been diligently building our company strength and foundation. As the only US-based large scale AI platform designer and manufacturer, we have been shipping our winning products in volume to our partners for more than a quarter. Couple of months ago, I was honored to have my close friend, NVIDIA CEO Jensen Huang, join me on stage at Computex to highlight our optimized new generation GPU solutions for this AI era. We are deploying not just systems, but complete rack-scale total solutions to large generative AI innovators. Supporting HGX H100 Delta-Next, H100 PCIe GPUs, and the just-announced L40S GPUs; our Delta-Next, Redstone-Next, and 8-GPU OVX systems have become the benchmarks of excellence for the industry. We are also getting good traction on Intel's Gaudi 2 and PVC, as well as AMD's MI-250 accelerator solutions.

The demand for artificial intelligence infrastructure is perfectly addressed by our building block server architecture, and due to our building block solution, we have the best and broadest application-optimized GPU solutions on the market. We plan to extend this leadership with the upcoming MGX platforms that we announced at Computex, bringing modular flexibility on acceleration with X86, Grace (C2), and Grace-Hopper (CG1) Superchip. The MGX platform is a shared vision of AI computing between NVIDIA and Supermicro, designed to be open, flexible and future proof for multiple generations of GPUs, CPUs, and DPUs.

We already have many customers engaged with these new MGX platforms as we speak, and soon they will be in volume production. With nearly half of our revenues this quarter based on AI-related designs, I expect this AI growth momentum to continue expanding our TAM across all customer types from major AI innovators, super large CSPs, Tier 1 datacenters, Tier 2 cloud, and to the general enterprise market.

As the performance of GPU, CPU, DPU and memory technologies increase, enhanced storage performance is also necessary to feed massive data sets to the applications without becoming a bottleneck that slows the entire system down. Supermicro's new PCIe Gen 5 based E1.S and E3.S petascale All-Flash storage servers offer industry-leading storage performance and capacity. Together, with our U.2 NVMe, top-loaded scale-out and traditional storage platforms, we are fulfilling customers' AI, compute, and storage needs with a one-stop total solution shopping experience, with the best optimization in performance, time-to-online, and cost.

With accelerated computing and storage, the power consumption and thermal challenges of these new technologies have risen dramatically. 40KW or even 100KW rack solution demand is rapidly expanding and required for compute-intensive datacenter, CSP, and other verticals. Having high power efficiency, free-air and liquid cooling expertise has become one of our key differentiators of success. We have made major investments across a variety of technologies to drive adoption of direct liquid cooling in datacenters to address the thermal challenges.

In addition to increasing computing density and reducing TCO, liquid cooling reduces the environmental impact of datacenters dramatically, aligning with our green computing mission. Similarly, our liquid immersion solution is making good progress, and we will be happy to provide that option to interested partners soon. The reality is that datacenter infrastructure is getting more complex, especially with the unprecedented demand for AI. The design complexity and integration skills, along with the requirement to deploy in a timely fashion are heavy burdens to many end customers. Supermicro Total IT Solutions approach saves them from the complications of design, validation, sourcing, and integration. It also streamlines network switching, firmware and software management in datacenter scale. Customers' peace of mind is topped off by the value of our 24/7 Global Service tiers.

Given the current infrastructure demand, we have been continuing to evaluate our footprint beyond our recently announced Malaysia expansion. We have recently added another new building close to our Silicon Valley headquarter and are aiming to further increase our current 4000 rack-per-month capacity in this fiscal year. To further support demand from key domestic partners, we are also planning to build another manufacturing campus in North America. At this moment, our U.S. headquarter and Taiwan facility can support at least $15 billion in revenue while the new Malaysia facility will further increase our total revenue capability by serving large scale builds on a reduced cost structure in the coming 3 quarters.

In closing, Supermicro is in a great position to continue our growth momentum with our leading AI portfolio, our infrastructure readiness, and our ability to deliver products in a timely manner. Due to the current key components supply shortages, we forecast revenue in the range of USD 1.9 billion to USD 2.2 billion for the September quarter. However, given the record high backorders, we see fiscal year 2024 revenue between USD 9.5 billion to USD 10.5 billion with room to deliver more depending on availability of supply.

