Micron Technology (MU.US) released its financial results for the fourth quarter of fiscal year 2026 after the U.S. market close on September 30, Eastern Time. Management struck an optimistic tone on the earnings call, stating that AI-driven memory demand remains strong, supply and demand will stay tight in 2027 and 2028, and the company's confidence in long-term growth has strengthened.
Management said that more than 75% of fiscal 2027 shipments are already locked in, Strategic Customer Agreements (SCAs) cover about 35% of sales through 2030, HBM is growing faster than conventional DRAM, and 2027 pricing has been raised significantly. On the supply side, constraints include cleanroom construction, diminishing returns from technology transitions and a rising HBM trade ratio, and the company does not yet see a point at which supply and demand will balance.
The core message from management is that AI-driven memory demand is far from reaching supply-demand equilibrium, the market will remain tight in 2027 and 2028, and 2028 could even be tighter than previously expected. President and Chief Operating Officer Manish Bhatia said server units will continue to grow into 2027, and the rapid rise of agentic AI is also generating CPU-driven memory demand. The company has locked in more than 75% of fiscal 2027 shipments and has extended customer allocation negotiations into 2028.
On the supply side, structural constraints include diminishing returns from technology transitions, HBM growing faster than conventional DRAM, a rising HBM trade ratio and long cleanroom construction cycles, which is why the company "cannot see when supply and demand will balance."
HBM remains the focus. Management said HBM shipment growth will outpace conventional DRAM, and its share of industry capacity will continue to rise through 2028. Calendar 2027 HBM pricing has been raised significantly and will be reset early in the year to narrow the profitability gap with conventional DRAM. On HBM market share, the company is not chasing a single number but aims to roughly maintain a share comparable to its overall DRAM share. On HBM4E, Scott DeBoer said the company has been co-designing NV HBM with Nvidia for more than a year, and this will be the first important custom HBM product, with differentiation in power consumption, speed and product quality, expected to deliver high value and strong ROI.
On Strategic Customer Agreements (SCAs), the company has signed 26 SCAs covering about 35% of sales through 2030, spanning DRAM and NAND, with DRAM volume slightly below 35% and NAND bits slightly above. The 10 newly signed customers range from small to large and cover all business units. About three-quarters of SCA revenue has a clear pricing framework, while one-quarter is open to periodic negotiation or market-based pricing, with most having floors and ceilings, but new agreements have been repriced according to current market conditions and expectations of future tightness.
On capital expenditure, CFO Mark Murphy said the increase in fiscal 2027 construction capex is mainly directed at cleanrooms coming online at the end of 2028 and beyond to accelerate capacity space, but the spending will not immediately translate into bits. The company will flexibly equip capacity based on demand trends and equipment supply agreements. On China, management said exposure continues to decline and is expected to fall to single digits in fiscal 2027. The company leads Chinese competitors by at least two nodes in technology. 1-gamma DRAM already accounts for most bits, 1-delta will enter mass production in the second half of next year, and EUV is a key differentiator.
For mobile and client, management acknowledged that bit shipments declined sequentially, but revenue grew on improved pricing and mix, with strong demand in high-end client and flagship smartphones. On NAND, the industry is expected to grow around the mid-20% range in 2027-2028, with the market remaining tight. Singapore cleanroom space is being used for advanced R&D and HBM pilot operations, causing supply growth this year to lag the industry, but the G9 ramp will provide high-ROI supply. Overall, management believes AI hardware demand, SCA lock-ins and supply constraints will support strong long-term financial performance, although startup costs and slower price increases may partially offset margin expansion.
The following is the transcript of Micron Technology's fourth-quarter fiscal 2026 earnings call:
Satya Kumar, Vice President of Investor Relations and Treasurer, welcomed participants to Micron Technology's post-earnings analyst call for the fourth quarter of fiscal 2026. Joining him were President and Chief Operating Officer Manish Bhatia, President and Chief Technology and Product Officer Dr. Scott DeBoer, and Chief Financial Officer Mark Murphy. He reminded everyone that today's discussion contains forward-looking statements regarding market supply and demand, market trends and drivers, expected performance and guidance, and other matters. These forward-looking statements are subject to risks and uncertainties that could cause actual results to differ materially from today's statements. Please refer to the company's SEC filings, including the latest Form 10-K and the upcoming Form 10-Q, for risks that could affect results. Although the company believes the expectations reflected in the forward-looking statements are reasonable, it cannot guarantee future results, activity levels, performance or achievements. The company undertakes no obligation to update any forward-looking statements to conform to actual results. The Q&A session then began.
