Transcription
These actually have billions and many billions of memory bits. 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 these. Memory chips.
And there's only one major memory maker based in the U.S. Micron. Micron. Micron. The maker of basic DRAM chips, the building blocks of all memory. Micron stock is up more than 30% since the start of the year, largely thanks to generative AI hype. And while memory is a simpler, cheaper type of semiconductor than the high powered central processing units and graphics processing units sending other chip stocks through the roof, multiple memory chips are needed to support each GPU or CPU, and that means making memory takes more space than other kinds of chips.
The new fab that we're building in Boise, Idaho, right now that we just poured concrete on in October will be over 600,000 sq ft. So over 12 times the size that you see behind me. In addition to that, in New York, we're going to build four of those 600,000 sq ft fabs for a total of 2.4 million sq ft. At $100 billion, over 20 years, Micron's investment in four chip fabrication plants, or fabs, in upstate New York, is slated to be the biggest U.S. chip project in history. It's a commitment Micron made with some big help from New York and a bid for National Chips Act funding. We have to make it worth their while. It's still a competitive environment; we're competing with the rest of the globe.
But being the only U.S. memory maker also comes with risk. Micron is the latest target of China's bans on U.S. chips as the two countries vie for technological dominance. Micron has absolutely become a pawn in what I consider this 15-year war between China and the U.S. CNBC went to Micron's giant fab in Boise, Idaho, to see the start of its major expansion into advanced memory manufacturing on U.S. soil and ask how it will pull off massive growth amid a market oversupply and geopolitical turmoil that's casting doubt on 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. By 1980, it was building its first fab and pumping out a revolutionary small 64k DRAM chip by 1981. These chips, used for storing bits of data that can be quickly accessed by a CPU, ended up in many of the early personal computers.
So there are two main types of memory: DRAM and NAND. DRAM stands for Dynamic Random Access Memory, and it is a volatile memory, which means that when the power is removed, it loses all of its information. It's very fast, and it sits near the CPU and it's used for real-time processing. NAND flash memory is what's in your SSDs or your storage cards, and NAND flash is nonvolatile, meaning it'll still store your memory even when power is removed.
Micron went public in 1984. Remember our name when choosing your career. Scott Gatzemeier joined as an intern in 1997. And at the time, DRAM was worth its weight in gold. This was kind of as the dot-com era was starting to take off, and it was incredibly exciting and something that I wanted to be a part of. Now he heads up Micron's expansion projects in both New York and Boise, where he gave us a tour.
The features that we build on our NAND and DRAM devices are some 15 nanometers. Well, a human hair is 3,000 times the size of those devices. So any, you know, hair follicle or dead skin could cause contaminations to our equipment, our wafers, causing our yield to lower. So this is why we wear the bunny suits and the gowns inside of the fab.
CEO Sanjay Mehrotra says the $15 billion Boise expansion and the four fabs in Syracuse, New York, are necessary because of how fast the entire memory industry has grown alongside compute and AI. At the turn of the century, memory used to be about 10% of semiconductors. Today, you know, it varies from 25% to 30% of the total global semiconductor industry revenue. Anything that has compute requires some amount of memory. This trend of AI, memory enables deeper insights, and that makes AI smarter. So as AI gets bigger, memory gets bigger.
Memory used to be a crowded field, but over the years it's whittled down to just three top players. But the name of the game is high performance and low cost at the same time. Otherwise, you're going to be blasted out of the market. And there used to be 20 memory makers that were out there that were relevant, and really, there's only three at this point. When it comes to the biggest type of memory, DRAM, South Korean giant Samsung is by far the leader in revenue, followed by SK Hynix, also out of Korea, and then Micron in the U.S.
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 fab, but they were still able to retain their market share by acquiring other smaller memory firms which were either going out of business, bankrupt.
Unlike most kinds of chips, memory wasn't in short supply during the chip shortage, but Micron and its competitors did see a major upswing during the pandemic-fueled boom in consumer electronics sales. Then, Micron's profits fell significantly in 2022 due to weakened demand for PCs and smartphones, a downturn that's affected much of the chip industry. Micron reduced its output to limit oversupply and laid off 10% of its workforce in the beginning of 2023. It now employs some 43,000 people globally.
