Micron Technology, Inc.

Good afternoon. My name is Josh, and I will be your conference facilitator today. At this time, I would like to welcome everyone to Micron's third quarter 2021 financial release conference call. All lines have been placed on mute to prevent any background noise. After the speakers are marked, there will be a question and answer session. If you would like to ask a question during this time, please press star, then the number one on your telephone keypad. If you would like to withdraw your question, press the pound key. In order to allow as many analysts to ask a question as possible, please limit yourself to one question. Thank you. It is now my pleasure to turn the floor over to your host, Farhan Ahmad, Vice President of Investor Relations. You may begin your conference.

Thank you, and welcome to Micron Technologies' Fiscal Third Quarter 2021 Financial Conference Call. On the call with me today are Sanjay Mehrotra, President and CEO, and Dave Zinsner, Chief Financial Officer. Today's call will be approximately 60 minutes in length. This call, including the audio and slides, is also being webcast from our investor relations website at investors.micron.com. In addition, our website contains the earnings press release and the prepared remarks filed a short while ago. Today's discussion of financial results will be presented on a non-GAAP financial basis unless otherwise specified. A reconciliation of GAAP to non-GAAP financial measures may be found on our website. As a reminder, a webcast replay will be available on our website later today. We encourage you to monitor our website at micron.com throughout the quarter for the most current information on the company, including information on the various financial conferences that we will be attending. You can follow us at Twitter at MicronTech. As a reminder, the matters we will be discussing today include forward-looking statements. These forward-looking statements are subject to risks and uncertainties that may cause actual results to differ materially from the statements made today. We refer you to the Form 10-K and 10-Q that we filed with the SEC for a discussion of risks that may affect our future results. Although we believe that the expectations reflected in the forward-looking statements are reasonable, we cannot guarantee future results, levels of activity, performance, or achievements. We are under no duty to update any of the forward-looking statements after today's date to conform these statements to actual results. I'll now turn the call over to Sanjay.

