Micron Technology, Inc. (MU) Earnings Call Transcript & Summary

Very well, good afternoon, everybody. My name is Raji Gill. I'm the Managing Director here in Needham & Company. I cover global semiconductors as well as automotive technology. We're very pleased to host a virtual fireside chat with Farhan Ahmad, the Director of Investor Relations for Micron. This will be about 40 to 45 minutes. If there are any questions that you guys have, just there's a chat box on one of the links so you can message me, and I'll do my best to try to answer those questions as well. So welcome, Farhan.

Thanks. Thanks, Raji.

So if we can maybe -- if we could get started on the Huawei issue to frame the conversation and then can go into some specifics, how would you -- in terms of percentage of sales, how much revenue are you generating from Huawei currently and what's the split? And can you talk about how the current measures against Huawei are different from the measures that were enacted before? And describe how you were able to potentially circumvent those existing de minimis rules. I just wanted to get a sense of Huawei and then we can go into a little bit further details.

Okay. Thanks. Before I start, I just want to say that I'll be making some forward-looking statements today. And for a discussion around risk and uncertainties associated with these forward-looking statements, please refer to our filings, specifically our 10-Q and 10-K filings that we filed with the SEC. And so with that, turning to your question on Huawei. So let's back up to maybe like before Huawei was added to the Entity List, and then I'll walk through like how our business has changed there. So at the time, I would say, more than a year now ago that Huawei was first added to Entity List, at that time, we stopped initially all shipments to Huawei and then we did an evaluation of what the regulations really meant. And based on the regulations, we concluded that most of our shipments that were being produced overseas we could continue shipping in the server and mobile markets and we -- without any major issues. However, we were limited in our ability to support our customers on warranty and also to engage with them on new product development. So like just qualification of new products, we had some challenges related to that. Since then, like what we have done, is that we applied for licenses to allow us to collaborate with Huawei. And we got these licenses after 2 months. And after we got the licenses, we were allowed to work closely with Huawei on qualification of new products. And we also were allowed to support them on warranty. And the last thing is that we were also able to sell products that were manufactured in our U.S. fabs, which are a small portion of our overall production. But nevertheless, prior to us, Micron, getting the license, we were unable to sell products manufactured in that fab. And after that, we were good. So after we got the licenses, our business largely was back to pre-Huawei Entity List addition in terms of our ability to engage with them in the mobile and server space. And -- however, like it was a lower portion of our business. Now like pre-Entity List addition, Huawei was like over 13% of our customer -- of our revenues and -- I should say 13% of our revenues. And now it is less than 10%. We don't disclose any customer less than 10%, but Huawei is no longer a 10% customer. We won't explicitly say how much it is, but it is still an important customer for us even though it is not a 10% customer. And now in terms of the recent regulation changes, like if you look at the press release from Department of Commerce, the measures are based on like what the Department of Commerce is saying is mainly directed -- seemed to be mostly directed towards foundry and like Huawei getting their design manufactured at foundries. At least in our initial assessment, we don't see that there would be any direct impact to us, but this is very preliminary and we are continuing to evaluate this situation. And so like as there is no impact to our ability -- or should not be an impact to our ability, it does not change, like whether any licenses are needed or not, like all of that -- our business with them, indirect relationship, is not really impacted. But then there could be indirect impacts because of like if Huawei's supply chain is disrupted, that will have an impact on the demand from them. That could have an impact on demand from them. So we are still evaluating all of those impacts. It's very early days. But that's our initial sort of view.

So as it stands now, is it fair to characterize this as status quo? With regarding your relationship with Huawei, minus the fact that the percentage of sales have come down a bit from 13% to less than 10%, that you still -- that you have got the licenses in the past to ship new products, plus to support warranty, plus to sell some small percentage of the products being manufactured in the U.S., those steps have not changed?

No. Nothing has -- nothing -- so our relationship on the server and mobile space is pretty much the same as it was prior to the Entity List addition. And the recent changes should not have a direct impact on our relationship or any direct impact on our relationship with Huawei.

Okay. And what type of -- this might be hard for you to answer, but what do you think some of the countermeasures China will take against U.S. companies just broadly and what kind of actions that China could possibly take that could potentially impact your business in China itself, outside of Huawei?