"When I look at this market over the past 30 years, it's always feast or famine. And we have an oversupply now, but guess what? Give it a couple of months and we will be in an undersupply."
Micron supplies memory in phones from Apple, Motorola, Asus, and more, and it's optimistic about growth. The mix of smartphones is going more and more toward higher-end smartphones, toward the flagship smartphones, which require more memory as well. So when we look ahead at 2024, we actually expect that year over year, total worldwide smartphone unit sales will increase. But Micron is also focused on rapid growth markets like automotive and generative AI. Its most advanced product, High Bandwidth Memory, is set for volume production next year. HBM helps AI models like ChatGPT remember past conversations and user preferences to generate more human-like responses. It is able to pack 50% more memory capacity in a memory cube. It is able to give you 50% faster performance, and is able to give you about two and a half times better power and performance efficiency. And these are all the elements that are critically important in AI applications.
Unlike the market-wide headwinds of oversupply and slow device sales, Micron is facing one major challenge that the Korean memory giants have not. China banned some of Micron sales in May, citing cybersecurity risks. About 25% of Micron's business revenue comes from China markets, and about half of that revenue is at risk given the CAC decision. Last year, the U.S. barred chip companies from supplying China with certain key technologies.
"Micron is absolutely just a pawn in this game right now. They weren't the first and they were not the last."
"Have you become a pawn in this in this geopolitical chip war between the two countries?"
"What I can tell you is that it's very important for U.S. and China to provide an environment to the businesses so that they can invest in a predictable manner. Micron, of course, is totally committed to our customers across various end markets in China."
Meanwhile, Micron has started construction on a $2.75 billion assembly and test facility in India. So Micron is obviously trying to diversify its base. It has testing and packaging facilities in China, and obviously they are trying to move, diversify out of China. India has been trying to lure and attract multinational chipmakers 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 instead. That's possible because memory is considered a commodity, meaning it's relatively easy to switch between products from different companies, although that's not guaranteed to last.
"What I'm going to find really interesting is when we get back to the boom days, and Hynix and Samsung can't fulfill all the volumes, you might see China diving back into Micron and suddenly lifting any restrictions."
As for whether Micron's tech does indeed pose a national security risk for China. "I believe it's a front compared to a CPU or a GPU or a system. It's pretty hard to embed something nefarious into something like storage or memory. That would be technology that I have never heard of. We think China was being very nasty about this to Micron."
In October, Schumer led a delegation of senators to visit China for a rare meeting with President Xi, in part to discuss the ban on Micron. China's upset with the Biden administration's very smart prohibition of selling certain types of chip manufacturing equipment to China, but we're going to stick up for Micron.
This also isn't the first time Micron has been at the center of U.S.-China tensions. In 2018, the U.S. cut ties with Chinese chip company Fujian Jinhua after accusing it of stealing IP from Micron, a claim the Chinese company denied. Micron told CNBC its IP is critical for revenue in the commodity business of memory.
So behind me you see Micron's patent wall and we have over 54,000 patents. "Our primary revenue source is we want to use our IP to manufacture our chips better than others."
The week before Schumer's China visit, Micron marked its 45th anniversary by pouring the first cement for its new Boise fab. Both Boise and New York will be producing these advanced 300-millimeter wafers with memory technology on them. The most advanced size wafer in the industry, correct? In volume production, yes.
Micron's U.S. footprint today includes its headquarters and R&D facility in Boise, a lagging-edge memory fab in Manassas, Virginia, and offices in five other states. Its leading nodes are made at fabs in Japan and Taiwan. By 2026, Micron plans to start leading-edge production at the new Boise Fab. And then in our state-of-the-art facility in New York, when it's fully built out, we'll produce over 100,000 wafers per month. Each of these silicon wafers goes through more than 1,000 steps over 90 days, traveling via automated overhead robotics. The vehicles that deliver these wafers travel over 23,000 miles annually in our fab.