Thank you, Farhan. Good afternoon, everyone. We delivered outstanding results in SU3. Our strong execution enabled us to achieve the largest sequential EPS improvement in our history and to set multiple revenue records. NAND hit record revenue propelled by record mobile MCP, consumer SSD, and client SSD revenues. Our embedded business exceeded $1 billion for the first time with record revenue across automotive and industrial markets. We also achieved key technology and product milestones with our industry-leading 1-alpha DRAM and 176-layer NAND, reaching a meaningful portion of our bid production and QLC NAND, accounting for a majority of our client SSD bid shipments. We expect DRAM and NAND supply to remain tight into calendar 2022 as the global economy rebounds. The strong demand for memory and storage across the data center, intelligent edge, and user devices puts Micron in the best position ever to fully capitalize on these exciting opportunities. We continue to make solid progress on our goals to deliver industry-leading technology and improve our cost structure, to bring differentiated products to market and improve our product mix, and to grow our share of industry profits while maintaining stable bid share. I will start with an update on our operations. Despite shortages across the semiconductor ecosystem in various assembly materials and assembly capacity, Micron delivered record assembly output this quarter, which helped fuel our strong revenue performance. Our assembly and test success was the result of a strategic decision we made several years ago to increase our captive footprint and strengthen relationships with suppliers and partners. We successfully mitigated the impacts of the drought in Taiwan with no reduction in our production output. Taiwan's rainy season has begun, bringing with it sufficient water supply to support our manufacturing requirements. While the drought in Taiwan is behind us, the rise in COVID-19 cases in Malaysia, India, and Taiwan are a risk to our manufacturing operations and R&D activities in these regions. We are also working with local governments to facilitate on-site testing and vaccination for Micron team members where possible. Additionally, in order to protect Micron team members at our Muar, Malaysia back-end facility, we temporarily reduced our on-site workforce early in FQ4, which reduced output levels. We have since started bringing back team members to the site as the situation has improved. While we ramp back toward full production levels in Moar, we will utilize our global supply chain, including subcontractor partners, to meet our customer commitments and minimize any disruption to delivery schedules. Earlier this year, we announced our decision to exit 3D cross-point development and manufacturing and to reprioritize our R&D investments toward new CXL-enabled memory solutions. CXL, or Compute Express Link, is a new industry standard interface that will significantly change data center architecture through high-performance connectivity between compute, memory, and storage. We are developing exciting CXL-enabled products and will have more to share on our roadmap in the future. As part of our exit from the 3D Crosspoint business, we announced our intent to sell our Lehigh Utah FAB. Today, I'm pleased to report that Micron has reached an agreement to sell the FAB to Texas Instruments in a transaction that we expect to close later this calendar year. We see this transaction as positive for our team members the Lehigh community, and our shareholders. The Lehigh site has been an important part of the Micron network and responsible for many technology and manufacturing innovations across NAND and 3D cross-point products. Texas Instruments will offer all Lehigh team members the opportunity to become TI employees at the Lehigh site upon closing. After the sale closes, we will be able to eliminate the remaining underloading costs we were incurring at Lehigh, enhancing our efficiency and strengthening our profitability. Dave will provide additional details. Now for an update on technology and products. Our industry-leading 1-alpha DRAM and 176-layer NAND process technologies are now in production and ramping according to plans. These nodes accounted for a meaningful portion of our bid production in FQ3 and are on track to become a meaningful portion of our revenue in FQ4. We expect that by end of calendar 2021, the combination of one alpha and one Z DRAM nodes will represent the majority of our DRAM bid production. And at the same time, 176 layered NAND will be the majority of our NAND bid production. In fiscal year 22, we expect these workhorse nodes to fuel bit growth and provide us with good front-end cost reduction on a like-for-like basis. However, there are two factors that will create cost headwinds for us next fiscal year. The first is driven by our strategic portfolio migration toward more advanced and higher-value products, such as DDR5 memory, high-density server modules, and SSDs. While this portfolio shift helps us increase profit share, it will also impact our costs next fiscal year. The second cost headwind is driven by several actions we have taken in our supply chain to increase resilience and provide business continuity to our customers across all product lines. While these actions will allow us to capitalize on robust market demand, they will also impact our costs. We are on track to support customers as they begin to introduce DDR5 enabled platforms in the second half of calendar 2021. DDR5 was designed to meet modern data center requirements, including improved performance through doubling of memory bandwidth and improved reliability and efficiency through integration of on-die ECC. DDR5 features a larger die size compared to DDR4. limiting DRAM industry supply growth and cost reduction as it ramps, starting from the second half of calendar 2021. In storage, we introduced the industry's first UFS 3.1 solution for automotive applications this quarter. We also announced volume production of client PCIe Gen 4 SSDs built on the world's first 176-layer NAND, and available in a variety of form factors. We are delivering 176-layer NAND in volume to OEM and channel customers across multiple markets and have several products in customer qualifications. We are also driving an increased mix of QLC NAND, which brings down the cost of SSDs, accelerating the replacement of hard drives. QLC SSD