So I would say like it is purely speculative at this point. And like I don't think I can add much color on what it could take and what -- like I don't think like speculating on what -- how a geopolitical situation may evolve in the coming months is going to serve any purpose because, frankly, we don't have any insights into it.

That's fair. You talked about the potential indirect impact to Huawei's supply chain and affecting demand. And two, the broader point of smartphone demand, in general, China's smartphone was, I believe, around 30% of revenue -- or smartphone, in general, was around 30% of revenue.

And that's correct.

Okay. So we have seen early signs that China handsets have started to recover in some recent data, that it's -- in the month of April, was up 15%, 17% year-over-year. You also talked about this on your call back in March -- late March. How would you characterize the China handset market in terms of overall shipments as we go throughout the -- through the second half? And is there -- your content in the smartphones, would that be able to offset any potential unit weakness given the slow start? For instance, if you look at your mobile business in total, will the content gains be able to offset the decline that you -- that the market experienced in February?

Okay. So obviously, the -- it's fair to say that the mobile market in the first half of the calendar year is a little bit weaker than -- is weaker, I should say, than what we had and especially going into the year. In particular, the China smartphone market, as you rightly pointed out, I think you characterized the market situation quite well. It was weak early in the first calendar quarter, and then in March, you started seeing the rebound that we talked about on our call. And as you rightly pointed out, the sell-through from those customers seems to be pretty good as well as China has a recovering of -- as China recovers from COVID, you are seeing the demand come back, which is fairly positive. For the 2nd half of the year, we are not really providing any color, and that's not to say that the market is really bad or good, but it's just that it is very uncertain. And by that, I mean, look, we are seeing pretty dramatic GDP declines across very major economies. And for us to really forecast what the market will do in those kind of market environment in the back half of the year is really something that we don't feel confident about discussing with investors because it could be that we are seeing certain demand, but what confidence do we ascribe to it given the macroeconomic uncertainty that we are -- that the overall economy is experiencing. So we don't think like it is really useful to provide that color on the second half. We will just say that the second half environment for the year is uncertain. That's how we described it on our last call. And like maybe on the next earnings call, if we feel comfortable, we may be able to provide you more color on maybe the third quarter or the second half of the year. But as of right now, we will just leave it there. It's a very uncertain environment. Now to your point on content growth versus unit growth, I mean, that's a very good point. If you look at our smartphone market overall, if you -- we have seen that -- within the smartphone space, our revenues have doubled from 2016 to 2019, roughly speaking. And that's in a market environment where the units have declined pretty much every year, right? So despite the unit declines, the content growth has been a big positive driver. Now as you think about 2020, obviously, content will be a positive driver as well. But at the same time, what is important to keep in mind is that from the beginning of the year, pre-COVID to post-COVID, you have had a clear decline in the unit sales and that news that the market will be weaker than what we had anticipated coming into the year and how it's like, will the bets be up, probably, like it's reasonable to think that it should be up. But like what ends up happening remains to be seen. And given that second half of the year is very uncertain, we should -- I would say that we would just avoid any full year commentary given the uncertainty. And now again, I would say that, look, like this is not -- you should not take that -- this news that the second half is bad in any way. But it's just that we have chosen to be -- to characterize the second half environment as one of uncertainty because, quite honestly, given the broader macroeconomic situation, we just feel that we can't really count on all the forecasts that we are getting from our customers. And there is the higher-than-usual risk for all of those forecasts, and that's why we don't want to really bank into the second half market.

No. I understand the inherent uncertainty in the forecast from the customers, whether it's mobile or not. But I'm just trying to frame the risk. And one way to understand it also is the 5G smartphones as a percentage of the overall smartphones. Based on the conversations I have had with other companies in the 5G space, the road maps, the smartphone road maps for 5G, appear to be unchanged. So they're still forecasting anywhere between 180 million to 230 million, 240 million of 5G phones. Where the weakness has occurred is really on the low-end phones, 2G phones and 3G phones. So we could be in a situation -- and again, I'm not telling you to bless this or not, but we could be in a situation where overall units might be down 10%, 20%, but the mix of 5G could jump to 40% of phones. So that, combined with the fact that your DRAM content on the 5G phones is going up, the 68 gigs, correct me if I'm wrong, that one will insulate you a little bit to some of the weakness in handsets. And the same thing on the NAND side. So I just wanted to get your thoughts on that kind of dynamic. Not saying whether mobile will be up or down, but is that your assessment or is that still pretty -- is that overly optimistic?

