원문 보기
Micron investors just received one of the strongest customer endorsements ever, and it came directly from Elon Musk. During Tesla's earnings conference, Elon Musk confirmed that Micron is allocating significant amounts of memory to them for many years to come. This is much more important than just announcing one client.
Memory has become one of the biggest bottlenecks in artificial intelligence, self-driving cars, and robotics. And now the world's largest companies are fighting for access to supplies for years to come. So today I want to talk to you about why I think Micron stock is still worth buying even after its meteoric rise.
Let's discuss this huge opportunity in the field of memory. First of all, we need to understand the amount of DRAM used in every Nvidia chip—and mind you, not just Nvidia. This applies to TPU, AMD, and every single custom silicon or ASIC, which is seeing a huge increase in memory requirements. Simply because it is necessary for the products we use.
Generative artificial intelligence and all these technologies are showing rapid growth. So, we expect about a tenfold increase between 2022 and 2028. With Nvidia's revenue growing 85% and probably 100% next quarter, reaching $93.5 billion, it shows that in addition to tenfold growth per chip, we are selling more chips.
So it's much more than just a tenfold increase. This is mainly due to the fact that the most popular models are the most demanding on HBM bandwidth. That's why all new chip designs, like Vera Rubin, are more focused on memory and its capital costs compared to previous architectures. Last quarter, we saw revenue of approximately $33.5 billion.
Gross margin was 81 %, and earnings per share were about $19.15. These were Micron's figures for the past quarter. We saw $41.4 billion, and mind you, that's almost identical to what we saw in the previous quarter. Were these results that exceeded Wall Street expectations and Micron's own forecasts by 20%?
They forecast revenue of $33.5 billion to $35 billion . These were their forecasts for the May quarter. And, obviously, Micron exceeded those expectations. Furthermore , the 81% margin they claimed actually reached approximately 84.56%. This means that the cost of producing what they sell is only about 15% of the total cost.
And that's up from 60% of the total value or 63% just a year ago. So, this really shows the limited supply. Note, this is not based on new volumes. The offer that we actually give to the market is approximately the same as last year, because there are no new plants. All production lines are operating at full capacity.
Now they need to build a new factory. The new plants will not start operating until the end of 2027. So that's all we have. So where does this gross profit come from? Average selling price. Different customers say, "Hey, we're willing to pay higher and higher prices. Please meet our demand. We'd be willing to pay more to get the product from you." This also results in the average selling price increasing, but the actual costs of operating the plants that are already in operation, supporting offices, sales departments, etc.—remaining about the same.
And so the costs of all this are falling. R&D costs, administrative and general expenses, stock compensation costs— all of these even go down because you don't have to pay more for all that growth. This makes operating profit skyrocket. You got 84% gross margin and 80% operating margin after all the expenses mentioned.
In addition, you have a consolidated net profit of approximately $28 billion and it remains high, at 66-67%. Extremely high growth. Again , compared to last year, net profit margin more than tripled—from 20% to the mid-60s. You grow a company from $1.8 billion to $28 billion in total net profit. This all affects the final result.
So that initial forecast that we saw last quarter—$19.15— actually turned into $24.67 per share, which beat Wall Street estimates and their own forecasts by 22-28%. The results here are simply impressive. You won't see this in many companies. If you look at traditional businesses, you can point to Microsoft or maybe Nvidia more recently.
They can exceed revenue by 3-4%, and EPS by 5-8%. EPS can be a little more difficult to estimate because there are tax implications and all these things that you can do with accounting shenanigans. Exceeding revenue by 18 % and earnings per share by 25% is something incredible. This is something that only the best companies in the world have.
If you look at their cash reserves, taking advantage of a period of higher average selling prices, Micron increased its cash from $10 billion last year to $26 billion. They also paid off their debts. The debt was reduced from 15 billion to 5 billion. So, $10 billion went to pay down debt, and another $16 billion was simply added to their cash reserves.
So when you look at this, what are they going to do with operating cash flow and everything else ? Capital expenditures for the construction of new facilities. We even have new facilities opening in New York, many in America and around the world. Let's try to get more access to the NAND, DRAM, HBM that we love so much.