adoption continues to grow and we delivered record QLC SSD revenue and BitMix in FQ3. Turning to end markets, in the data center, integration of AI into data-centric workloads will drive long-term growth, with memory and storage becoming an increasing portion of server bomb cost. Propelled by the transition to DDR5, strong capabilities in graphics memory and the introduction of HBM and NVMe SSD product offerings, Micron's strong product roadmap across DRAM and NAND solutions positions us for success in the data center. We will enhance our NVMe SSD portfolio with the introduction of new products with internally designed controllers in the coming months. In the fiscal third quarter, data center DRAM revenue grew quarter over quarter driven by strong demand from cloud customers and increases in module density. Data center SSD bit shipments and revenue grew sequentially, driven by both cloud and enterprise. Data center demand is expected to be strong in the second half of the calendar year as cloud demand picks up and enterprise demand improves due to broad economic recovery. In addition, we expect that the new CPUs featuring more memory channels will accelerate server memory demand starting later this year and continuing into 2022. The PC market continues to benefit from the trend toward greater mobility as people embrace a work or learn from anywhere culture. Industry expectations for calendar 2021 PC unit demand growth have increased to the high teens, driven by robust notebook sales and the recovery in the desktop market. In the fiscal third quarter, we achieved several customer qualifications for our OneAlpha-based DDR4 products across various PC platforms. Our client SSD bit shipments were up sharply quarter over quarter and year over year. In graphics, bit shipments increased sequentially and year over year, driven by strong next-generation game console and graphics card shipments. Micron has an excellent position in the graphics market with a broad product portfolio and deep customer partnerships. Mobile business achieved record MCP quarterly revenue. We made strong progress with our 1-alpha LP DRAM products and 176-layer UFS 3.1-enabled solutions. We have already completed customer qualifications for some of these products. While COVID-19 has softened mobile demand in parts of Asia, supply shifts to address stronger demand in other regions are keeping the global market in tight supply-demand balance. Mobile unit sales are expected to show healthy growth this year with some variability across geographies, driven by an expected doubling of 5G units in calendar 2021 to more than 500 million units. These 5G phones also feature rich content demanding significantly higher DRAM and NAND. We are also encouraged to see bold OEM innovation in new devices like gaming smartphones featuring 18 gigabyte of DRAM. Our automotive business delivered a third consecutive record quarter, driven by continued manufacturing recovery and increased LPDDR4 and EMMC content for in-vehicle infotainment and driver assistance applications. Auto unit sales are expected to grow significantly from last year. Auto memory and storage content growth trends remain strong, particularly as EVs, which have significantly higher memory and storage content requirements, grow much faster than the broader auto market. We are continuing to see record automotive and industrial segment demand, Yet, despite our best efforts, we may be unable to meet all the demand from these customers over the next few months due to certain non-memory semiconductor component shortages in our supply chain. Turning to our market outlook, while the pandemic remains a risk factor, calendar 2021 is shaping up to be a strong year fueled by the macroeconomic recovery combined with secular drivers such as AI and 5G, that are creating sustained demand increases across broad end markets. As a result, our expectations for calendar 2021 DRAM and NAND bid growth have increased since our last earnings call, and we now expect calendar 2021 DRAM bid demand growth to be somewhat above 20% and NAND bid demand growth in the mid 30% range. There is currently unmet demand for DRAM and NAND due to end market strength. This unmet demand would have been even larger had it not been for the non-memory component shortages influencing our customers' ability to manufacture their products, particularly in the PC, automotive, and industrial markets. These shortages can cause variability in demand patterns as customers experience challenges sourcing matched set of non-memory components. We are hopeful that foundry capacity coming online can begin to alleviate some of the component shortages in the second half of calendar 2021 and support robust memory and storage growth. Additionally, as a result of strong end market demand trends, the lessons of the pandemic and ongoing geopolitical uncertainty, some customers will change their inventory management strategy from just-in-time to just-in-case and increase the target level of what they consider normal inventory levels. Long-term, we see a DRAM bid demand growth CAGR of mid to high teens and a NAND bid demand growth CAGR of approximately 30%. Turning to Micron Supply, we are targeting to align our long-term bid supply growth CAGR with the industry bid demand growth CAGR across DRAM and NAND. However, we expect year-to-year variability caused by no transition timing. In both DRAM and NAND, we expect our calendar 2021 bid supply growth to be below the industry bid demand growth, and we have used our inventory to add to our bid shipment growth this year. Before handing over to Dave, I have one more important announcement to share regarding our DRAM technology and manufacturing strategy. Based on our assessment of the progress EUV has been making and aligned with our technology strategy and industry-leading DRAM scaling roadmap, we plan to insert EUV into our DRAM roadmap starting in the 2024 timeframe. Micron has placed purchase orders for multiple EUV tools from ASML as part of a long-term volume agreement. The prepayments for these systems will contribute towards the fiscal year 21 and fiscal year 22 capex. We have increased our fiscal year 21 capex to be somewhat above $9.5 billion, mostly from areas that do not impact calendar year 21 and calendar year 22 bid growth, such as these EUV prepayments, construction spending, and other R&D and corporate items. I will now turn it over to Dave.