Benjamin Reitzes of Melius Research LLC congratulated Scott on his new role and said he wanted to talk about 2028. He noted the company added commentary on 2028 and said supply and demand would be tighter than this year and 2027. He wanted to understand what had changed and what it meant for margins. He said the company had given more margin commentary around 2027 than before, which was good, but he did not think 2028 margins would deviate from the level implied for 2027. He asked for further explanation on 2028.
Manish Bhatia, President and Chief Operating Officer, said he would answer first and Mark would add if needed. He also thanked Benjamin for the congratulations. On 2027 and 2028, he said the company is seeing stronger demand drivers than before. Server units will continue to grow into 2027. At the same time, agentic AI is growing rapidly and creating a CPU-driven demand stream. As a result, when planning fiscal 2027, the company has locked in more than 75% of full-year shipments, reflecting stronger demand and enabling it to extend customer allocation negotiations into 2028. That is one source of confidence: the demand picture for 2027 is stronger than previously expected. At the same time, as the company conducts more long-term agreement (SCA) negotiations with customers, including completed extensions, it is also more confident in the long-term outlook. Together, these give the company a very strong demand outlook through 2028. On the supply side, the structural constraints previously discussed remain: diminishing returns from technology transitions; HBM growing faster than conventional DRAM through 2028, meaning HBM's share of industry output capacity rises; and not only the trade ratio for current HBM, but also the higher trade ratio of more complex future HBM, which will limit supply. In addition, new cleanrooms across the industry take a long time to build, certify and equip, and even once in production, it takes several quarters to generate meaningful shipments. Therefore, combining demand and supply prospects, the company currently does not see a point at which supply and demand will balance.
Mark Murphy, Executive Vice President and Chief Financial Officer, added that, as Benjamin noted, for 2027 the company has said full-year margins will expand versus the first quarter because prices continue to rise, but at a more moderate pace. The company has also said many times that price increases will eventually slow. At that point, beyond slower price increases, mix will improve, and the company will optimize mix based on technology and product leadership. As stated, market conditions are expected to remain tight and provide support through 2028. Part of what offsets price and favorable mix is startup cost. But these are matters the company is managing and it expects to sustain strong financial performance.
Benjamin Reitzes of Melius Research LLC followed up. He noted Sanjay had a prominent seat at Trump's lunch and asked whether he was optimistic about industry growth and self-regulation, whether that supported the optimistic guidance, whether there was more information from the meeting, and whether memory was frequently mentioned.
Manish Bhatia, President and Chief Operating Officer, said Sanjay was not on the line and he would only speak to what he knew about the conversation. The company was of course pleased to take part in the forum. Micron was invited alongside model companies and accelerator companies, showing the importance of memory. The white paper framework released at the meeting has been signed by several model companies, and the framework is constructive for continuing to advance AI infrastructure, especially AI hardware infrastructure. One concept discussed frequently, as he understood it, is that one way to manage some security issues is to adopt security solutions, which will require more advanced hardware, including higher-performance, lower-latency memory and higher-bandwidth memory. This is because the responsiveness of gateways that may be set up in the future to manage security issues will depend heavily on the availability of high-performance and low-latency memory.
Melissa Weathers of Deutsche Bank Research also congratulated the two newly promoted executives. She noted that over the past few quarters the company had given views on total HBM TAM for 2028 and 2030. Clearly the pricing environment has changed. She asked whether the company was willing to give a new official HBM TAM number, whether it could help estimate directionally how much comes from bits and how much from pricing, and for the latest view on market size.
Manish Bhatia, President and Chief Operating Officer, thanked Melissa. He said the company is not updating that TAM outlook at this time. But it has said HBM shipments are expected to grow faster than conventional DRAM, meaning HBM's share of industry capacity will continue to rise through 2028. On pricing, at least for the company, it has significantly raised calendar 2027 HBM pricing and will reset it early in the calendar year to narrow the profitability gap with conventional DRAM. Beyond that, the company has not commented on specific HBM TAM prospects. The market continues to grow and remains an important driver, and HBM deployment continues to be key to unlocking the broader potential of AI. So it is an important part of the market.
Melissa Weathers of Deutsche Bank Research said she understood. On market share, she asked whether there was an update to the HBM share target. In the past the company said it wanted to raise it to around the company average, meaning a market share in the low 20s. She asked whether that was still the case. She also noted that in prepared remarks the company mentioned HBM4E and cooperation with Nvidia. She asked for more information on 2027 and HBM4E customer progress.