Our role as the leader of rack-scale Total AI and IT Solutions has only just begun. We are ready to deliver optimized AI infrastructures to existing and emerging markets, along with our value-added software and services. To say it plainly, our foundation and capacity are fully ready, and our demand is growing strongly. With LLM, Large Language Model, and other AI applications booming, I now expect the $20 billion annual revenue target to be just a couple of years away.

Before passing the call to David Weigand, our Chief Financial Officer, I want to say thank you to our partners, customers, Supermicro employees and to our shareholders for your continued support. David?

Thank you, Charles.

Fiscal fourth quarter revenues were $2.18 billion, up 34% year-over-year and up 70% quarter-over-quarter. Q4 revenues exceeded initial guidance range of USD 1.7 billion to USD 1.9 billion and were at the high end of the recently updated range of USD 2.15 billion to USD 2.18 billion. For fiscal 2023, we reported revenues of $7.12 billion, representing 37% growth over fiscal year '22 revenues of $5.2 billion. Next generation CPU and AI platforms continue to drive record levels of design wins and orders. Exiting fiscal year 2023 with a record backlog, we are well positioned for fiscal year '24 with an outlook for continued revenue growth and profitability. We expect diversified growth driven by top-tier datacenters, emerging CSPs, enterprise AI build-outs, CPU upgrades, and edge/IOT/telco markets. We are also targeting new opportunities in adjacent markets such as Storage, Switches, Software & Services. We note that our shipments against a record backlog may continue to be constrained by supply chain bottlenecks for key new components for our advanced AI server platforms.

Q4 results were driven by our high-growth AI/GPU and rack-scale solutions which represented 52% of our total revenues. We had 2 10% customers for Q4 and did not have any 10% customers for fiscal year '23. During Q4, we recorded $976 million in the Enterprise and Channel vertical, representing 45% of revenues versus 50% last quarter, this was up 19% year-over-year and up 51% quarter-over-quarter as we ramped several key enterprise programs. The OEM appliance and large datacenter vertical achieved $1.17 billion in revenues, representing 53% of Q4 revenues versus 47% last quarter, up 59% year-over-year and up 94% quarter-over-quarter as we gained momentum from existing and new datacenter, CSP, and OEM cloud appliance customers. Our emerging 5G/Telco/Edge/IoT segment achieved $43 million in revenues, representing 2% of Q4 revenues versus 3% last quarter.

For the fiscal year 2023, Organic Enterprise and Channel and AI/ML revenues grew 10% to represent 48% of total revenues. The OEM appliance and large datacenter segment grew 102% and represented 49% of total revenues. The emerging 5G/Telco/Edge/IOT segment decreased 36% and represented 3% of total revenues.

The mix of complete systems and rack-scale Total IT Solutions has increased over the last 2 years. Server and Storage Systems comprised 93% of Q4 revenue and Subsystems and Accessories represented 7%. ASPs continued to increase on a quarter-over-quarter and year-over-year basis driven by the value and complexity of our rack-scale Total IT Solutions.

By geography, U.S. represented 76% of Q4 revenues, Asia 11%, Europe 10%, and Rest of World 3%. On a year-over-year basis, U.S. revenues increased 55%. Asia decreased 17%, Europe increased 1%, and Rest of World increased 12%. On a quarter-over-quarter basis, U.S. revenues increased 112%, Asia increased 10%, Europe decreased 1%, and Rest of World increased 11%.

The Q4 non-GAAP gross margin was 17.1%, down quarter-over-quarter from 17.7% in Q3 as we focused on winning strategic new designs and market share. For fiscal year '23, the non-GAAP gross margin of 18.1% versus 15.4% for fiscal year 2022 was driven by an increased rack-scale product and customer mix and higher manufacturing efficiencies.