Thanks, Sanjay. Micron delivered outstanding FQ3 results. Revenue in EPS grew by a record amount sequentially on an organic basis, and we generated over $1.5 billion in free cash flow in the quarter. Total FQ3 revenue was approximately $7.4 billion, up 19% quarter-over-quarter and up 36% year-over-year. Revenue growth was driven by stronger DRAM and NAN pricing, and by robust customer demand for Micron's products. FQ3 DRAM revenue was $5.4 billion, representing 73 percent of total revenue. DRAM revenue increased 23 percent sequentially and was up 52 percent year-over-year. Bid shipments increased in the low single-digit range sequentially, and ASPs were up approximately 20 percent quarter-over-quarter. FQ3 NAND revenue was approximately $1.8 billion, representing 24 percent of total revenue and an all-time high for the company. NAND revenue increased 10 percent sequentially and was up 9 percent year-over-year. BIT shipments increased by low single-digit sequentially, while ASPs increased in the high single-digit percentage range quarter-over-quarter. Now, turning to our revenue trends by business unit. Revenue for the compute and networking business unit was approximately $3.3 billion, up approximately 25 percent sequentially and 49 percent year-over-year. CNBU revenue growth was driven by broad-based sequential pricing increases. Revenue for the mobile business unit was $2 billion, up 10 percent sequentially and 31 percent year-over-year. Mobile demand remained healthy as 5G handset sales continued to ramp. Revenue for the storage business unit was $1 billion, up approximately 19% from the prior quarter, and approximately flat year over year. Both client and consumer SSD revenues set records. And finally, the embedded business unit generated record revenue of $1.1 billion, which was up 18% sequentially and 64% year over year. Automotive and industrial revenues were at an all-time high for the company. The consolidated gross margin for FQ3 was 42.9%, up 10 percentage points from the prior quarter. DRAM and NAND price increases helped drive the margin expansion in FQ3. Gross margins also benefited by 100 basis points from $75 million less depreciation at our Lehigh FAB, which is classified as assets held for sale. Operating expenses were $821 million in FQ3, which we continue to tightly manage. Operating expenses also benefited from approximately $21 million of gains from the sales of certain assets. FQ3 operating income was $2.4 billion, resulting in an operating margin of 32% compared to 20% in the prior quarter and 18% in the prior year's quarter. FQ3 EBITDA was $4 billion, resulting in an EBITDA margin of 53% compared to 45% in the prior quarter and 44 percent in the prior year. Net interest expense was $31 million in FQ3, and we expect it to be roughly flat going forward. Our FQ3 effective tax rate was 8.4 percent. We expect our tax rate to be in the high single digits for FQ4. Non-GAAP earnings per share in FQ3 were $1.88, up from 98 cents in FQ2, the 90-cent sequential improvement was the largest in Micron's history. EPS included approximately 5 cents from the sale of certain assets, investment gains from Micron ventures, and one-time tax items. Turning to cash flows and capital spending, we generated approximately $3.6 billion in cash from operations in FQ3, representing 48 percent of revenue. Net capital spending was approximately $2 billion during the quarter. As Sanjay mentioned, we now expect our FY21 capital spending to be somewhat higher than $9.5 billion. Most of this capex increase that we are highlighting today will not increase our CY21 and CY22 bid supply. We expect that while we invest in the EUV infrastructure and initial deployment, our capital intensity will increase to mid-30 percent of revenues. Once we get past the investment period of EUV adoption, we expect that these tools will boost our competitiveness and help drive productivity of our fabs. As a result of the strong market environment and Micron's extraordinary execution, we generated positive free cash flow of $1.5 billion in FQ3. The increased cash flow was driven by strong revenue growth, higher margins, and efficient working capital management. We expect free cash flow to continue to improve in the fourth quarter, driven by continuing growth in revenue and earnings. We completed share repurchases of $150 million, or approximately 1.7 million shares in FQ3. From the inception of the share repurchase program, we've repurchased $3 billion worth of Micron stock, representing 55 percent of our cumulative free cash flow. In addition, since FY19, we have used approximately $2 billion in cash to settle conversions of our convertible notes, including approximately $800 million to settle the convert premiums. Combining the share repurchases and convert premiums, we've used $3.8 billion, or 69% of our cumulative free cash flow, towards reducing our share count. We plan to continue repurchasing shares in FQ4. Ending FQ3 inventory was $4.5 billion, or 98 days. We remain in a very lean inventory position as demand continues to outstrip our supply. We ended the quarter with total cash and investments of $9.8 billion and total liquidity of approximately $12.3 billion. FQ3 ending total debt was $6.7 billion. Our balance sheet is rock solid with investment grade ratings from all three rating agencies. In the last three months, Fitch and Standard & Poor's both raised their outlook from stable to positive for Micron debt. These upgrades to the outlook for our debt ratings are further evidence of the financial transformation underway at Micron. Before providing the financial outlook, I want to cover the financial implications of the sale of our Lehigh FAB. We are pleased with this transaction and believe that it is good for our shareholders as it frees up capital and enhances our ongoing profitability. The economic value for Micron from the sale is $1.5 billion, comprised of $900 million in cash resulting from the sales transaction, and approximately $600 million in value for select tools and other assets that Micron will retain for redeployment to its other manufacturing sites or that are sold to other buyers. We're taking an impairment charge of $435 million, or approximately $330 million on an after-tax basis, as the $900 million sale price is below our book value of the assets being sold. Note that the tools that we are keeping have largely been depreciated, but have substantial future value in our manufacturing network. As we previously disclosed, we stopped depreciation of the Lehigh FAB assets last quarter, and this benefited our costs by approximately $75 million in FQ3. Once the sale is completed, we will further improve our profitability by entirely eliminating our underload charges. Now turning to our near-term outlook, both DRAM and NAND markets are tight, and we expect pricing increases for both markets in the fiscal fourth quarter. In FQ4, we're qualifying 1-alpha and 176-layer nodes with several customers. We expect these nodes to support a modest level of bid growth and face cost headwinds that are common at this stage of the ramp. Additionally, we also expect cost headwinds from product mix and COVID mitigation. Despite cost headwinds, we expect strong improvement in our financial performance in FQ4. Our growth opportunity is healthy and market momentum heading into fiscal year 2022 is strong. With all these factors in mind, our non-GAAP guidance for FQ4 is as follows. We expect revenue to be $8.2 billion, plus or minus $200 million, gross margin to be in the range of 47% plus or minus 100 basis points, and operating expenses to be approximately $900 million plus or minus $25 million. Finally, based on a share count of approximately 1.15 billion fully diluted shares, we expect EPS to be $2.30 plus or minus 10 cents. Micron's relentless focus on execution positions us well to generate solid returns for our shareholders. Measuring our performance trough to trough across the cycle from FY16 to FY20, we substantially improved our EBITDA margin and our revenue grew by more than 70%. During this time, we delivered average gross margins of 40%, EBITDA margins of 50%, and return on invested capital of 20%. We believe Micron's strong financial performance will continue cross-cycle, and over the long term, our revenue growth will outperform the broader semiconductor industry. Our industry-leading technology, dramatically improved product portfolio, and financial strength position us well to capitalize on the long-running demand trends driving the memory and storage industry. I will now turn it back to Sanjay.

