How Micron’s Building Biggest U.S. Chip Fab, Despite China Ban

In reality, these have billions and many billions of bits of memory. Bits that store information in ones and zeros in them. Data is at the center of all our devices and generative AI, and accessing and storing that data requires it. Memory

chip

s. And there is only one major memory manufacturer based in the United States: Micron. Micron. Micron. The maker of basic DRAM

chip

s, the

building

blocks of all memory. Micron stock has risen more than 30% since the beginning of the year, largely thanks to the generative AI hype. And while memory is a simpler and cheaper type of semiconductor than the high-power central processing units and graphics processing units that make other chips skyrocket, multiple memory chips are needed to support each GPU or CPU. , and that means making the memory take up more space. than other types of chips.

The new factory we are

building

in Boise, Idaho, right now and just poured concrete in October, will be over 600,000 square feet. More than 12 times the size you see behind me. On top of that, in New York we're going to build four of those 600,000-square-foot factories for a total of 2.4 million square feet. Micron's investment in four chip manufacturing plants in upstate New York, valued at $100 billion over 20 years, is expected to be the largest chip project in U.S. history. It's a commitment Micron made with a big helping hand from New York and a bid for National Chip Act funding.

We have to make it worth it. It's still a competitive environment, we're competing with the rest of the world. But being the only memory maker in the United States also comes with risks. Micron is the latest target of China's bans on American chips as the two countries compete for technological dominance. Micron has absolutely become a pawn in what I consider this 15-year war between China and the US. CNBC went to Micron's giant factory in Boise, Idaho, to see the start of its big expansion into memory manufacturing advanced on American soil and ask how to achieve massive growth amid a glut in the market and a geopolitical turmoil that is calling into question the quarter of its revenue that comes from China.

Micron was founded in 1978 by three chip engineers and one of their twin brothers in the basement of a dental office in Boise, Idaho. In 1980, it was building its first factory and in 1981 it was launching a revolutionary small 64k DRAM chip. These chips, used to store bits of data that can be quickly accessed by a CPU, ended up in many early personal computers. So there are two main types of memory DRAM and NAND. DRAM stands for dynamic random access memory and is volatile memory, meaning that when the power goes off, it loses all of its information. It is very fast, sits close to the CPU, and is used for real-time processing.

NAND flash memory is what's on your SSDs or storage cards, and NAND flash memory is non-volatile, meaning it will continue to store your memory even when the power goes out. Micron went public in 1984. Remember our name as you choose your career. Scott Gatzemeier joined as an intern in 1997. And at that time, DRAM was worth its weight in gold. This was like when the dotcom era was starting to take off, and it was incredibly exciting and something I wanted to be a part of. He now leads Micron's expansion projects in both New York and Boise, where he gave us a tour.

The characteristics that we incorporate into our NAND and DRAM devices are about 15 nanometers. Well, a human hair is 3000 times the size of those devices. So any hair follicles or dead skin could contaminate our equipment, our wafers, which would reduce our performance. That's why we wear bunny suits and dresses inside the factory. CEO Sanjay Mehrotra says the $15 billion Boise expansion and four factories in Syracuse, New York, are necessary because of how quickly the entire memory industry has grown along with computing and artificial intelligence. At the beginning of the century, memory used to make up about 10% of semiconductors.

Today, as you know, it ranges between 25% and 30% of the total revenue of the global semiconductor industry. Anything with computing requires a certain amount of memory. This trend of AI, memory, allows for deeper insights and that makes AI smarter. So as AI grows, memory gets bigger. Memory used to be a crowded field, but over the years it has narrowed down to just three of the best players. But the name of the game is high performance and low cost at the same time. Otherwise, you will be eliminated from the market. And there used to be 20 memory makers that were relevant and there's really only three right now.

When it comes to the largest type of DRAM memory, South Korean giant Samsung is by far the revenue leader, followed by SK Hynix, also outside Korea, and then Micron in the US. Micron has made 11 acquisitions since 1998. Numonyx, Elpida, Inotera, TI's memory business. For a very long period, they had not invested in a new factory, but they were still able to maintain their market share by acquiring other smaller memory companies that were closing or going bankrupt. Unlike most types of chips, memory was not in short supply during the chip shortage, but Micron and its competitors saw a big rebound during the pandemic-fueled boom in consumer electronics sales.