Manish Bhatia, President and Chief Operating Officer, said he would answer the first question and Scott would answer on HBM4E since he leads that product development. On HBM market share, about a year ago the company reached a milestone when HBM market share matched overall DRAM market share. At that time the company said the target would move based on various factors. Apart from saying HBM share is expected to be roughly comparable to overall DRAM share, the company has not updated a specific share target and does not treat a particular number as an absolute goal. Share will fluctuate for various reasons. But as noted in the previous question, HBM is a very important part of the market, allowing the company to stay at the forefront of accelerator platforms designed and deployed by customers, and it is also a key enabler for other AI promises and potential.
Scott DeBoer, President and Chief Technology and Product Officer, added regarding the Nvidia cooperation. This will be the first important custom HBM product on the market. The company has worked with Nvidia on HBM4E, the so-called NV HBM, for more than a year. The company sees a major opportunity because co-designing the product with a key customer makes its value significantly exceed standard HBM4E. He believes this will have a real impact on the industry, showing how future systems can optimize products.
Atif Malik of Citigroup Research asked the first question about the 26 SCAs covering 35% of sales through 2030. He asked whether these cover both DRAM and NAND and whether the company could break it down.
Manish Bhatia, President and Chief Operating Officer, said yes. The company does not break it down, but the SCAs do cover DRAM and NAND and run through 2030. Specifically, DRAM volume is slightly below 35%, while NAND bit volume is slightly above. Looking ahead, the company has more available volume. As stated, as SCA negotiations continue, that number could be higher in the future.
Atif Malik of Citigroup Research said he understood. On the impact of Chinese competition, he first asked whether the company could confirm that China sales exposure is small, and whether Scott could comment on Chinese competitors narrowing the technology gap.
Manish Bhatia, President and Chief Operating Officer, said on the first question, yes. Over the past several years and past several quarters, the company's China exposure has continued to decline and is expected to fall to the single-digit range in fiscal 2027.
Scott DeBoer, President and Chief Technology and Product Officer, said on technology, the company currently leads Chinese competitors by at least two nodes. He emphasized that the company is focused on maintaining technology leadership and truly differentiating its products and competitive approach. As Manish previously said and as noted on the call, the company's 1-gamma DRAM technology already accounts for most of its bits and is expected to become the largest node in company history. That technology relies on EUV. The next generation, 1-delta, is progressing well, and the company is focused on mass production in the second half of next year. EUV technology will be critical for all future advanced DRAM nodes. The company's expertise in this area, including working with technology suppliers and mask technology, continues to be a key differentiator for Micron.
Karl Ackerman of BNP Paribas Research noted that the company is seeing strong demand across most of its portfolio, but mobile and client bit shipments appeared to decline for a second consecutive quarter. He asked whether higher memory prices are weakening demand in that area. He also noted that SCA adoption is slower in that area and asked whether the company's SCA growth is coming from that customer group.
Manish Bhatia, President and Chief Operating Officer, thanked Karl. The company did see mobile and client bits decline sequentially, but revenue grew on higher pricing and improved mix. He noted that the high-end segment of client and flagship smartphones has strong demand for higher-content, higher-performance solutions, which is a key focus area for the company. Driven by this, the company expects PC and mobile industry revenue to grow even as unit sales decline. On SCAs, the company has SCAs in all business units, including mobile and client. It does not break them out specifically, but maintaining SCAs helps sustain diversified supply across all end markets.
Karl Ackerman of BNP Paribas Research said he understood. He asked Scott a question. He asked how Scott views the competitiveness of the company's internally optimized base die on HBM4E relative to peers. Some customers are adopting custom solutions. He asked whether the complexity and economic value mainly flow to compute customers or HBM suppliers.
Scott DeBoer, President and Chief Technology and Product Officer, first clarified one point. On HBM4E, the company co-designed with Nvidia, but it does not use an internal base die as with HBM4. The HBM4E co-design is based on a foundry process, used for both custom products and JEDEC standard products. On differentiation, as the company has shown in past HBM products, it ultimately comes down to power consumption, speed performance and margin in product and customer cooperation. In previous generations, different suppliers performed differently, and the company believes this will continue to be a Micron advantage in product quality and capability.
Manish Bhatia, President and Chief Operating Officer, added on economic value that HBM is a high-end product. As Scott said, NV HBM with customization is also expected to be a high-value product. The company believes HBM will continue to contribute and be a high-ROI product.
James Schneider of Goldman Sachs Research congratulated Manish and Scott. He wanted to understand the 10 new customer SCAs signed this quarter and what customers are asking for. Clearly they want supply and longer terms. But he asked whether the pricing structure has changed. Given expected tightness in calendar 2027 and 2028, if more upside can be captured in coming years, he asked whether the company is less inclined to require fixed floors and ceilings.