Thank you, Dave. Micron's fiscal third quarter results demonstrate the strength of our business and we expect to achieve continued strong results in the future. Demand for memory and storage is solid across market segments, and industry trends like artificial intelligence, edge computing, and 5G continue to create new opportunities for Micron. Our team is building on our technology leadership to deliver bold new solutions that offer valuable differentiation for our customers. Micron's business is healthier and more robust than ever and we are energized to seize the opportunities ahead at a truly exciting time in the semiconductor industry. We are also leveraging our success to deliver results for all our stakeholders. In April, we released our sixth annual sustainability report, highlighting progress towards our environmental, social, and governance goals. I'm pleased to report that we are on track to achieve the environmental and sustainability goals we set last year despite the challenges posed by the pandemic. In fact, our ESG risk scores have improved to the top 10% of the semiconductor industry, according to the third-party rating agency Sustainalytics. We are also making good progress on achieving 100% renewable energy consumption in the U.S. by the end of 2025. In calendar 2021, we continue to focus on emissions abatement, transition to renewable sources, water restoration, and increased efforts to reduce, reuse, or recycle waste. We will pursue these goals with the same focus with which we have created sustained momentum in the business, and I look forward to providing updates on our progress on future calls. We will now open for questions.

Thank you. As a reminder, to ask a question, you'll need to press star 1 on your telephone. To withdraw your question, press the pound key. Please stand by while we compile the Q&A roster. Our first question comes from CJ Muse with Evercore. You may proceed with your question.

Yeah, good afternoon. Thank you for taking the question. I guess, you know, and market demand question, you know, there's clearly fears out there around PCs peaking, volatility around handsets, and whether, you know, there's any inventory build on the cloud side yet. You know, here you're talking about DRAM and NAND remaining tight into calendar 22. So I guess can you walk through what you're seeing out there from a demand perspective? And then also, I think very importantly, particular to the DRAM side, you know, how you're thinking about supply, which clearly, you know, seems to be constrained both this year and next year.

Thanks, CJ. So on the demand side, we certainly see strong demand across almost all N markets. PC, on a year-over-year basis in calendar year 21, the growth is in high teens. And, of course, the SSD, attach rate, average content continues to increase on the NAND side, and PC continues to drive healthy demand for DRAM as well. Our data center, after the digestion period earlier in the year, in the second half, driving strong demand for us as well. Smartphone, 5G trends driving unit sales as well as average content growth Automotive, we can't meet the supply. We can't meet the demand that is strong there. Industrial markets, the demand is strong. So across the board, almost across all end markets, we are seeing strong demand. In fact, in the industry, there is unmet demand. And you know that there is semiconductor shortage across the technology ecosystem. And as that semiconductor shortage gets alleviated over time, that actually is going to create more demand for memory and storage because every end application today, you know, whether it's analog IC related or memory CPU cores related, all of them actually require memory and storage. So semiconductor shortage, which is actually impacting some of the demand, as that gets alleviated over the course of next several quarters, That, too, will bring about increased demand. So, demand trends are strong. The supply, as we see through the year, through the end of the year and into calendar year 2022, is tight as well. And you know that CapEx in the industry has been, on the DRAM side, extremely disciplined. The producer inventories, the supplier inventories are running extremely lean as well. I can certainly speak for our inventory. days of inventory are at 98, you know, extremely low as well. And capital intensity is increasing as well in the industry, and that all bodes well for discipline supply growth as well. We talked about that how DDR5 as a spec that is there in the industry as a JEDEC spec actually requires more on-chip ECC that results in bigger die sizes for everybody. in the industry for DDR5 over DDR4. So that again, as you can well understand, as the industry transitions to DDR5 over the course of next several quarters in 22 as well as in 23, that too means less supply growth availability from the wafers even with technology transitions. And so all of those trends from the demand side as well as from the supply side bode well for our industry.

That's very helpful. If I could follow up on the gross margin side, Dave. You talked in the prepared remarks around higher cost mix and investment in the supply chain. Can you walk through, you know, the moving parts there for fiscal 22 gross margins?

So, obviously, you know, one of the bigger components of the margins for fiscal 22 is going to be around pricing. And, you know, we don't provide pricing beyond the next fiscal or beyond this quarter. other than to say that we think pricing will be up next quarter, and we are suggesting that it will be tight at least into 22. So that's as much as I can give you on the pricing side. On the cost front, when you look at the front-end cost reductions that we'll see next year on a like-for-like basis, driven by, as Sanjay mentioned in the prepared remarks, the ramp of 1 alpha and the ramp of 176 on the NAN front, we do feel like those costs will be good. The only counter to that is we will see a higher mix of products that carry higher costs. Sanjay mentioned like DDR5, higher density server modules, more SSDs, all those things will be a bit of a headwind on the cost front. And we will likely go into the year with some COVID mitigation costs. That also will be a bit of a headwind. Now, hopefully that over the course of the year alleviates and that starts to you know, help on the margin front. The only other factor is we will have a little bit of a lift in Q4 from Lehigh as the full amount of the depreciation goes away in the fiscal fourth quarter. And then once we close on the sale, all the underload charges will also go away. The way I think I'd model it is, you know, maybe about $20 million of benefit in the fourth quarter and probably another $20 million in the first fiscal quarter, and assuming we close somewhere close towards the end of the first fiscal quarter, that should be behind us.

Thank you.

Thank you. Our next question comes from John Pitzer with Credit Suisse. He may proceed with your question.

Yeah, guys, just two quick questions. Dave, maybe to follow up on CJ's questions about cost, I want to make sure I understand the messaging here. I get that these higher value-added parts have higher costs, but shouldn't they also have higher gross margins, or am I thinking about that incorrectly, and then I have to follow up?