No, I think like it's fair to say that the content growth is a positive driver for us and will continue to be a positive driver for our business. But where I'm a little bit hesitant to say is that will it be big enough to offset the weakness in unit growth or not because it really impacts -- it depends on how much the unit growth will be transitioned. 5G, if you think about like 150 million to 200 million units, like something in that range, we have said that it's about 10% to 15% of the overall market this year. That was our assessment, like I would say, pre-COVID even. And so like the numbers that you are talking about, with 180 million, 200 million, will still be in that ballpark, like not like very far from that. And obviously, the units on 5G are not going up. The units on 5G are what they were, maybe down some. The other units are down a lot more. And so the demand is net-net down a lot before from pre-COVID to now. The mix certainly is bigger. Now one thing I do want to point out is that, and this is more like I'm stepping out of the mobile and 5G. But if you were to take a step back and look at what's happening really right now, the overall economy is slowing down and there is less money in consumers' pockets that they are -- in their spending now going on. Like retail spending is down like 40%, 50% in some geographies. So the spending is down less, but what is happening at the same time is that we are using our electronic devices a lot more. We are buying a lot more on the Internet. We are spending a lot more time on entertainment on the Internet. We are talking remotely. We're doing a lot more video conferences like we are having right now. We are doing it socially even as well. So a lot of these digital technologies are really being adopted right now. And the -- there is acceleration in the trend towards digitization of the overall economy. And that applies to consumers, that applies to the overall data center and cloud markets as well, where Microsoft talked about the demand in the cloud space being pulled forward like by a couple of years because of the acceleration that they have seen related to COVID and work-from-home. So all of that is setting us up for an economy, where a lot of the share of economy is going towards digital economy, and that is a very positive thing for our market. So now when you think about in that backdrop, as the economy recovers from COVID, I think it's fair to assume or expect that we will be coming out as a beneficiary on that -- in that transition and the fact that on the supply side, you have seen some incremental actions by the different suppliers to curtail their growth -- capacity growth plans. So as we emerge from COVID, I think there is probably going to be better market dynamics. But like in the near term, while we are facing COVID uncertainty, is while there are still these work-from-home orders and shelter-at-home orders, there is a level of uncertainty that is much higher than usual.

And so to your point, Farhan, the stay-at-home economy is driving more demand for memory, and that bodes well for your compute and networking segment, which is roughly 40% of sales. The larger portion of that is server and client DRAM modules, and the rest is formed by networking graphics. The -- I wanted to get your thoughts on the demand for cloud infrastructure. Last quarter, it accelerated based on underlying hyperscaler demand for these things we've talked about, whether it's gaming or e-commerce. Do you think this will continue throughout the year? Do you think we're hitting a peak and will kind of subside? Clearly, last year, we started to see data center recover in the second half after a long inventory correction. Then you had COVID that really accelerated the hyperscaler plans. Just qualitatively, what's your thoughts on hyperscaler as we sit almost in the middle of the year?

So I think like it's fair to say that for the first half of the calendar year, the data center demand has been strong and that applies also to cloud companies. As -- I mean as I said before, on the second half of the year, we are not really providing much color on it. Some of the other companies have talked about second half, like at least in the third quarter, trends looking -- continuing to remain strong, but like we have not really commented on it and we will stay away from sort of commenting on the second half of the year. What I would say is that the long-term demand trajectory in the cloud and data center space continues to be very positive for us. And I -- and we feel good that longer-term, this will be a growth market. We are still in the very early innings of the cloud and data center buildout. So we do feel confident about it. But second half of the year is something that we really don't plan to have visibility on at this moment because if you think about even the demand for cloud customers and if that weakens for some reason, if the enterprises see a macroeconomic slowdown and as a result, they start spending less on the cloud services, that can have a ripple effect that could have an impact on us. But it is also possible that the data center could continue to be strong. So I just wanted to be -- to kind of view the second half environment as uncertain without really ascribing any particular view to us.

In the fiscal second quarter, your -- you talked about the data center SSDs experienced more than 50% growth. And you also started to see supply shortages for some of the server DRAM modules that go into data center, which is a large portion of data center is the server-class DRAM modules. And then you were shifting the supply from smartphones, both on the NAND and DRAM side, to the data center to accommodate this demand. I just want to get your thoughts. I'm not talking about the second half, but I just want to get your thoughts on that dynamic. And also some of your customers you had mentioned were building inventory to be prepared in the case of supply shortages, and I wanted to get a sense of what the inventory situation is in that area.

