저자: Alex · 외국인
요약 보기
AI주가 급락 속 코스피 폭락(서킷브레이커)의 원인은 중국 DUV 노광장비 양산 소식이다. 메모리 기업은 장기 고정가 계약으로 인해 가격 상승 이익을 못 누리고 있어 과도한 매도라는 주장. 메타의 클라우드 사업 진출 우려로 네오클라우드(CoreWeave, Nebius, IREN)가 급락했으나 메타는 오히려 이들의 순매수자이며 계약도 늘고 있다고 설명. 양자컴퓨팅(IonQ, D-Wave, Rigetti)도 실적·계약 성과에 비해 과도하게 하락했다고 주장. 결론적으로 하락장에서 분할 매수하라는 메시지.
다듬은 글
AI stocks have lost over a trillion dollars in combined value in the past two weeks, with some of them collapsing so quickly that circuit breakers were triggered , halting trading entirely. But while most of Wall Street is panicking, I see a huge opportunity to buy great stocks at even deeper discounts.
Here's a great way to get rich without relying solely on luck. My name is Alex, and this is how I made so much money in great stocks like Nvidia, Micron, and TSMC over the past 10 years. Let me show you what's really going on under the veil of panic and how I invest in it. Your time is valuable, so let's get straight to the point.
The main stock index of South Korea is called Kospi, it's something like our S&P 500, but for the Korean stock exchange. The Kospi index has fallen by more than 15% in the past few days and by more than 30% in the past month alone. In fact, it fell so hard and fast that it triggered a market breaker—an automatic halt that freezes trading across the entire exchange.
Imagine the panic if this happened to the S&P 500 index. It has happened to the KOSPI index at least eight times this year. KOSPI tracks 833 companies, but the two we are focusing on are at the very top. Samsung makes up over 26% of the total index, and SK Hynix another 24%. That's more than half of the entire index, represented by just two stocks.
And this is after their recent massive fall. The same thing that is causing this catastrophic decline in the Korean stock market is also affecting American stocks. Not because of the companies themselves, but because reports are emerging that China has begun mass production of its own DUV lithography machines, the first of which are expected later this year.
Lithography machines are the size of a small apartment, and they contain hundreds of thousands of parts. These parts work together to print microscopic circuits on chips using ultraviolet light. These machines can only operate in clean rooms, specialized sealed rooms with filtration and virtually zero dust content, as a single speck of dust falling on the wafer can obstruct the light and ruin the chip.
In fact, these machines are so specialized that practically only one company in the world can manufacture them. ASML is a Dutch company that controls almost the entire global lithography market. At least until now. Deep ultraviolet, or DUV, lithography is not precise enough to create cutting-edge GPUs or AI processors with sufficient yield of usable products, but it is precise enough to make memory.
NAND flash, DRAM, and even the crystals inside the high-bandwidth memory used in AI chips. But stacking these crystals is a completely different problem, much more complicated. And if China can establish mass production of memory, then Samsung and SK Hynix are in big trouble, because they control about 70% of the global DRAM market and almost 80% of the high-speed memory market.
That's why they're falling and dragging all the other AI stocks with them . Remember what I said a week ago when I warned about this market shock. These machines arrive in hundreds of crates and take months to assemble . And after that, they still need to be calibrated, tested, and tuned for the specific chips they will produce.
And this is not the only equipment in the process. Creating memory requires dozens of separate machines and hundreds of individual steps. So China needs to replicate the entire production process, not just one machine, and with a high enough output to really compete. This raises a completely logical question.
Should all AI company stocks fall just because competition in the memory market may intensify in a few years ? The best way to answer this is to look at the data . I believe there are three types of companies that are currently selling too much during this correction. Memory manufacturers, which is where it all started, AI cloud companies that rent out computing power instead of manufacturing memory, and quantum computing companies, which are a completely different type of business.
The thing is , all three of these types of companies are achieving significant success while their stocks continue to plummet. Let's start with memory. On July 13, SK Hynix shares fell more than 15% in a single trading day —the worst day in the company's more than 40-year history. Here's what actually happened.
An analyst from Korea Investment & Securities published a report explaining that most of SK Hynix's high-speed memory is locked into long-term contracts for a period of 3 to 5 years . Prices in these contracts are fixed 12–36 months before the first chips roll off the assembly line. This means that when memory prices rise, SK Hynix cannot increase the cost of products that are already contracted.
And memory prices have increased significantly over the last quarter. Prices for standard DRAM increased by about 30 %, and for NAND flash memory - by about 50% compared to the previous quarter . So, the main idea of the analytical note was that memory companies lost this profit by pre- fixing a large part of HBM sales with contracts.