Turning to operating expenses, Q4 OpEx on a GAAP basis increased by 14% quarter-over-quarter and 19% year-over-year to $145 million driven by headcount and related expenses. On a non-GAAP basis, operating expenses increased 14% quarter-over-quarter and 17% year-over-year to $133 million. Our non-GAAP operating margin increased significantly to 11% for Q4 versus 8.7% last quarter due to operating leverage driven by higher revenues that outpaced increases in operating expenses. For fiscal year '23, we achieved a non-GAAP operating margin of 11.4% versus 7.2% in fiscal year '23 due to higher gross margins and operating leverage from higher revenues and expense controls.

Other Income and Expense for Q4 was an expense of approximately $1.5 million consisting of $3.5 million in interest expense offset primarily by a net foreign-exchange gain of $2.0 million. Our interest expense increased sequentially as we utilized our short-term credit lines for working capital requirements along with higher short-term interest rates on our borrowings.

The tax provision for Q4 was $31.3 million on a GAAP basis and $36.6 million on a non-GAAP basis. The GAAP tax rate for Q4 was 13.9% and the non-GAAP tax rate was 15.4%. The GAAP tax rate was 14.7% for fiscal year '23 versus 15.7% in fiscal year '22 and the non-GAAP tax rate was 15.9% versus 17.7% in fiscal year '22.

We delivered strong Q4 non-GAAP diluted EPS of $3.51 which exceeded the high end of the original guidance range of USD 2.21 to USD 2.71 and the recently updated range of USD 3.35 to USD 3.45. The increase in Q4 EPS was due to a combination of higher revenues and operating leverage. For the full fiscal year 2023, we reported non-GAAP diluted EPS of $11.81, up 109% year-over-year versus fiscal 2022 non-GAAP diluted EPS of $5.65 and higher than the fiscal year '23 guidance of USD 10.50 to USD 11.00.

Cash flow used in operations for Q4 was $9 million compared to cash flow generated by operations of $198 million in Q3 due to higher accounts receivable, offset by lower inventory and higher accounts payable from backend loaded shipments in the quarter due to supply constraints. CapEx was $8 million for Q4 resulting in negative free cash flow of $17 million versus positive free cash flow of $190 million last quarter. For fiscal year '23, we had cash flow from operations of $664 million, CapEx of $37 million resulting in free cash flow of $627 million. We have $50 million remaining under the authorized buyback program which expires on January 31, 2024.

The closing balance sheet cash position was $440 million, while bank debt was $290 million resulting in a net cash position of $150 million down from net cash position of $176 million last quarter as we utilized our bank lines of credit to support higher revenues and accounts receivable as we ramped up production of new AI/GPU design wins.

Turning to the balance sheet and working capital metrics compared to last quarter, the Q4 cash conversion cycle improved to 77 days versus 126 days in Q3. Days of Inventory was 75, representing a decrease of 51 days versus the prior quarter of 126 days. Days Sales Outstanding was down 13 days quarter-over-quarter to 38 days while Days Payables Outstanding came down by 15 days to 36 days. Faster inventory turns contributed to the improvement in our cash conversion cycle.

Now turning to the outlook, we remain enthusiastic about our diversified business model covering a wide range of AI, core computing, storage, 5G telco/edge and IOT applications. Design wins span across enterprise/channel, large datacenter/emerging AI CSP, and Telco and OEM customers. We are carefully observing the global macro-economic situation and continuing supply-chain constraints especially for leading AI platforms. For the first quarter of fiscal 2024 ending September 30, 2023, we expect net sales in the range of USD 1.90 billion to USD 2.20 billion, GAAP diluted net income per share of USD 2.02 to USD 2.80 and non-GAAP diluted net income per share of USD 2.75 to USD 3.50. We expect gross margins to be similar to Q4 levels. GAAP operating expenses are expected to be approximately $185 million and include $48 million in stock-based compensation and other expenses that are not included in non-GAAP operating expenses. GAAP and non-GAAP operating expenses are expected to increase due to continued investments in R&D and higher personnel costs adding to our great engineering and sales teams.