Okay. So let me start with inventory and then come to the supply between mobile and server. So on the -- in terms of the inventory, the inventory customer is a lot better than what it was in, say, second half of 2018 -- or I should say, by early 2019, middle of 2019. The customer inventories were in a much worse place like late 2018, early 2019. And now the inventories are not that bad, like they are in a much better place. Now we have talked about 2 areas where there might be some inventory builds. It's not of the magnitude that we talked about in 2018. First, in the area of China, like just the geography, China, more broadly, all the end markets there, some customers have increased their inventory level possibly because of the trade tensions that are there between U.S. and China. And so we have talked about inventory builds in that particular part of the market. And then the second part of the market where we have said is, just more broadly, where certain customers may be building some inventory because of the COVID supply uncertainty, right, like I mean, our fabs are operating in regions where there was outbreak of COVID and similarly, our competitors' fabs were operating in regions of which were at risk of COVID outbreak. And so as a result, like if you are a customer, it is natural for you to be like -- to try to build some inventory to allow you to have some redundancy in case like some supply goes off-line, you can -- you don't want that your entire operation is held up by some particular incompetence. So in general, I would say that that's what we think would be the logical thing for some of them to do. We don't have very high visibility in our customers' inventory. So we do anticipate that there might be something going on related to this. But overall, when we think about the inventory situation, the data centers are in a far healthier place than they were in at the end of 2018, still like a relatively healthy portion of the market to -- compared to the other segments. Mobile generally doesn't tend to have a lot of inventory builds, but maybe there might be some inventory builds there at some of the customers. And in some parts of the market, the demand has weakened, right, like in mobile or in autos. So if you are planning for a certain level of production and suddenly the production tanks, you may end up with like a high level of inventory as well. So autos is another area which is like that. And by the way, like we have -- one thing I do want to mention is that we have cut our production, our utilization rates in the auto-related capacity to make sure that we can control our supply in that part of the market as well. So that is an action that we have already taken. Manish Bhatia, who's our EVP of Global Operations, provided that update recently as well. So I just want to make sure you are aware of that.

And just on the auto part, what percentage of the revenue is coming from auto?

It's a small portion. We have much higher market share than our competitors. But it is a single-digit number. It is not a double-digit number.

Okay. Got it. So you've been cutting up your production of that market. So just going back to the data center though and just to my initial question, you talked about supply shortages last quarter that led you to shift your NAND and DRAM supply from smartphones to data center applications. Is that continuing, that dynamic?

So look, like it is always a very dynamic process. We look at our forecasts for mobile and server. And we ship capacity as our demand forecast changes. So like if demand for server goes up, we continue that. If mobile goes down, we reduce it. And there was a big change, like one of you got the COVID-related impact, and as that demand changed, we adjusted our supply. And we adjust -- we make adjustments to our production on a weekly basis, quite honestly. And what we said on the last earnings call was that we move -- we are moving some supply from server -- from mobile to server.

In terms of the increased need of server-class memory for data center -- for hyperscaler, how has the workloads changed and affected the demand for memory, pre- and post-COVID? Have you seen an increase in AI workloads, for instance, post the COVID situation and therefore the memory requirements have changed? Or is it just about scale, we just need to scale more?

That's a good question. I would -- but to be completely honest, I don't know the answer to that question. I think it is fair to say that there are certain types of workloads that have gone up a lot, like some companies have talked about video conferencing and that being -- equals up a lot. Cloud services have gone up as well. I would think that it is more broader rather than like only one kind of workload really increasing. So that would be my guess, but to be completely honest, I'm -- I have not really looked into it and I don't know the answer to that question.

And so if we look at the revenue guidance range that you provided, it was a wide revenue range, it was about a $600 million range. Excluding COVID, the revenue would have been on the higher end of the range. I think you had talked about $4.6 billion to $5.2 billion with the range. You would have been on the higher end and then likewise, on the gross margin, would have been higher than 31%. So just 2 questions there. On your own internal inventory, you had talked about carrying higher levels of inventory, raw materials, flexing your supply chain to ensure it has redundancies. There's obviously been a lot of shelter-in-place issues across the world, Malaysia, et cetera, companies into the smartphone supply chain or others who have plants and facilities. Their ability to get parts has been affected. Can you talk about any situation that you're experiencing? When you reported back in late March, you anticipated some of these issues happening by carrying higher levels of raw materials. Just describe your own internal inventory situation, any supply chain issues that you're seeing?

