// php echo do_shortcode (‘[responsivevoice_button voice=”US English Male” buttontext=”Listen to Post”]’)?>
Micron Technology executives discussed the company’s momentum on Investors Update Day last week, revealing plans to increase production of the sixth-generation 232-layer flash by the end of 2022, among other updates.
Micron has spent the last few years trying to lay claim to being a leader in DRAM and NAND flash memory development. The company announced in late 2020 that it was delivering its 176-layer NAND-based Replacement Technology (RG) with its CMOS-sub-array architecture, and introduced its 1-alpha DRAM node in early 2021. improving memory density by 40% compared to its 1z DRAM node. However, Micron chose not to pursue the 3D Xpoint market and instead focused on developing Express Computing (CXL).
Roadmap of NAND technology
The most significant technology news from the session was that Micron expects to increase production of the sixth-generation 232-layer flash by the end of the 2022 calendar year, according to Scott De Boer, executive vice president of technology and products.
“This node will provide increased density, power and bandwidth compared to our 176-layer node.” He said that external and internal controllers continue to be a strong element in the focus of Micron’s vertical integration.
Micron also continues to focus on QLC (4 bits / cell) NAND, while expanding its CMOS process architecture – sub-array and two stacks, which stacks two 3D NAND arrays on top of each other. Also known as ‘string stacking’, the process can address difficulties in semiconductor manufacturing, such as etching multi-layer bonding holes. This is because when the depth of the hole deepens, the sides of the holes can be distorted and prevent NAND cells from working properly.
Progress of DRAM
In terms of DRAM, DeBoer said Micron will increase its 1-beta node in late 2022 using its advanced CMOS technology, with 1-gamma expected to follow in 2024 with the introduction of extreme ultraviolet (EUV) lithography. The company is expanding its planned DRAM roadmap while making continuous investments in 3D DRAM research and development.
Micron CEO Sanjay Mehrotra explained that over the past five years, the company has made tremendous progress in establishing a competitive advantage in technology, product and manufacturing. This makes the company well-positioned for what he called the “next era of leadership”, as the digital transformation creates more opportunities for data creation, which was accelerated by the pandemic, as well as hyper-connectivity activated by 5G in data centers. smart advantage devices and smart consumer devices.
Add to this the amount of data generated by artificial intelligence (AI) and machine learning (ML), and the amount of data generated is expected to double by 2025. “As AI algorithms become more complex, they work on more data and then they need more memory to process and extract insights, ”Mehrotra said.
He, meanwhile, stressed that the “5G train” is leading to greater use of DRAM content in smartphones that do computational photography and advanced video editing thanks to AI algorithms. These AI algorithms also increase the demand for content in memory in autonomous vehicles as they become data centers on wheels.
Search for DRAM and NAND
Mehrotra explained that what is even more exciting is that the demand for memory and storage is in many large end markets. “This diversity of large end markets has a positive impact on the stability of the demand environment.”
Demand is no longer anchored by computer development; it is also driven by mobile centers and data centers, while the personal computer market is reviving with the sudden shift to remote work. He said the data center is growing faster than the industry average due to the cloud and the adoption of AI to support new business models. This will stimulate growth for DRAM as well as NAND, which will also have high demand for growth in industry and the automotive industry, but slightly below the average for the PC and mobile industry, with smartphone devices stabilizing on the basis of global sales.
All of this diverse memory and storage capability means they will grow faster than the wider semiconductor industry, Mehrotra explained, while the industry’s profitability continues to grow. Demand leads to a healthier foundation for the industry, while delays in Moore’s Law increase the discipline of supply in the industry. There is an increased capital intensity required to implement a new node, as there is more expensive equipment and sophisticated tools are needed to implement increasingly complex functions, as well as all aspects of materials and structure.
Said Mehrotra Micron’s strategy is to increase basic supply in line with demand and maintain the discipline it has demonstrated over the last few years, while making the necessary capital investments and given the growing complexity of the industry.
Sumit Sadana, Micron’s executive vice president and chief business officer, also highlighted the company’s focus on boosting overall supply growth in line with market demand. “Our overall share in the supply of bits will remain equal over time, but we are focused on stimulating profits on preferred stocks and important parts of the market.
Turning to the data center market means that understanding current architectures cannot meet the needs of tomorrow’s workloads. As a result, the market is becoming increasingly heterogeneous, with the bandwidth between the CPU and memory becoming even more critical, Sadana explained. By bringing memory closer to the CPU, more specialized processors and the CXL protocol can address bottlenecks in data traffic – the CXL will “usher in a whole new era of memory connectivity,” Sadana said.
Micron’s recent landmark move was to leave the 3D Xpoint market with a view to developing CXL solutions. No specific product details were provided, but Sadana said CXL 1.1 will begin implementation later this year as proof of concept.
“The real deployment of memory in CXL will begin in late 2024, when CXL 2.0 is introduced, and then accelerate even more when the CXL 3.0 standard allows memory to be consolidated in a very real and practical way,” he said. . “Combining memory and disaggregating memory are critical aspects that our customers are looking for.”
Even more important is the expansion of the amount of memory that can be connected to a processor, which ultimately allows CXL in a very material way. This will result in more average DRAM content on the server.
How much this DRAM will cost may depend on different prices for Micron customers, as the company has launched a new pricing model called Forward Pricing Agreements (FPAs). Instead of entering into long-term agreements based on volume rather than prices – as they are now – the FPA will see prices based on both volume and prices. The aim of the model is to reduce price fluctuations, although profits are traditionally made through price volatility.
The new pricing model comes at a time when Micron is facing some strong “inflationary cross-winds”, including higher energy prices, as well as the entire semiconductor industry, as it invests capital in instruments to meet demand and focuses on the current challenges of the supply chain. This includes AI and ML platforms that help predict operational challenges.
Manish Bhatia, Micron’s executive vice president of global operations, said the collaboration between its technology development and manufacturing organizations, as well as equipment suppliers, allows it to identify and address variations in processes and equipment at the start of production growth.
During a period of high capital investment, Bhatia explained that it is especially important to strive for higher productivity. “Micron’s commitment to intelligent manufacturing has really allowed us to achieve this higher productivity, whether it’s improving tool performance or increasing tool life by implementing predictable maintenance.”
As equipment delays or delivery times are longer and more common, Bhatia said it is even more important to boost equipment productivity so that Micron can maintain production ramps for leading technologies and provide possibility for fast deliveries for its customers.
– Gary Hilson is the editor-in-chief who focuses on memory and flash technology for the EE Times.
Micron pulls forward on DRAM