And yet, despite the large capital expenditures, free cash flow still grew to over $17 billion . Last quarter it was less than one-tenth of that . We have come a long way from 1.6 billion, and a year earlier from 400 million. So: 400 million, 1.6 billion, 17.5 billion dollars. All of this would be extremely worrying if demand peaked and began to fall precipitously.
An interesting aspect is the forecast for the fourth quarter. The forecast for the 4th quarter expects a figure of $50 billion, plus or minus $1 billion. Gross profit margin remains at 86%, and diluted earnings per share is around $31. If you look at the big picture, growth continues. Going back to Q3, they beat revenue forecasts, which were already expected to be 33.5 billion.
They exceeded that by 18%. So, this is their current forecast. Imagine what an extra 10%, or an extra $5, would look like. Or if the overperformance is 18%, we could crush all forecasts, approaching $60. However, Wall Street—and these are the current figures for today— predicts that this growth will not slow down until at least mid-2029.
So throughout 2027 and until 2028, these indicators will grow, and we will only talk about a possible decline in the years after that. They also state that they plan to establish new DRAM and NAND production in the second half of the 2027 calendar year. They are using the massive scale-up of HBM 4, so their 12-layer stack , which is growing twice as fast as HBM 3E, has already generated over a billion dollars in revenue.
So, it's not just about their two flagship products—DRAM and NAND— they are also expanding, and these areas are also showing good results. They also announced the signing of 16 strategic agreements with customers totaling about one hundred billion dollars, and have since signed many more contracts with automakers.
It seems that four more automakers have agreed to supply memory, since modern cars are essentially computers. But all this affects the actual cost estimate. A growth leap that brings us today to a forward P/E ratio of approximately 6.4 times. 6.4 times—this massive collapse was driven by profits and an overestimation of future growth, and now we are growing again as the share price has clearly corrected from $400 to today's $970.
As we continue to increase the overall price and revenue is growing rapidly, this forward P/E is almost unchanged, as EPS growth makes Micron shares cheap, even despite their rapid appreciation. This is the strangest thing about this situation, because usually when stocks grow by 150% in half a year, right?
What we actually see here is that stocks usually become more expensive. But we don't observe this. UBS just announced that, thanks to the amount of cash generated, the company can buy back about 40% of its shares by the end of 2028. So while investors are trying to estimate when this development will slow down , we don't know for sure.
We are only making assumptions, but this requires knowledge of where agents will be and how much we will pay them in the future, which is difficult to predict years in advance. But while this is happening, they are getting real cash flow. UBS expects Micron to generate more than $40 in free cash flow per share by 2028, and after the buyback cap expires on December 9, the company could potentially use those funds to buy back more than 40% of its shares at the current price.
So if they think the stock is too cheap, and their cash reserves are already around $26 billion and growing rapidly, they may well put that money to work. And I don't think people really realize how massive these buyouts can be. So, we now see Micron's stock price as estimated by 29 analysts, with one recommending a hold, none recommending a sell, and the average price is around $1,568 per share, giving us roughly 81% potential upside.
Although I have partially withdrawn my initial investment from Micron and am now playing only with market money , we initially formed this position at a price of around $370. As I said, earnings have consistently exceeded our expectations, so a price below $1,000 is starting to look interesting, especially with reports that Kimmy K3 is fully loaded and they are pausing new subscriptions.
All of these new models also require a huge amount of memory. Regardless of the type of model and the ability to measure their intelligence, one thing remains the same: memory is still necessary for agent work, right? Just like the CPU. While GPUs certainly do the bulk of the complex work in AI, now with the use of AI tools you need CPUs for all the basic agent needs, right?
So, it scales with them. It will be exactly the same with memory. Whether you use Anthropic, Kimmy K3, or something else, you're going to need a ton of memory because it's going to scale based on the tasks you're actually doing with these AI agents. I am optimistic about Micron. I think if we continue to see a sell-off and people start to doubt this stock, I might decide to buy more.
Maybe not 970, but closer to the $800 range. I think I can afford to be a little greedy and try to find such points to buy. If not, I've already formed a pretty decent position and will likely hold it for the next year or two, and then we'll see how I feel when there are updates about new plants and other such things that might change my thesis; but now I act according to the circumstances.
Comments 0