So SK Hynix experienced the worst day on the market in its history because memory became more expensive, not cheaper— it became more expensive. And speaking of costs, you probably know that there are hundreds of data brokers that collect and sell your personal information. But did you know that it can cost you a lot of money?
Online stores are implementing personalized pricing, using your browsing history, location, and income to set prices just for you. And it’s easier than ever for them to get your data, as websites, apps, and AI tools request access to more of your information more often than before. But what surprised me most was that these data brokers had more than just my private data.
They also had my wife's and my entire family's data. This is where Delete Me, the sponsor of this video, comes to the rescue. Delete Me is an automated subscription service that removes your personal information from the databases of these online brokers. They provide a quarterly privacy report that shows everything they have done.
And at this point, they've checked almost 60,000 records for me. They even have a family plan so I can protect myself and my entire family. So, if you care about your family's privacy and like to save money, you can get 20% off any consumer plan with my code Symbol 20 by going to joindeleteme.com/symbol20 or the link in the description.
OK. So, SK Hynix had its worst market day ever because memory became more expensive and their contracts did not allow them to fully take advantage of this increase. And this applies not only to SK Hynix. SanDisk has disclosed information about the minimum value of approximately $42 billion in contracts they signed in the last quarter.
These contracts cover more than a third of the memory bit volume that SanDisk expects to produce this fiscal year. And Micron has 16 strategic customer agreements totaling over $100 billion in minimum contract value. This covers approximately 20% of their DRAM and a third of their NAND flash memory .
These are huge contracts that prevent such large profits. So shares of memory companies have fallen in the past few weeks. But look what's happening now. The main reason for the current downturn is that China may be able to produce its own memory using domestic DUV lithography machines. This increases supply, and therefore memory prices should fall.
But the same contracts that prevent companies from making excess profits also protect them from falling prices. Exactly what the entire market is worried about. SK Hynix shares have fallen about 25% in the past few weeks on the Nasdaq and about 50% in the past month on the Korea Exchange. Micron shares have fallen about 35% in the last month and 25% in the last week alone.
And SanDisk shares have literally halved in the last 30 days, even though contracts protect all three companies from falling prices. And it will be many more years before China can use its DUV lithography machines to mass produce competitive chips. I'm not saying that these memory stocks have no risks at all, but I'm arguing that it's too early to price these risks in, at least in my opinion.
Another risk that is taken into account too early is Meta Compute. Earlier this month, Bloomberg reported that Meta is planning its own cloud business, called Meta Compute, to lease out to third-party companies any additional AI infrastructure they build. This report sent Meta shares up nearly 9% in a single trading day, while sending shares of “neo-cloud” companies like Coreweave, Nebius, and IREN down 14–17% each.
But the decline for these companies did not stop. Coreweave shares have fallen more than 35% this month and nearly 25% in the past week alone. IREN also fell by more than 35%, and Nebius by more than 40%, making it one of the steepest declines these three companies have ever seen. It's worth asking the same question here.
Did these three companies really have to fall so badly? Let's think about it from a fundamental perspective. “Neo- cloud” companies rent computing power to firms that want to use AI but don’t want to spend billions of dollars building and maintaining their own physical infrastructure. So, these “neo-clouds” look for grid connections, build or buy data centers, fill them with racks of GPUs and networking equipment, and make money by renting it all out when the system is up and running.
This means that “neo-clouds” have to risk a lot of money up front and hope that demand remains sufficient once everything is up and running. If a company is still young and unprofitable, it needs to borrow money, usually by taking out loans at high interest rates or diluting shareholder equity, both of which are bad for the stock.
But Meta Platforms does n't have that problem. Because it's already a trillion- dollar tech giant with huge profits. So Meta can build as much computing power as it wants and rent out what it doesn't use itself, without having to raise additional capital or take out loans on unfavorable terms. So, it's the end of the game for "neo-clouds," right?
Meta gets all the benefits with almost none of the downsides or risks. Well, not exactly . First, Meta just reported earnings on July 29th. During a previous earnings conference call, Mark Zuckerberg said that entering the cloud market is definitely under consideration. Regarding this, he said that they are receiving offers to rent out their computing power at a large premium to what they paid for it .
But he believes it would be unwise to simply sell off your computing power for short-term profits. Meta can earn much more by selling ready-made intelligent services, such as agent models and programming tools, instead of renting out “ raw” hardware. As a result, Meta uses all of its servers on its own and does not have any additional computing power to rent.
