Investment Thesis
We already know memory chip manufacturer Micron Technology, Inc. (NASDAQ:MU) is benefiting from the AI enterprise wave. With more enterprises increasingly deploying AI in their operations and services, the company is taking advantage of the growing need for high-performance memory technology. Their latest financial forecasts reflect a significant uptick in demand, with revenue projections for the financial 3Q 2024 marked between $6.4 billion and $6.8 billion, higher than the anticipated $5.99 billion.
Powering this, Micron released their HBM3E products to respond to the high-bandwidth memory (HBM) market demand, which requires rapid data processing capabilities to handle complex algorithms and massive data sets. I believe that the less commoditized nature of HBM allows Micron to command higher prices to support them in realizing higher profit margins compared to standard memory chips.
Micron’s partnership with NVIDIA Corporation (NVDA), for example, makes them one of the most important players in the development of new processors for AI applications. The company has started the mass production of their HBM3E chips that are used in Nvidia’s H200 Tensor Core graphic processing units (GPUs). These chips can reduce power consumption by 30% compared to competitors to help address critical energy efficiency concerns in data center operations where these GPUs are primarily deployed, as claimed by the company. I believe energy optimization will be a big purchasing decision going forward for GPU customers. Micron is leveraging this.
With projections indicating several hundred million dollars in revenue from HBM products alone, I believe Micron’s financial outlook appears strongly tied to the success of their AI-focused products and partnerships. The AI-driven demand for these advanced memory solutions has also enabled Micron to differentiate themselves from other industry players like Samsung and SK Hynix, which also provide alternative short-term memory solutions.
However, as AI continues to evolve, so will the need for local AI compute in consumer PCs, smartphones, and even work laptops need to jump to handle more edge computing for AI. I wrote about this in December for Intel Corporation (INTC) on the CPU side. Memory units will be just as important. I believe Micron is going to benefit immensely from this as well.
Companies like Apple Inc. (AAPL), and Alphabet Inc. (GOOG), (GOOGL) are integrating advanced AI capabilities into their devices to push into smaller AI systems, even on less powerful devices, thereby expanding the AI ecosystem beyond high-end models. These enhancements will likely require consumers to upgrade their devices to fully benefit from new AI-driven features. These new devices will require new memory features. Collaborations between tech giants like Apple and OpenAI exemplify this trend towards embedding AI directly into hardware on local devices.
As AI functionalities become standard features in smartphones and PCs, the requirement for advanced memory that can quickly process and analyze large volumes of data in real-time grows. With this, I think Micron continues to be a “strong buy” going into earnings (expected during market hours on June 26th). They have performed well since the last quarter. I believe AI on smartphones will help drive them further.
Why I Am Doing Follow-Up Coverage
In my previous post, I highlighted Micron’s recent performance, partly buoyed by the U.S. government’s CHIPS Act grants aimed at revitalizing America’s semiconductor manufacturing capabilities to reduce dependence on overseas production and enhance national security. Micron was granted $6.1 billion to support the expansion of their production facilities in New York and Idaho. I refer to this government-backed financial support as “rocket fuel” that boosts Micron’s stock by lowering capital expenditure burdens and fostering an environment ripe for sustained growth and profitability.
Following the CHIPS Act funding, Micron’s stock has jumped, reflecting the impact of the government’s commitment to revitalizing the U.S. semiconductor industry. The company plans to invest up to $100 billion over the next two decades, which is expected to create thousands of jobs and mitigate previous supply chain vulnerabilities brought about by recent global events and supply chain disruptions.
However, I have observed how mobile AI is setting the stage for shifts in smartphone technology, particularly memory requirements for edge computing. The market is geared for expansion, which is projected to grow by $39.91 billion between 2023 and 2028 at a CAGR of 26.78%. This is driven by advancements in AI capabilities within smartphones, such as enhanced natural language processing and machine learning.
In 2024 alone, shipments of generative AI smartphones are anticipated to reach over 100 million units, and by 2027, this number is expected to soar to 522 million units, indicating a CAGR of 83%. These smartphones require substantial computational power and memory to handle sophisticated AI tasks directly on the device to reduce latency.
Micron’s role in supplying memory products, especially those suited for high-bandwidth and quick-access applications, will be more critical for supporting the intensive data processing and real-time analytics required by mobile AI technologies, in my view. For instance, MediaTek and QUALCOMM Incorporated (QCOM) have emphasized the importance of integrating AI capabilities in flagship smartphones that will boost demand for powerful and efficient memory chips.
Earnings Estimates and What I Am Looking For On The Call
As Micron heads into their fiscal 3Q 2024, market analysts have set expectations for earnings per share (EPS) at $0.52 and revenue forecasts at $6.67 billion. These figures mark significant year-over-year growth, with EPS turning positive from a negative $1.43/share this quarter last year.
I believe this projected increase in revenue highlights a robust recovery and growth trajectory for Micron due to the strong market demand for their memory and storage solutions. I will be looking
At the Goldman Sachs’ Global Semiconductor in May, Chief Business Officer Sumit Sadana remarked:
So, when we think about enabling AI PCs, AI smartphones, a lot is focused on how to increase the amount of DRAM that’s addressable in these products and how to lower the power consumption. And of course, lots of interesting things [are] happening on the power consumption side. And new architectures, like processing and memory type of approaches being contemplated in order to figure out what kind of approaches in like laptops and smartphones could ultimately bring value of higher bandwidth, more DRAM addressability, and enablement of AI workloads on these devices without creating a lot of cost and without creating a lot of power consumption that HBM would create. -Goldman Sachs Conference.