Micron's profits then fell significantly in 2022 due to weakening demand for PCs and smartphones, a slowdown that affected much of the chip industry. Micron reduced its production to limit oversupply and laid off 10% of its workforce in early 2023. It now employs about 43,000 people worldwide. When I look at this market over the last 30 years, it's always feast or famine. And now we have an oversupply, but guess what? Wait a couple of months and we will have a shortage of supply. Micron supplies memory to phones from Apple, Motorola, Asus and more and is optimistic about growth. The smartphone mix is ​​increasingly leaning towards high-end smartphones, towards flagship smartphones, which also require more memory.

So as we look ahead to 2024, we actually expect that year-over-year, total smartphone unit sales worldwide will increase. But Micron is also focused on fast-growing markets like automotive and generative AI. Its most advanced product, high-bandwidth memory, will be produced in volume next year. HBM helps AI models like ChatGPT remember past conversations and user preferences to generate more human-like responses. It is capable of including 50% more memory capacity in a memory cube. It is capable of giving you 50% faster performance and can give you about two and a half times better power and performance efficiency. And these are all elements that are critically important in AI applications.

Unlike market headwinds like oversupply and slow device sales, Micron faces a major challenge that Korean memory giants do not. China banned some of Micron's sales in May, citing cybersecurity risks. About 25% of Micron's commercial revenue comes from China markets, and about half of that revenue is at risk given the CAC's decision. Last year, the United States banned chip companies from supplying China with certain key technologies. Micron is absolutely just a pawn in this game right now. They were not the first nor the last. Have you become a pawn in this geopolitical chip war between the two countries?

What I can tell you is that it is very important that the United States and China provide an environment for companies so that they can invest in a predictable way. Micron, of course, is fully committed to our customers in various end markets in China. Meanwhile, Micron has begun construction of a $2.75 billion test and assembly facility in India. So obviously Micron is trying to diversify its base. They have testing and packaging facilities in China and obviously they are trying to move and diversify out of China. India has been trying to attract and attract multinational chip makers to set up projects in the country through multiple schemes and projects, and this is the first major one.

For now, Micron is still in China, but China is turning to memory from Samsung, SK Hynix, and smaller Chinese memory makers. This is possible because memory is considered a commodity, meaning it is relatively easy to switch between products from different companies, although it is not guaranteed to last. What I'm going to find really interesting is that when we get back to the boom days and Hynix and Samsung can't meet all the volumes, we may see China dive back into Micron and suddenly lift any restrictions. As to whether Micron's technology really poses a risk to China's national security.

I think it's a front compared to a CPU, a GPU or a system. It's quite difficult to insert something nefarious into something like storage or memory. That would be a technology I had never heard of. We think China was being very nasty to Micron about it. In October, Schumer led a delegation of senators to visit China for a rare meeting with President Xi, in part to discuss the Micron ban. China is upset with the Biden administration's very smart ban on selling certain types of chipmaking equipment to China, but we're going to defend Micron. This is also not the first time Micron has been at the center of tensions between the United States and China.

In 2018, the United States cut ties with Chinese chip company Fujian Jinhua after accusing it of stealing intellectual property from Micron, a claim the Chinese company denied. Micron told CNBC that its intellectual property is critical to revenue in the memory business. Behind me you can see Micron's patent wall and we have more than 54,000 patents. Our main source of income is that we want to use our intellectual property to make our chips better than others. The week before Schumer visited China, Micron celebrated its 45th anniversary by pouring the first cement for its new Boise factory. Both Boise and New York will produce these advanced 300-millimeter wafers with memory technology.

The most advanced size wafer in the industry, right? In volume production, yes. Micron's current U.S. presence includes its headquarters and research and development facility in Boise, a state-of-the-art memory factory in Manassas, Virginia, and offices in five other states. Its main nodes are manufactured in factories in Japan and Taiwan. By 2026, Micron plans to begin cutting-edge production at the new Boise Fab. And then at our state-of-the-art facility in New York, when fully built, we will produce over 100,000 wafers per month. Each of these silicon wafers goes through more than 1,000 steps over 90 days, traveling through automated aerial robotics. The vehicles that deliver these wafers travel more than 23,000 miles a year in our factory.