The outlook for Q1 of fiscal year 2024 fully diluted GAAP EPS includes approximately $44 million in expected stock-based compensation and other expenses, net of tax effects of $9 million, which are excluded from non-GAAP diluted net income per common share. We expect other income and expense, including interest expense, to be a net expense of approximately $1.5 million. The company's projections for GAAP and non-GAAP diluted net income per common share assume a GAAP tax rate of 14.7%, a non-GAAP tax rate of 15.4%, and a fully diluted share count of 56 million for GAAP and 57 million shares for non-GAAP. We expect CapEx for the fiscal first quarter of 2024 to be in the range of USD 10 million to USD 12 million. For the fiscal year 2024 ending June 30, 2024, we are giving guidance for revenues in the range USD 9.5 billion to USD 10.5 billion.

Okay, Michael, we're now ready for Q&A.

Operator?

(Operator Instructions) Now we have the first question come from the line of Nehal Chokshi from Northland Capital Markets.

And congratulations on an amazing quarter and amazing fiscal year '24 revenue guidance of USD 9.5 billion to USD 10.5 billion. I definitely want to dig into that. First off, Charles, I think you said you had 50% revenue exposure to AI for the June quarter. That's incredible. It makes a very strong statement that Super Micro indeed is the leader in AI systems. What do you think is a sustainable differentiation that you guys are building with rack-scale AI systems that's driving that incredibly fast increase in revenue exposure to AI systems?

Yes, AI have been our major focus since a few years ago. And we work with AI chip company and like NVIDIA very closely. And we could develop a lot of platform, including our MGX, now coming MGX, as you know, to support the C2 and CG1. So our AI solution with rack-scale or cloud-scale (inaudible) solutions that make our customers deploy even much easier and fully have their solution for energy cost savings for time to online. So I believe our AI product line business will continue to grow, in the near future maybe continue to be more than 50% of our revenue.

Next question comes from the line of Mehdi Hosseini from Susquehanna International Group.

And Charles, I think it would be very helpful if you could (technical difficulty) $10 billion of revenue. And perhaps any kind of color on ASP increase that would capture your increased content would give us an insight unless you have other ways of explaining to get to $10 billion.

We continue to gain lots of new customers. All partners continue to like our solution, and we continue to gain new customers with our rack-scale, cloud-scale total solution, including air cooled and liquid cooled, right? So we see a very strong demand and growing very fast back order. So $10 billion should be a kind of a short-term target for sure.

The question, your competitors also argue for an innovative cooling and they also argue for share gain. Is there something with your cooling technology that offers better cost performance compared to some of the emerging cooling that your competitors are marketing or talked about?

Yes. Indeed, our engineering team is strong and dedicated. Everyone can design systems, our system simply better than others, not just better in quality performance, but also earlier time to market and target with service, software, system management. So it's a total solution and service. So we feel pretty confident that we will continue to gain much share.

Okay. Maybe move on, you talked about AI becoming more than 50% of your revenue, how is your pipeline as you look into second half of calendar year, do you already have a pipeline that gives you confidence you're going to hit that 50% plus AI as a mix of revenue? Or do you still have to go and qualify and win the business to hit that target?

Indeed, the pipeline has been very strong. And at the moment, the visibility is very clear. So we are very confident. And especially we continue to gain the customer too.

Your next question comes from the line of Jon Tanwanteng.

Congrats on a really strong quarter outlook again. My first one is I just want to dig into your confidence for the outlook. Can you just, in various parts, maybe talk about your confidence in supply chain, being able to supply that maybe 40% growth guidance at the midpoint, maybe talk about your ability to get allocations from key suppliers. And then within your order book and backlog, what is your assurance? Or how can you identify if there's double or triple ordering? And how do you protect yourself against that if they do exist?

Yes, it's a complicated question. We have a very good product, many product lines for data mix, for [restore] mix, for Grace, for Grace (inaudible) and other solutions, and including (inaudible). So the solution is really strong and supply chain we work with our vendors very closely every day. And so hopefully, the situation will be consistently improved, but it's not 100% controlled ourselves. So although we have a good partnership with vendor, with customers. So we work very, very closely. And the situation will be continuing to improve a bit.

And on the order side...

Are we ready for the next question?

Jon, did you have a follow...

On the order side, protection mixed up on...