You know, it somewhat depends on the product itself, but I would say in general, you know, we are trying to drive towards higher value products, which, you know, arguably on a like-for-like basis, or at least on a comparable basis to other products, would carry better gross margins.

Perfect. And then as my follow-up, you know, two quarters ago, Dave, you didn't buy back any stock. This quarter it was $150 million, which was, I think, 10% of free cash flow. Notwithstanding what you've done over multiple quarters, I'm just kind of curious as to the message you're trying to give us here, especially if you look at sort of the cross-cycle risk-reward approach in the stock, why not be more aggressive with the buyback here? And does that portend something about next fiscal year's CapEx? And as you talk about that, CapEx, is next fiscal year a year that you should outgrow BIT relative to industry?

Okay, a lot of questions on that. Okay, so on the buyback, I wouldn't read anything into the $150 million. Some quarters will have higher levels of buyback than others. We did have an eye on our net cash position. That was something that we were trying to move in the right direction. I think you'll find in the fourth fiscal quarter that our buybacks are, you know, meaningfully higher than our third fiscal quarter. So nothing to read there. We do feel like this price is obviously a good price to be buying the stock back. And we are committed to, you know, what we've talked about previously, which is to return at least 50% of our free cash flow in the form of buybacks. And as I talked about on the prepare remarks, I think the metric we've hit so far is 55%, so we've done pretty well. And that doesn't even account for the converts, which I think by the end of this fiscal quarter, that is the fourth fiscal quarter, will be completely done with converts. Converts will be completely off the balance sheet, and we will remove that dilution as well. As it relates, no message on the CapEx as it relates to buybacks, other than to say that... You know, given the EUV investments, it does appear that we will operate maybe at a little bit of a different level from a percent of sales perspective than perhaps we previously were operating. We were thinking more in the low 30s. I think with EUV, it's safe to say that we probably are operating in the mid-30s as a percent of revenue for CapEx, at least as we, you know, build out the EUV part of the tool set.

Perfect. Thank you very much.

Thank you. Our next question comes from Shannon Cross with Cross Research. Can we proceed with your question?

Thank you very much. I had a question about DRAM ASPs. The 20% quarter-over-quarter growth was the highest in several years. So can you talk a bit about the drivers of the growth? How much of it was like-for-like price increases given the current tight supply versus, say, benefit from mix? And how should we think about sustainability?

So certainly on a like-for-like basis, pricing increased across the board in the DRAM industry and, again, driven by the strong demand, as I mentioned earlier, pretty much across all of the end markets. So we enjoyed price increases across all end markets. And, you know, as we have mentioned, that even in FQ4, we see price increases not just in DRAM, but we also see that in NAND.

Okay. And then I guess given the comments you made about customers moving to just-in-case inventory management, can you unpack that a little bit just in terms of magnitude? I mean, is this sort of a one-off conversation you're having with people as they're dealing with, you know, the supply issues that are out there right now, or do you think this is something that's going to be sort of a meaningful transition within the industry? Thanks.

So we see it as an emerging trend in the industry. When you think about it over the course of the last couple of years or even maybe a somewhat longer time frame, there have been challenges with respect to geopolitical considerations. Certainly COVID brought into stark relief the need for a resilient, flexible supply chain. And when you look at all the acceleration of the digital transformation and the surge in demand that has occurred, And on top of it, impose semiconductor industry shortages that are leaving a lot of the unmet demand across multiple industries here. All of that is really leading the customer ecosystem as well as us, the suppliers, to really absolutely prepare for supply chain so that we can meet the demand. I mean, Micron itself has taken actions in this regard in terms of securing supply capacity, for example, for assembly operations, and that has really enabled us. For example, when our Muar operation, we had to bring down our team members there because of a recent COVID outbreak in Malaysia. Because we had made changes to our capacity, assembly capacity footprint, we had secured more external supply, assembly capacity. That enabled us to quickly shift our production to other parts of our manufacturing footprint. These are the kind of considerations that customers in general and suppliers in general are considering to make sure that they're able to manage their supply chains to be able to meet the end customer demand. And that's why certainly some of the just-in-time aspects of inventory management have proved to be costly over the course of last few quarters, particularly as the world has struggle to respond to the needs during the COVID timeframe. And yes, there is an emerging trend toward considering just in case, you know, whether it is related to geopolitical considerations, whether it is related to acts of God, you know, that can result in supply chain disruptions or just responding from challenges of COVID. So this is an emerging trend. And some of the customer's may have already reacted faster in terms of building stronger inventory positions. Other customers perhaps still scrambling to meet the requirements, but this is definitely a trend that we think will likely persist with companies as they think about their own supply continuity considerations in the future, just like Micron itself has taken the steps necessary to address its own customer requirements and fulfilling their demands.

Great. Thank you.

Thank you. Our next question comes from Timothy R. Curie with UBS. You may proceed with your question.

Thanks a lot. I had two questions, first on EUV and then on cost down. So, Sanjay, I guess the first question on EUV is sort of what's changed on EUV? I mean, it's not like there's been a sea change in, you know, progress made on EUV? Is it simply maybe that there's another big chip maker trying to get in the queue and taking up some slots, and so you felt like you had to get in the queue? So I'm just sort of curious what changed on EUV, and then I had a follow-up.