So those are the approaches that are being worked on. It will take many years for some of these to come to market. So, the easiest path is to simply expand the amount of DRAM. And that’s what I was talking about earlier about smartphones getting to 12 gigabyte is really one important place where you can start running, for example, 7 billion parameter models. Some companies are trying to run like 10 billion, 13 billion parameter models, but at least that 7 billion parameter model could be run using a 12-gigabyte smartphone on the DRAM side. And then on laptops, more like 16 gigabytes but preferably more 24 to 32 to do it well. -Goldman Sachs Conference.
Mobile AI Is Big
The integration of advanced AI into smartphones is creating a significant shift in the memory market that continues to drive a spike in demand for mobile DRAM. This trend is starting to permeate the broader market, that will continue to set new standards for what smartphones can achieve with AI. Current smartphones that support basic AI functionalities are typically equipped with up to 12GB of RAM. However, with the introduction of more complex AI features such as on-device generative AI, which includes text-to-image models and LLMs, the memory requirements are expected to surge. As detailed in recent analyses, upcoming Android smartphones will likely need to standardize at least 20GB of RAM to manage these AI operations that require computational and data storage demands of LLMs. This estimates an increase in DRAM requirements by as much as 7GB per device, depending on the complexity and the nature of the AI tasks being performed.
In addition, with the continual push for more advanced AI capabilities, smartphone manufacturers are expected to soon incorporate AI-driven chips capable of supporting upwards of 30 tera operations per second (TOPS) on neural processing units (NPUs), which are designed to improve the speed of processing AI tasks. This development will drive larger DRAM capacities to support the high-performance requirements of these NPUs.
I expect that as smartphones incorporate more DRAM to meet AI requirements, the cost per unit for manufacturers increases, which in turn could lead to higher retail prices for consumers. Yet, this also presents an opportunity for exponential growth in revenue per unit sold from the added value that advanced AI capabilities bring to smartphones.
Valuation
Micron’s forward price-to-cash flow (P/CF) ratio stands at 17.73, which is significantly below the sector median of 23.13. This discrepancy suggests a -23.32% undervaluation compared to their peers.
Considering Micron’s strong utilization rates for their facilities (offsetting the depreciation concerns I had in earlier research on them), I think that the free cash flow (FCF) provides a robust proxy for assessing the company’s operational efficiency and long-term sustainability. Their management of production capabilities and capital expenditures has historically led to higher FCF making it a critical metric for investors focused on the fundamental value generation of the company. Eventually, companies are most sustainable on Free Cash Flow, not profits.
I continue to believe the current valuation gap implies a potential upside in Micron’s stock price if it were to realign with the sector median Free Cash Flow. If Micron’s P/CF were to converge with the sector’s median from their current position, this would imply a 30.45% upside.
Where This Fits With My Previous Valuation Estimates
In my previous estimate, I argued that Micron had up 59.28% based on the upside based on Micron at the time converging on the sector median Price/Cash Flow ratio.
Since then, shares have increased by 24.75%. In essence, the remaining 30.45% upside recognizes the remaining upside potential of the stock, converging on the sector median Price/Cash Flow.
Risks
Recent developments have seen China impose bans on Micron’s products, citing national security risks, without specifying the exact concerns. China’s actions could significantly impact Micron’s revenue, as sales to Chinese firms contribute notably to their overall earnings. The company has already warned of a more substantial revenue hit than previously anticipated, expecting an effect on about half of their revenue from China-headquartered firms, or a low double-digit percentage of their total revenue. This situation stems from heightened tensions between the U.S. and China over technology and trade, with Micron being caught in the middle of an international dispute.
The implications of this ban are also critical because of the potential escalating restrictions. Should China extend these prohibitions, it could further alienate Micron from a significant portion of the global semiconductor market. There is an ongoing risk that Chinese customers might shift their business to competitors like Samsung and SK Hynix.
In response to these challenges, Micron has expressed a commitment to continuing their operations in China and is seeking to engage with Chinese authorities to surpass these regulatory hurdles.
Despite these, I believe that Micron is going to benefit a lot from the AI wave. Samsung, a competitor, last year announced a huge reduction in their memory chip output, which comes at a time when the industry grapples with one of their worst downturns in the non-AI memory division, driven by an oversupply that has severely depressed prices. Samsung’s operating profit dropped by more than 95% in that quarter. By scaling back production (with effects still being felt today), Samsung aimed to stabilize the market and improve their pricing, which, I think, could help all players in the industry recover from current lows. Micron has been able to take advantage of this over the last 15 months. AI DRAM will allow them to leverage this further, offsetting risks from a China ban.
Bottom Line
Heading into earnings, I believe that Micron is poised for continued robust growth as we approach their earnings announcement. The company’s strategy in the AI market through advancements in high-bandwidth memory contributes to their market leadership. With revenue projections for the 3Q 2024, ranging between $6.52 billion and $6.94 billion, investors are confident that the demand for Micron’s memory solutions will be sustained in the coming years. This growth is anchored in their recent rollout of HBM3E products and solidified by a lucrative partnership with Nvidia. I believe the AI opportunity for mobile will help the stock go further
As AI becomes more embedded in mobile devices, I expect that the demand for Micron’s sophisticated memory solutions is expected to grow further, propelled by the broader adoption of AI technologies in smartphones and PCs. These devices require advanced memory architectures to efficiently handle increased data processing and real-time analytics, areas where Micron’s products excel.
Given the upside potential, I think the stock continues to be a strong buy.
Read the full article here