Yes, and see the order side, I mean, the back order has been continuing growing strongly. Every month, I see the order growth.

Also, Jon, we have a lot of [NCNR] orders as well. That protects against double and triple booking.

Your next question comes from the line of Ananda Baruah from Loop Capital.

Yes, congrats on the strong results and the ongoing momentum. I have a couple if I could. Charles, you talked in the prepared remarks about the unprecedented demand you're seeing, and you guys have talked about actually adding new customers, including top-tier datacenter customers, I think you mentioned. And some of these Gen AI server forecasts for 2024 calendar are really, really strong, and you're also talking about gaining share, et cetera, et cetera. And so I guess the first question is what's the opportunity you see to maybe even do keep stronger than the fiscal '24 guidance, I guess, what would be the puts and takes there? And if you were to be able to exceed the 2024 guidance, what would be some of the things you think would need to occur? And then I have a quick follow-up.

Yes. Thank you. Very good question. Again, we continue to gain generative AI innovator, and we have a very good partnership. We started to ship them some (inaudible). And for sure, they need 20x more system. And we just cannot ship at this moment because of supply chain. And at the same time, we also continue to engage with large CSP and large datacenters. So we continue to gain more customers from, I would rather say, (inaudible) vertical. So they had to work out with the supply chain, the 3-party relationships, our supplier, our customer and ourselves. And that has been our major focus now, although we continue to improve our software total datacenter solution, including (inaudible) solution. So I mean, we are on the right track. Yes, expecting supply chain can improve so that we can grow our revenue.

That's really helpful. And so Charles, just to make sure that I understood that accurately. Is that to say if you can get more from the supply chain, well, actually, is to say this way, you have order visibility such that if you can procure more, you would have the ability to share gains exceed the fiscal '24 range that you provided? Is it really a supply chain issue, I guess, is what I'm asking. Did I hear that accurately?

Yes, absolutely. I mean we continue to prove to our customers and our vendors, hey, we have a basic solution. So work with us. That's the way we went for everyone.

I got it. And then let me just ask quick follow-up to that one. Is there any way you guys could provide any context for us around, I guess, how constrained you are like what might the demand outlook sort of look like if you were not constrained. So not like a guidance, right? Obviously, you gave your guidance, but just context for us to really get a sense of what the structural positioning of the company is if they were not constrained.

Yes. Again, we have based on the product, and we have a very good part of our customer, and we push our vendor. And we know our vendor doing their pace to support us at the same time as well. So we just to continue work together.

All right, Charles. I have one more quick one. I'll move away from demand. I appreciate it. It sounds like the gross margin guidance for the September quarter is a pretty solid most attractive guide. For fiscal '24, should we assume that same 17% kind of gross margin range? Or what's the right way to think about that? And that's it for me.

This is David. So we're targeting to hold our margins, and that's all the guidance that we'll give right now.

Your next question then comes from the line of Aaron Rakers from Wells Fargo.

Yes. And also, congrats on the results tonight. So first of all, I just want to clarify, I want to make sure I heard the number right. Was it 50% or 52% of the revenue that was from your AI and rack-scale solutions this last quarter?

Yes, Adam, Charles in his script said approximately 50%, and I clarified 52%.

Okay. And so I guess the question that leads to that is that one of the questions I get asked a lot about is the spend around AI being so strong, and it sounds like, obviously, you're carrying a pretty good backlog looking out over the next several quarters given the supply dynamics. But if I take that number and I say the non-AI business, how has that trended? Or basically, are you reallocating capacity away from more of the non or traditional compute side to satisfy demand on the AI side? Or have you seen spending slow down outside of these AI investments that you're clearly benefiting from?

Yes. on AI server storage, IOT telco, we keep about a little bit flat. I understand the industry and that'll be declining. We although not declining above flat, and we try to grow as well. And that's why we are growing in many direction kind of manufacturing kind of like Taiwan, Malaysia, and we may have another North America campus. So that's all to increase our capacity so that we can grow traditional datacenter business as well instead of only focus on growing AI.