Memory isn't a high-cost product like logic chips, computing powerhouses like CPUs and GPUs, but that's actually one reason Micron's fabs are even bigger than those needed to make logic. So memory is very cost-sensitive, and we have to get economies of scale to mass produce our chips on a level that meets the market demands. Another reason the fab is so large is to accommodate huge, expensive machinery like this $250 million extreme ultraviolet photolithography machine for printing the chips' tiniest features. So the sheer size of this tool is absolutely massive. There's over 100,000 parts. The tool weighs 20 tons. It has to be delivered on three 747s, then shipped to us in 40 shipping containers on the back of 20 semis. EUV machines are only made by one company, ASML, in the Netherlands, and they're infamous for taking incredible amounts of power. Micron says each of its new fabs will use the equivalent of 25,000 homes' worth of energy.
"Renewable, reliable energy is absolutely critical for our fabs. In the U.S., we have access to that renewable, reliable 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're building for 600,000 sq ft fabs. Each one of those fabs will use 25 Olympic-sized swimming pools' worth of water a day. 75% of that water will be reused and recycled right on site. Water and power were both big reasons why Micron settled on U.S. expansion. We have an area of plentiful fresh water, not just the Finger Lakes, but two Great Lakes, Lake Erie and Lake Ontario, and low-cost power generated primarily by hydroelectric and wind and solar. So we're ready for it. We know it's going to be a transition. The energy costs are, interestingly enough, lower in the United States than most parts of the world. People are more expensive in the United States, and so is the materials and the cost to build that factory. But that gap is narrowing over time.
Still, in Arizona, the world's advanced chip leader, Taiwan Semiconductor Manufacturing Company, recently blamed a shortage of skilled labor for delays to its massive $40 billion fab under construction 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 to us. We're attacking this from K through 12 all the way through post-graduate studies. We are big donors to the STEAM School in Syracuse, New York. We also run chip camp. In fact, my daughter, she's 13 and she went to the chip camp in Boise, Idaho this summer.
Still, the U.S. share of chip manufacturing has plummeted in recent decades. It costs at least 20% more to build and operate a new fab in the U.S. than in Asia. Labor is cheaper there, the supply chain is more accessible, and government incentives are far 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 thought up the idea for the Chips and Science Act at the gym in 2019 while talking with Republican Senator Todd Young. "And all of a sudden, when it came to chips so essential to everything we do, we had lost that edge. And if we didn't get back that edge, and not just on chips, but on science broadly, we would no longer be the number one economic power in the world. The prosperity in America would decrease. The number of jobs would go down. So I felt a passion about this."
The U.S. has especially lost that edge on the memory side of chips. Now, Micron is making big promises about changing that. Today, only 2% of the world's total memory production, DRAM production, is coming from the U.S., and all of that comes from Micron's fab in Manassas, Virginia. With Micron's investments through Chips support in Boise, Idaho, as well as in Syracuse, New York, that 2%, over the course of nearly 20 years, will be changing to about 15% of the worldwide production coming from the U.S.
In New York, Governor Kathy Hochul signed the $10 billion Green Chips Act last summer to entice companies like Micron, which says it's eligible for up to $5.5 billion from the bill. "If they hadn't passed the Chips and Science Act first, I don't think it would have been as many as incentives as necessary to bring them home or to encourage existing U.S. manufacturers to continue to expand, so I knew I had to woo them, talk about our incentives, but also we get out of it 50,000 jobs."
To make good on its promises in New York and Idaho, Micron is betting big that as the world relies more and more on technologies with vast data needs, they'll also need more and more memory. "The large language learning models and other things like that that continue to increase large demand. We're now moving into things like FaceTime, higher resolution images, movies on demand. All of that requires more and more memory to be made available."
Micron says construction in New York will begin at the end of 2024, and chip production will start in 2027, with some major promises about what that means for the country's share of memory manufacturing. "10% of Micron production today is produced here in the U.S., as these projects advance and complete, nearly 60% of Micron's production will be coming over the course of next couple of decades from here in the U.S."