So we had always said that, you know, we monitor EUV progress. We have actually engaged in EUV evaluation. We have had EUV tool in the past. So we had always said that we will intercept EUV in our roadmap at the right time when we see the EUV platform as well as the ecosystem becomes more mature. That's when we plan to intercept EUV in our roadmap. And that's what our plan is that in the 2024 timeframe, and again, aligned with our technology and leadership DRAM scaling roadmap that we'll be implementing this in 2024. So it's consistent with how we have always approached it, and of course, EUV has continued to make good progress, and we really think that with our EUV technology capability from 2024 onward timeframe, coupled with our multi-patterning expertise that Micron has the leadership in the industry, we really will have unique, differentiated and absolutely feel confident about continuing to lead our DRAM scaling roadmap through our, of course, currently one alpha, but then one beta, and then one gamma, and beyond. And initially, we will deploy EUV in limited layer count in 2024 timeframe with our one gamma node, and then we will broaden it to the one delta node with greater layer adoption. And just keep in mind that we will combine it with our immersion multi-patterning techniques as well. And so we really believe that we will have a very strong roadmap, and this is pretty much along the lines of how we always intended to insert EUV in our roadmap in the future, basically keeping in track of cost effectiveness, productivity, as well as our overall scaling roadmap, and we feel really good about our leadership DRAM scaling roadmap ahead.

Thanks, Sanjay. I guess, Dave, my follow-ups on cost downs, you sound a little more negative or a little more cautious on your fiscal 22 cost downs than you were last quarter. I think This year in DRAM, you're going to be close to roughly 10% this fiscal year. And I think OneAlpha was supposed to help you next year, so the feeling was that you could do better than 10% next year in fiscal 22. But it sounds like maybe some of these mixed issues are going to result in you doing worse next year than you did this year. Can you sort of give us what next year is relative to sort of what you've done this year? Thanks.

So I don't think I'm ready to – we haven't completely – finish the plans on next year, so maybe it's a little premature for me to talk about next year specifically on cost downs. I would say that when you look at 1-alpha's cost declines, they are very good. You know, the timing in which, you know, the 1-alpha ramps is certainly an impact, and when it gets to its mature yield state is certainly an impact. And then, of course, it's hard to call these mixed elements that drive some headwinds, But suffice it to say, when you kind of think about what our strategy is, we do feel that we'll see many of these things enter into the equation. So I think when you look at it on a front-end basis, it's quite good and quite comparable. I think when you look at it on a mixed basis, somewhat dependent on how the market unfolds, but based on our early view into next year in terms of next, we'd expect some headwinds.

Okay, Dave, thanks.

Thank you. Our next question comes from Joe Moore with Morgan Stanley. You may proceed with your question.

Great. Thank you. I wanted to follow up on the just in time to just in case inventory question. You know, you're talking about demand not being fulfilled in the short term. So is the message here that the customers don't really have inventory, but that they want to put that inventory into place? Or are there pockets where there's inventory kind of waiting for other components?

So again, this really varies from customer to customers. Some customers may have reacted fast and would be carrying adequate level of inventory or inventory in line with their strategy in terms of how to cope with the current environment with respect to demand and supply for their own components. Whereas some other customers may have less level of inventory. So really it varies from customer to customer. So what I'm saying is that regarding the just-in-time, shifting toward just-in-case kind of mindset, it really is that customers focus on managing their supply chain so that they can have sufficient inventory to meet their end market requirements. Some customers may have moved more in that direction, and some other customers may have yet to move in the direction of from, you know, just-in-time mindset toward just-in-case. So, for example, the car production, we have seen that auto market has suffered through significant supply chain shortages and, of course, have incurred significant costs to that industry as well in not being able to fulfill all their supply requirements. And, of course, that then drives a different mindset on how to avoid this kind of situation in the future. It varies from end market to end market. It varies from customer to customer. But overall, what we are saying is that with the lessons of the geopolitical considerations, with the lessons of the pandemic, and the lessons of the recent supply chain shortages, in the backdrop of digital transformation requiring more and more of semiconductor solutions, customer ecosystem, parts of the customer ecosystem likely is approaching their inventory considerations in a different manner compared to before. And, again, we look at it as an emerging trend in the industry.

Great. Thank you very much.

Thank you. Our next question comes from Chris Daly with Citi. You may proceed with your questions.

Hey, thanks, guys. Just to follow up on that previous question, if you look at the three main end markets for DRAM, PC, cell phone, server, and your, I guess, your best guess of inventory in each, where would you say it's lowest? And then when do you think that those end markets will achieve their, you know, whatever the heck normal is these days level of inventory?