So like we are not providing an update to the guidance here, but what Manish did say last week is that there have been no major surprises as it relates to the guidance. In terms of our overall inventory, we are operating with a little bit high level of inventory in the raw materials to deal with the uncertainties surrounding the supply chain during this COVID period. And then obviously, like the inventory will change depending on the demand and production. There will be some changes. Our expectation was that, last quarter, we were at 132 days, which is a little bit elevated compared to our normal 110 days of inventory. We would anticipate that, this quarter, inventory will also be elevated, and then it will start to decline in the second half of calendar year. In the year, we do have a high level of NAND inventory because of the RG transition that we are going through at the moment. We are targeting -- I should say, pre-COVID, we were targeting to ship about 30% bit growth for the year, for the calendar year. And that -- we were not getting a lot of bit growth from production. So the only way to meet that demand growth from our customers and to support them was to study inventory and use that to meet the demand growth. And then eventually, next year, we will get the second-generation replacement gate coming. So noninventory is elevated because of that dynamics around the replacement gate transition. And the DRAM inventory then, we talk about the server market being in shortages, so clearly, in that area, the inventory is -- were even below normal at the end of last quarter and were, for us, and whereas like maybe in some other markets where the demand has been weaker, our inventory was a little bit higher than the mean levels that we would like to operate at. So that's kind of like the -- just some more color on the inventory.

I see. And in terms of kind of pricing dynamics for NAND and DRAM there, last quarter, there was an uptick in the ASPs for NAND. DRAM was around flat. There's obviously a lot of puts and takes in terms of the demand in the end markets. Any thoughts in terms of kind of the pricing dynamics? And just when you did your update last -- a week ago, on the strategy side, you talked about, very interestingly, the structural profits have improved quite significantly despite all the cycles. And to me, that would imply some sort of pricing leverage given the fact that there's a tightness of supply, but also the demand is pretty strong to enable you to have structural profits. So just, number one, broadly, how do you look at your pricing kind of leverage and dynamics going forward? And then how has that changed in this environment?

Yes. So we don't really comment on pricing too much, but like to your point, on the company becoming more valuable to the overall ecosystem, I think that part is definitely true. We are providing products with a lot higher differentiation and pairing value for our customers to product innovation, through giving them choices on sourcing of their supply from different geographies and giving them more value-added products, all of which is translating into better profitability for us. And we are also -- at least relative to the industry. And we are also like definitely focused on advancing our technology, which can also allow us to differentiate from our competitors, also give us good cost structure. So overall, we are continuing to improve our profitability. And that update that you had mentioned, Manish did say that on the cost side, we will continue to have at least in line or better cost reductions than the rest of the industry, even going forward for both NAND and DRAM. And so that's a positive. And then also, we will continue to mix towards more value-added products. And we obviously have a lot of focus on making sure that we are being disciplined on pricing. We are not giving up pricing easily and trying to drive towards a greater discipline across the entire organization. That's a big focus area for Sumit and Sanjay and what they are trying to drive at the company. And I think those results have frankly translated into the significant improvement that the company has shown relative to the industry, right, like so the strategies that the company has taken on cost reduction being faster than the industry, a higher-value solution. And then our pricing strategies have all contributed to our structural profitability improvements that you talked about.

The cost reductions on the NAND side are really going to take effect when you transition to the second-gen RG node, which is starting in November of this year, I believe, November...

So we haven't quite said when but it will be like in next fiscal year.

Next fiscal year, which starts November...

Yes. It could be, like we have not really given a timing...

Okay. It's going to start next fiscal year. You'll migrate a significant amount of capacity to the second gen, and this was always going to be on a 128-layer and will be charge trapping.

Well, it's not 128 layers. It's a higher layer count.

It's a higher layer count? Okay. So...

Yes. The first generation is 128 layers. The second generation is a higher layer count. We have not talked about publicly on what that layer count is, but it's a significantly higher layer count than 128. So a big step-up.