Manish Bhatia, President and Chief Operating Officer, thanked Jim. The SCA framework is similar, but the difference is that negotiations reflect current market conditions and future pricing prospects. As a result, the pricing direction has been upward. These factors have been reflected in customer discussions, whereas previous agreements were set under CQ2 market conditions. The company has said that about three-quarters of SCA revenue has a clear pricing framework, while about one-quarter is open to periodic negotiation or market-based pricing. The overall framework is similar, with most pricing frameworks having floor and ceiling ranges, but new agreements already considered current market conditions and the future tightness the company sees at the time of negotiation.
James Schneider of Goldman Sachs Research said he understood. He asked whether that means all price floors and ceilings have been reset to current higher market conditions, or whether the nature of the pricing terms is also different. He also asked whether the company discloses signed initial hyperscaler customers and whether these are included in TAM.
Manish Bhatia, President and Chief Operating Officer, said there are multiple frameworks. What he meant is that most frameworks with pricing have floors and ceilings, but the company continues to use multiple frameworks. Since the last earnings call, the 10 new SCAs signed cover customers from small to large. The company does not break them out specifically, only saying that SCAs now exist in all business units and customer sizes range from small to large, including the 10 most recently signed. Total is 26.
Christopher Caso of Wolfe Research, LLC asked the first question about capital expenditure and the comparison between fab construction capex and tool purchases. He said from the commentary he heard that construction capex is growing faster and was not sure whether he understood correctly. Cleanroom space constraints are limiting the ability to bring in tools this year. But the increase in construction capex is noteworthy because it will not generate bit production until at least 2029, possibly longer.
Manish Bhatia, President and Chief Operating Officer, said he would answer first and Mark would add. That is indeed the case. The main industry constraint is cleanroom space, because the strong growth in AI demand appeared relatively recently, and building cleanrooms takes a long time. So the company is focusing on that area. Christopher correctly noted that fiscal 2027 construction capex is higher than previously planned, most of it for cleanrooms coming online at the end of 2028 and beyond. This shows both how long the construction cycle is and why construction investment must start now, while also showing confidence in long-term demand, both from recent market trends and from the SCA structure and discussions with customers extending commitments beyond 2030. These SCAs are transformative for the company, matching supply with future demand and supporting confident investment.
Mark Murphy, Executive Vice President and Chief Financial Officer, added one point for Christopher. Most of the increase is indeed construction capex, mainly to accelerate cleanroom space availability in 2028 and beyond. The company expects this trend to continue beyond 2027. An important point Christopher raised is that this spending will not immediately translate into bits. These fabs will be equipped and produce wafers when needed, based on the company's view of the market and SCAs. These SCAs help the company keep a pulse on the market and ensure returns on capital expenditure.
Manish Bhatia, President and Chief Operating Officer, added another point for Christopher. The company will equip cleanrooms and build capacity based on demand trends. It has been executing long-term supply agreements with equipment suppliers to ensure equipment is available when needed. But of course, it will still equip cleanrooms and build capacity based on demand trends at the time.
Christopher Caso of Wolfe Research, LLC said he understood. His follow-up question was about the impact of CPU intensity on overall bit demand and supply-demand balance. This may be the biggest incremental surprise since the start of the year. CPUs do not have the same trade ratio effect as HBM. He asked how much this contributes to the supply-demand imbalance.
Manish Bhatia, President and Chief Operating Officer, said the recognition that agentic workloads run on CPUs is an important driver. LP, DDR memory and SSDs all have very high attach rates to support these agentic workflows. These workflows have begun to be implemented and are driving real value in enterprise and consumer settings. This is also one reason server units are growing strongly, as the company said, at a high-teens percentage rate. It also shows that logic chip growth is another vector to leverage AI trends. Therefore, this pushes up logic chip demand and enters total AI compute demand, while also placing more constraints on DRAM, showing that DRAM is the main constraint, not logic or data center power. He noted that there are multiple software implementations of enterprise agentic workloads. Progress after Meta's Muse release has been rapid within just a few weeks, showing how quickly agentic workloads are delivering real value to consumers.
Joseph Moore of Morgan Stanley Research said the slowdown in supply growth next year is somewhat surprising given the capex backdrop. The company mentioned the HBM suppression effect, but the gap should not change that much. He asked for an explanation of why supply is slowing despite high capex and what the factors are.
Manish Bhatia, President and Chief Operating Officer, asked Joe whether he was asking about both DRAM and NAND.