So we're not going to go there in terms of trying to break it down by market by market. You know, of course, you sometimes see different moving parts in different parts of the market. You know, for example, in mobile, you saw that with the India COVID situation, as well as April and May in China, you know, there was a reduction in demand in certain parts of the smartphone market. However, in other parts of the world, you know, the smartphone suppliers moved to supply the increased demand in the other parts of the world. And, of course, you know, some of the demand, because supply is in shortage, you know, some of the supply in the industry got shifted toward other parts of the market, too. So we're not going to break it down. I mean, I gave you mobile just as one example. And, you know, this situation can vary from customer to customer. But all in all, you know, when you look at the end markets, almost all end markets are seeing shortages. And in aggregate, there is. tight supply today. That's what is resulting in increase in prices in the industry that we reported for FQ3. And we guided to an FQ4 also for DRAM and NAND. We see price increases. And overall, we see supply tightness continuing through the year and into 2024 timeframe as well. Well said. I meant 2022. I like 2024 as well.

2024 is great, but 2026, I'll take it.

We'll definitely be talking about that one of these days, too.

Yes. One quick one, Sanjay. What do you think is going to be the chipset impact to Micron and just the memory ecosystem in general?

So I think when you say chipset impact, oh, chipset, I see. Okay. So with respect to chipset, we definitely – First of all, you know, it's really great that U.S. government is recognizing the importance of semiconductors and how important semiconductors are to national economic consideration as well as national security considerations. And, of course, semiconductors are important to all global economies today. So we are certainly, you know, look forward. to greater support for U.S. leadership in semiconductor research as well as semiconductor manufacturing in the years to come. And, of course, Micron, as the only player in semiconductor memory and semiconductor storage in the industry, is well engaged with the U.S. government. And I know that the U.S. government also recognizes the importance of memory and storage as a strategic part of the semiconductor industry. So we really look forward to opportunities in terms of addressing our future needs. We continue to stay engaged. We stay engaged with the governments in all global sites where we have major operations. And we look forward to the opportunities here in the U.S. as well. And we are really glad that the funding has crossed the finish line in the Senate. and we certainly hope that in the House as well this will pass and U.S. industry can get on with the business of really strengthening U.S. leadership in research and manufacturing in semiconductors for the years to come. And we definitely remain always committed to growing our own supply in line with the industry demand, and we remain disciplined in that regard.

Thanks, Sanjay.

Thank you. Our next question goes from Toshi Ahari with Golden Facts. You may proceed with your question.

Hi, guys. Thanks a lot for taking my questions. I have one on DRAM and one on NAND. On the DRAM side, I wanted to ask about your ability to grow bits over the next, call it, four to six quarters. I think, Dave, at a couple of conferences, you talked about bits being flattish into the August quarter. just given where you are in the transition and given low inventories. But as you progress and sort of transition to one alpha, at what point should we expect your bit supply to accelerate in the DRAM business? And to the extent you can't meet demand, call it over the next couple of quarters, how should we think about your willingness to increase capacity in DRAM? And then on the NAND side, at a very high level, I think, Sanjay, to me, it feels like you sound a little bit better on NAND supply-demand or less cautious on NAND supply-demand. I'm curious what's changed over the past couple of quarters. Is it purely demand being better? Is it sort of the shortages around controllers, yields on higher layer count nodes, or all of the above? Just curious what's changed in NAND over the past – couple of months, couple of quarters. Thank you.

Okay, so I'll take the DRAM question first. I think you would model, for sure, we are thinking pretty modest sequential growth in the fourth quarter in terms of DRAM. I think that will carry into the first fiscal quarter, quite honestly. I would expect a relatively gradual increase as we ramp 1-alpha and that there wouldn't be necessarily an inflection point where we see a big step up in the growth rates. We've been very focused on the supply-demand balance from our perspective, and so we have been investing in OneAlpha with that in mind. And then just the follow-on question you had, I almost say the same thing. As we look at DRAM, and actually as we invest in DRAM and NAND, we take a long-range view in terms of the growth rates of DRAM and NAND. Sanjay mentioned that We think DRAM growth rates are in – for DRAM long-term growth rates are in the mid to high teens, and we think NAND is growing – you know, should grow around 30% over the long term, and that's how we invest our CapEx. And we're all – now, year to year, things might be a little different than that, but we're investing over the long run to grow our supply in relationship to that demand growth, and we have not deviated from that strategy. And then on the NAND front.

And on the NAND front, yes, as you noted, that we have increased our outlook in terms of year-over-year NAND industry growth to now mid-30s. At the prior discussion, NAND industry was somewhat in oversupply. What we have seen is that NAND certainly has stabilized and the trends have improved. In fact, we talked about price increases that we experienced in FQ3 for NAND, as well as have guided to price increase in NAND in FQ4 as well. So overall, we see tightness in NAND as well through the remainder of this calendar year and into 2022. And NAND demand is being driven by elasticity and certainly, you know, continuing strength in PCs and also data center and smartphone markets as well. Overall, you know, our outlook has changed because the supplied inventories we believe are healthier, and, you know, certainly micron inventory in NAND also is running lean. And certainly our 176-layer NAND industry-leading node is ramping well, and overall we expect our long-term supply growth category to be in line with the market there as well. So...

Thank you.

Thank you and that concludes today's conference call. Thank you for participating. You may now disconnect.