Okay. So just to be clear, I mean, you're limited on your supply on the traditional compute side because of the AI demand that you're seeing. Is that fair?

Not exactly impacted but not exactly. Indeed, as you know, the traditional server, traditional datacenter business that had been flat from the demand side. So we are facing to a similar experience as well, but not declining to...

Yes. Okay. And then the final question for me is just thinking about the AI opportunity, and I think you alluded to this a little bit in your prepared comments. You mentioned positioning around, I think it was the Intel Gaudi chip and then also, I think another vendor coming to market, the MI 300 from AMD. How do you see that playing out for your opportunity? Does it help satisfy demand? Are you seeing indicators that this AI opportunity is going to be much more diversified here as we move into '24 versus what you've seen on the NVIDIA side at this point? Just curious how you're seeing kind of the competitive landscape or maybe opportunity set expand for you guys with those new or other solutions coming to market?

Yes. NVIDIA solutions (inaudible) leader and people really are with NVIDIA, more supply and other solutions from (inaudible) AMD, because our building block solution, we have a solution ready for all of them. So that's the advantage we have. So we just waiting for their solution to be available in potential.

Your next question comes from the line of Jon Tanwanteng.

Dave, I was wondering if you could talk about your working capital needs in the sort of environment. Can you generate positive cash flow going forward? Or are you going to be using cash as you try to fulfill this outside demand?

Yes, Jon, we see the business generating good cash flows as it has historically. And we think that especially in this constrained supply market where we could deliver more if we had more supply. But really the constrained supply ends up moderating the working capital. And so we grew our business last quarter quite a bit and grew our AR. So that utilized a lot of working capital. But we have no concerns about working capital.

Okay. Great. And then could you guys give a little bit more detail on the capacity expansion and when those various facilities come online? And what exactly they add?

Sure. We have Malaysia, which is expected to come online in around 12 to 15 months. And that will eventually double our capacity. And we also have additional capacity coming online in our Building 23 here in our Silicon Valley campus, and we've also added, as Charles mentioned, another site in San Jose with intentions to add another site in the Americas.

Got it. And then last one for me, maybe a more detailed question. What percentage of your AI sales right now are liquid cooling base? And do you expect that percentage to increase as you go forward?

This is very new questions. So we have a very good [DLC], direct touch liquidity solution ready to go now. And it's all dependent on customer demand. And at the same time, the (inaudible) solution also getting ready. So we have an older 3D solution and depends on customers' demand. (inaudible) for sure, (inaudible) still a majority as you know.

Your next question comes from the line of Nehal Chokshi from Northland Capital Markets.

Yes. Thank you for the follow-up question. So NVIDIA effectively guided their July quarter datacenter revenue at 2x Q-o-Q, what does that mean for Super Micro in terms of GPU systems, especially what I'm trying to drive at is that is there a lag between when NVIDIA gets recognized revenue and when the rest of the server OEMs get to recognize revenue?

We believe their capacity are growing. And that's why we talk to them every day, asking for more. So hopefully, we can gather more support from them. And hopefully, their capacity can grow mostly quickly. So that's all I can say now.

Okay. Do you have a guess on what is (technical difficulty)

We're not going to offer a market share for the June quarter. You can make some assumptions by looking at the results of others. But we sell different GPUs, as Charles mentioned. There is a direct correlation.

Yes. Basically, not for capacity. We are waiting for more AI chips, that's a situation.

Yes. Understood. And how should we think about distributing the remaining $8 billion across the final 3 quarters here?

We're not going to announce quarter-by-quarter our guidance. But we're expecting this to be a robust year. And as tempered by the natural supply constraints because of the popularity of these new platforms.

Yes. Basically, in media, we have more capacity for sure, and we're really happy to wait for that. And besides other than the (inaudible), is completely ready. And our CG1 solution is pretty much ready as well. So we believe we can shift much more and that's the situation. So in the next 4 quarters, I believe we will continue to grow quarter after quarter. And even the supply condition better, I believe we can surpass $10.5 billion for sure easily.

Thank you, everyone, for joining, and your time for questions at this time. And this will conclude today's conference call. You may now disconnect.