Joseph Moore of Morgan Stanley Research said yes, but mainly DRAM.
Manish Bhatia, President and Chief Operating Officer, said the company has stated that HBM is growing faster than conventional DRAM. As the industry moves from HBM3 to HBM4 and later to HBM4E in late 2027, the trade ratio keeps rising. These two things happening at the same time suppress the ability to grow bits. At the same time, bit growth for the company and other industry participants comes from new technology nodes, and transition timing affects bit growth, with diminishing returns from new nodes. These are all factors limiting supply. But the main factor remains cleanroom space across all manufacturers. Although the company's Idaho fab will produce its first wafers in the middle of calendar 2027, and other industry players may also open cleanrooms, meaningful supply growth after that will still take several quarters. These are the reasons the company believes DRAM industry supply shipments will decline next year.
Joseph Moore of Morgan Stanley Research said that makes sense. If the company is conservative on industry supply and there is more supply next year, there appears to be a lot of pent-up demand. Some AI racks have been downgraded as necessary, and if supply increases, would specifications be raised again? He asked whether he was being too optimistic.
Manish Bhatia, President and Chief Operating Officer, said that is exactly the point. Sanjay mentioned on the main call that the company does believe customers are choosing to maximize compute chip shipments given available memory supply. This creates potential demand for more memory to attach to these compute chips, improving system performance and end-customer experience. Therefore, if more memory is available, it can easily be used for higher content growth in AI workloads, whether attached to accelerators or CPUs.
Mehdi Hosseini of Susquehanna Financial Group, LLLP, Research, asked several follow-up questions. He noted the company emphasized that 2026 NAND bit shipments were below the industry average. He asked how 2027 to 2028 should be viewed and whether NAND bit demand or shipments could grow at the industry average of 25%.
Manish Bhatia, President and Chief Operating Officer, said frankly that the company has not commented on DRAM or NAND that far out. The company does expect overall NAND conditions to be tight, even if the industry is expected to grow at a mid-20% rate in calendar 2027 and 2028. The company's supply growth is affected by several factors: part of Singapore cleanroom space is currently used for an advanced R&D line for future NAND growth; at the same time, the company is preparing for the Singapore HBM fab ramp next year, and some existing cleanroom space is being used for pilot HBM operations. This causes supply growth this year to be below the industry. But the company is confident in its technology and expects continued ramp to G9 to provide high-ROI, cost-effective supply. In addition, the new cleanroom that broke ground earlier this year will come online in the second half of calendar 2028.
Mark Murphy, Executive Vice President and Chief Financial Officer, interjected for Mehdi due to time constraints. He asked him to ask his second question first, and then he would comment.
Mehdi Hosseini of Susquehanna Financial Group, LLLP, Research apologized and said he would ask quickly. He wanted to understand the team's view. Looking at DRAM, especially at the wafer level, the equipment is the same. He believes wafer-level DRAM is fungible, with differentiation at the back end. This wafer-level fungibility has not been seen before because past cycles were driven by a single product with highly concentrated customers. He asked whether his assumption is correct that wafer-level DRAM fungibility allows the company to better manage DRAM costs.
Scott DeBoer, President and Chief Technology and Product Officer, said first that he thinks it is partly correct, but he may take the opposite view. If you look at front-end variability, or the different things the company does in DRAM processes to optimize HBM, high-performance SOCAMM, LPDRAM and DDR6, the differences today may be greater than at any time in DRAM history. Different products must be built on the same node, with more differentiation built in. He then invited Manish to add.
Manish Bhatia, President and Chief Operating Officer, said that in the short term, running different products on the same manufacturing line does help adjust mix without running them in different fabs, since they can run on the same line. But as Scott said, each product has its own optimization vector. HBM optimizes bandwidth and TSV and requires unique process steps; DDR and LP also have their own unique process steps. The most important part of fungibility is that the company can flexibly adjust to where demand is, and even adjust mix to meet customer requirements. This is all within the same fab and is the most important part. But he does not think it necessarily helps cost.
Mark Murphy, Executive Vice President and Chief Financial Officer, said for Mehdi that he thought the call was nearing its end. He was pleased that questions focused on the long-term solid fundamentals and performance of the technology and business. He had expected some routine questions earlier. He added that the first-quarter guidance includes sequential single-digit unit growth in DRAM and NAND bits and double-digit cost growth, so please note that when modeling. In addition, excluding the impact of incentive compensation, R&D growth year over year will exceed the $1 billion mentioned last quarter, and will exceed $1 billion in 2027, as the company has increased R&D activity. He again asked everyone to ensure this information is used for modeling.