So -- yes. And so -- and you did this in a different way, when you're transitioning from floating gate to charge trapping versus other memory producers for a variety of reasons, yield issues that are -- but you are going to be on that approach. So how do we think about the cost reductions because there seems to be a lot of NAND tailwinds that, in your business next fiscal year, when you move capacity over to this environment?

Yes. So when you think about the -- in terms of the nonbusiness fee, as we ramp our second-generation RG node and as that technology matures, we should start to get very good cost reductions. We should start to get good cost reductions. And we expect that like in the time frame that we will introduce this technology node, it should have a very competitive layer count. It should be competitive in terms of the number of layers with the best that's out there. We will also have the CMOS under the array architecture. And we should be, overall, in -- like in a very good cash cost position across the industry. We should be among the -- we should be very close to the best in the industry on a cash cost basis. So we feel good about the cost reductions overall, but like you have to get past this period where we go through the first generation of RG and really start to ramp the second generation.

But no comments in terms of kind of what the potential cost reduction could be given transition?

I would say like it's a good reduction, but I will just leave it at we don't comment on specific numbers because it's -- like it's -- it impacts our business and for competitive reasons, we would avoid giving that information on. But it will be good cost reduction.

So it would be fair to assume that on the -- for gross margins overall, that DRAM margins could continue to benefit from demand that you're seeing on the data center side and relatively healthy pricing. And then I think this is a high level. In NAND margins, ASPs might have not have been as high as we expected in the second half versus the first half going into the year, but are still not declining at the rate that we saw last year when there was a lot of excess NAND inventory. And then you have the NAND cost reductions coming in the next fiscal year, so the NAND margin should start to -- they should at least stabilize -- it should be fairly stable and then kind of potentially move up.

So I mean, like obviously, the biggest driver of margins are pricing, and we don't really comment on that. But apart from that, the other big driver is mix. And we are driving towards higher mix of high-value solutions, which is a positive for margins, and that should continue to grow. We have talked about mix of high-value solutions at about 70% of the overall bits. And that will continue to move higher towards 80% in fiscal '21. So first half of fiscal '20 was 70%, and it is going to 80% by fiscal '21. And that should be a positive driver. Also, we will continue to move within the high-value solutions towards a richer mix of enterprise and data center, which tends to be better profitability as well. So the mix will be a positive driver for margins, and then tightening that on like is a wildcard and anybody's guess.

And just last question on the mix. So when Sanjay took over, there was a concerted effort to kind of get into the SSD market, client SSDs. There was a strong effort to get into UFS and smartphones, where there's a higher margin. Last year, you didn't have the product for PCIe. And so have we -- and there was excess inventory, too. So you might have been lucky in that sense that there wasn't much of an adoption. But this year has been cleared out. We are moving to PCIe broadly, and I believe you do have a product for that. Is that going to help contribute to the higher value-added service in managed NAND? And then also, on the smartphone side, what's the thought process at Micron of trying to sell more higher-value UFS solutions?

And so we are growing our mix of high-value solutions, both in SSDs and in managed NAND mobile products, mobile managed NAND products. We -- in the SSD space, we introduced our NVMe products last year, and we've raised the product portfolio. Now the product portfolio is more complete, and we have products in all the major categories. A lot of those products are right now undergoing qualification. In the client space, we have actually started to ramp the revenues. And in the data center, we've also started to see some revenues, and we expect that both areas will continue to grow through rest of the year and be a growth driver for the company. Managed NAND, like in the managed NAND space, we talked about, even last quarter, with all the weakness in the mobile space, we have record revenues in the MCP for us, which is really a testament to how far the company has come along in terms of adoption of these high-value solutions product in the mobile space. I should mention that this Friday, Raj Talluri, who's our Senior Vice President of the Mobile Business Unit, he will be doing a webcast, along with Dave, to provide an overview of our mobile business and what we have been doing there. So you can get more details on how we have been doing in the UFS and MCP space and how we're progressing in that area. And like we have UFS products also. We are growing that. We are growing uMCPs. We are growing -- we are the first to market on LP5 products. We are the first to market with uMCP5 products as well. So like there are a lot of new technology and exciting products that we are introducing.

All right. Great. I think we're coming up to the end of our time. So I'll stop here. Thank you so much, Farhan. This was very insightful. Thank you, everyone, also for joining the call. And hopefully, we'll be able to see each other face-to-face in the not-too-distant future.

Thanks so much, Raji.

Thanks so much.

Thank you. Bye-bye.