In addition to this, Iren and Nebius signed billion-dollar contracts after this report was released. On July 14, Nebius agreed to sell over a billion dollars worth of computing power to Reflection AI. The contract will run until 2029 and provides access to Nvidia GB300 Blackwell Ultra chips . Then, on July 20, Iren signed new multi-year contracts for cloud AI services worth $2.8 billion and raised its annual revenue forecast.
And don't forget that Meta actually has $35 billion in commitments to Coreweave and up to $27 billion in additional contracts with Nebius. This makes Meta a net buyer of Neo's cloud computing, not a net seller as the market currently believes. So, Meta's computing power is a great way to hedge against the excessive costs of your own AI data centers, but it won't happen anytime soon.
This is why I continue to buy shares of Iren, Nebius, and Coreweave, especially as their price continues to fall. And I'm not the only one. On July 20, Envidia filed a report with the SEC, stating that it owns more than 22 million shares of Nebius, representing about 9.3% of the company. To me, this means that Envidia is betting not only on Nebius, but also on the broader idea that neo- cloud is worth investing in directly.
And that brings me to the third group of stocks— quantum computing. And if you think I deserve it, consider hitting the like button and subscribing to the channel. This really helps and makes me realize that I should make more videos like this. Thank you. Now let's talk about IonQ, D-Wave, and Rigetti, as they have all fallen 30–40% in the last few weeks.
On July 13, the same day the Korean Stock Exchange halted, all three quantum computing companies fell by almost 10%. No news, no disappointing earnings reports or delays. So let's ask the same question a third time. Did these three stocks really have to fall so much? First, there is no significant connection between quantum computing equipment and Korean memory .
These machines do not compete with memory, do not buy it in large volumes, and do not sell services to the same customers. Quantum computing stocks simply moved along with the rest of the market . Second, all three companies had significant achievements during this downturn. On July 27, AT &T signed an agreement to expand the use of D-Wave quantum systems in its network.
AT&T plans to use these systems to detect outages, route technicians, manage network traffic, and even plan construction. In one of the first applications, AT&T reduced the time it took to complete a network optimization task that would normally take an hour to less than 15 seconds. This is about 240 times faster thanks to the D-Wave quantum processor.
On the same day, Rigetti expanded its collaboration with HPE to create a hybrid quantum-classical testbed at the Pittsburgh Supercomputing Center. Construction will begin on September 1. And a day later, IonQ cleared the final regulatory hurdle to acquire Skywater Technology, an American semiconductor factory; the deal is expected to close just as this video is being published .
So, IonQ is spending money to secure a fully internal supply chain to accelerate its own roadmap. All three companies report earnings in early August. So let me know in the comments if you want me to make another video about quantum computing. Okay, here's a table summarizing everything I've said. As you review it, consider a few points.
I compiled this table myself and tried to make the comparisons as accurate as possible, but it is not perfect. For example, I show price changes over the last month, but SK Hynix only went public on Nasdaq on July 10th. Therefore, I use the price on the Korean Stock Exchange, which is about a day ahead of the American one.
Also, all of these companies have different fiscal years, so I'm using the trailing 12- month revenue growth rate. And, of course, they have different contract terms with different clients and fundamentally different technologies. So, take this as a reliable summary table, not as official audit data.
Here's what stands out after I put it all together. Each of these stocks fell by 30–55% in just one month. Memory manufacturers, cloud AI companies, and quantum computing. Three completely different markets with three completely different types of achievements and risks. And almost every one of them has grown their income at a triple-digit rate.
Nebia grew by 453%. IonQ grew by 335%. These are not the companies that have problems. Their stock price fell by half while their business doubled. And then there's Rigetti, with a 34% year-over-year drop in revenue , but that figure hides something important. Their last quarter actually grew 199% thanks to on-premises system sales and government contracts .
Revenue from early- stage quantum technologies is always volatile, as it comes from the sale of individual systems, research grants, and cloud access milestones, rather than from stable, regular revenue . A large system may ship one quarter and not the next. So, if you're investing in quantum computing, you already know that it's going to be a long and bumpy road.
By the way, about thorny paths—a small note regarding this decline. If you look at the price dynamics over the last month, you will see many days when stocks soared by 10, 15, or even 20% in one day, only to lose it all the next week. A green day in the midst of a decline can seem like a bottom, and sometimes it is, but sometimes it's just a respite before another collapse.
That's why I always gradually build up these positions and always keep some money in reserve in case of further declines. This is a great way to get rich without relying on luck. And if you want to see more stocks I buy to get rich without getting lucky, watch this video below. In any case, thanks for watching, and see you next time.
It was the ticker U, my name is Alex, and I remind you: the best investment you can make is an investment in yourself.
Comments 0