Memory isn't a high-cost commodity like logic chips or computing powerhouses like CPUs and GPUs, but that's actually one of the reasons Micron's factories are even larger than those needed to create logic. . Memory is therefore very cost sensitive and we have to achieve economies of scale to mass produce our chips at a level that meets market demands. Another reason the factory is so big is to house huge, expensive machinery like this $250 million extreme ultraviolet photolithography machine to print the smallest features on chips. So the size of this tool is absolutely huge. There are more than 100,000 pieces. The tool weighs 20 tons.

It must be delivered on three 747s and then shipped to us in 40 shipping containers on the back of 20 semi-trailers. EUV machines are only made by one company, ASML, in the Netherlands, and are reputed to consume incredible amounts of energy. Micron says each of its new factories will use the energy equivalent of 25,000 homes. Renewable, reliable energy is absolutely critical for our factories. In the United States we have access to that reliable, renewable energy, and the cost of that energy is about 25% of the cost of equivalent energy in Asia. Many of those 1,000 steps also require a lot of water.

So in New York we are building 600,000 square foot factories. Each of those factories will use 25 Olympic-size swimming pools of water a day. 75% of that water will be reused and recycled directly on site. HeWater and energy were big reasons Micron opted for American expansion. We have an area of ​​abundant fresh water, not only the Finger Lakes, but also two Great Lakes, Lake Erie and Lake Ontario, and low-cost energy generated primarily by hydroelectric, wind and solar power. So we are ready for it. We know it's going to be a transition. Interestingly, energy costs are lower in the United States than in most of the world.

People are more expensive in the United States, and so are the materials and the cost of building that factory. But that gap is narrowing over time. Still, in Arizona, the world leader in advanced chips, Taiwan Semiconductor Manufacturing Company, recently blamed a shortage of skilled labor for delays at its massive $40 billion factory being built there. That won't happen in New York because we already have a legacy, we have Wolfspeed, we have GlobalFoundries, so this is not a new industry for us. We're attacking this from kindergarten through 12th grade and into graduate school. We are big donors to the STEAM School in Syracuse, New York.

We also organized a chip camp. In fact, my daughter is 13 and went to chip camp in Boise, Idaho this summer. Still, American participation in chip manufacturing has plummeted in recent decades. It costs at least 20% more to build and operate a new factory in the United States than in Asia. There, labor is cheaper, the supply chain is more accessible, and government incentives are much greater. That's why the Chips and Science Act sets aside $52.7 billion for companies to manufacture in the U.S., and it's no surprise that Micron and more than 460 other companies have applied for those funds.

Senate Majority Leader Chuck Schumer came up with the idea for the Chips and Science Act in the gym in 2019 while speaking with Republican Sen. Todd Young. And suddenly, when it came to chips so essential to everything we do, we had lost that advantage. And if we do not regain that advantage, and not only in chips, but in science in general, we would no longer be the number one economic power in the world. Prosperity in the United States would decline. The number of jobs would decrease. So I felt a passion for this. The United States has especially lost that advantage on the memory side of the chips.

Now, Micron is making big promises about how to change that. Today, only 2% of the world's total DRAM memory production comes from the US, and all of that comes from Micron's factory in Manassas, Virginia. With Micron's investments through chip support in Boise, Idaho, as well as Syracuse, New York, that 2%, over the course of almost 20 years, will become approximately 15% of global production coming from the In the US, Governor Kathy Hochul signed the $10 billion Green Chips Act last summer to attract companies like Micron, which says it is eligible to receive up to $5.5 billion from the bill. If they hadn't passed the Chips and Science Act first, I don't think there would have been as many incentives as needed to bring them home or encourage existing American manufacturers to continue expanding, so I knew I had to court them, we talked about our incentives , but we also get 50,000 jobs.

To deliver on its promises in New York and Idaho, Micron is betting big that as the world increasingly relies on data-intensive technologies, they will also need more and more memory. Big language learning models and other things like that continue to increase demand. Now we're moving toward things like FaceTime, higher resolution images, and movies on demand. All of this requires more and more memory to be available. Micron says construction in New York will begin in late 2024 and chip production will begin in 2027, with some major promises about what that means for the country's share of memory manufacturing. 10% of Micron's current production is produced here in the US, as these projects move forward and are completed, nearly 60% of Micron's production will come over the course of the next two decades from here in The USA.