MU 因当前盈利高企、远期市盈率仅 6.8 倍而被视为被低估的“价值股”，AI 需求提供支撑，但若供给超过需求，未来利润率存在被压缩的风险。

This is Micron, a $1.2 trillion semiconductor company that's benefiting from the current AI data center buildout. And it's only trading at 6 times 2027 earnings.

So today I'm going to dive in why Micron is trading at 6.7 times next year's earnings and why despite going up 1,400% over the past 5 years and nearly 500% over the past year, the stock is actually cheaper today based on its earnings.

The company just reported record earnings per share up 1,000% year-over-year in the latest quarter. And this growth doesn't look like it's going to stop in the near term.

So today I'm going to give my thoughts on Micron stock after their latest earnings report. I'll explain what's driving the massive growth in earnings and cash flow and whether this high-rising memory stock is part of the potential AI bubble or if the market's not being aggressive enough and is pricing this stock too cheap.

Micron, ticker symbol MU, is a well-known semiconductor company. They make RAM, DRAM, flash memory. This includes the HBM memory that is in the AI infrastructure products that companies like Nvidia make.

And the three main memory makers are Micron, Skhinx, and Samsung. All of which are now trillionoll companies, and they have been making a killing.

Right now, we're in the middle of a memory shortage. The memory business has not been a good business for many decades. It had boom and bust cycles. It typically sold at a very low margin.

It was considered a commodity product that's replaceable.

So there was not a lot of investment in increasing infrastructure and production. Micron, Skhinx and Samsung, they actually manufacture the memory.

Unlike the chip designer companies, the fabous semiconductor companies like Nvidia, AMD, Broadcom, these guys actually make the memory and they're investing to build out some new plants, which I'll get into.

So what happened is that companies like Micron, SKH Highix and Samsung realized they could charge a higher price for their memory and that the demand was much higher than the supply.

And you know raising prices is very easy to do when it's being packaged and sold into the biggest hit product in the history of capitalism which is these AI data centers where hundreds of billions of dollars is flowing to.

So here you can see the margins for Micron historically and it's gone up and down with the different cycles. As I mentioned, there's boom and bust cycles in this commodity, but probably on average, they've had a 30% gross margin over the decades.

Well, something crazy just happened this year. The company raised their margins, and it's now 86.8% gross margins, so more than Nvidia, 69.5% net margins.

So, they have realized that demand is far ahead of their supply available. and they have basically been selling the same amount of product and now they're starting to invest some of those profits into building out more supply.

But for now, they just rad the prices up and they are making a killing. And the entire game right now with forecasting these memory stocks is if those margins stay this high, if they continue even going up, which is crazy to say, and for how long.

Also, the thing that most people are not factoring in is the fact that when they bring more supply online, there still might be an exponential growth in demand that causes them to keep these very high margins, at least for their AI focused business.

And I think the market is definitely discounting that potential bullcase.

So, in this latest quarter, they just reported, they had 54.23 billion of revenue, up $379% year-over-year. So, Q4 last year was 11.3 billion of revenue. And now, mostly just from raising prices and charging more, they have increased it to $54 billion.

If we look at the earnings per share for Micron stock, it's kind of crazy. They just reported earnings. It was $3342. that beat consensus by 5.2%.

And look how crazy this number is. That is up 1,000% year-over-year. Q4 2025 was $3.3 in earnings per share for that quarter. Q4 2026 just came in at $3342.

Over the trailing 12 months, so the past four quarters, they've had $75.51 of earnings per share, up 810% from the prior trailing 12 months. and they trade at roughly a 14 trailing PE ratio.

But as you can see, earnings are going up every single quarter still. And when they bring more supply online at these higher margins and if they continue raising prices, which you know, there's only so much room to do and companies like Nvidia, they lock in longerterm contracts and agreed upon price.

But there's definitely room for this to keep going up over the next year or two.

So, compared to those trailing 12 months for the full year 2026 that I mentioned earlier, fiscal 2027's projected to go up 116% in earnings per share to $1605.

And this actually seems pretty reasonable cuz the latest quarter was $33. So, if you just annualize that and expect even no growth next year, which is not going to happen, that would be $133 for the full year of earnings per share.

But here's the thing. Based on today's stock price, 4 micron, it's trading at 6.8 times next year's earnings. This is actually like unheard of in most public market stocks, especially if you're a $1 trillion valuation.

And this will lead me into giving you the case of how Micron might actually be a value stock in 2026 going into 2027. But I will also give you the downside, which I think the market is pricing in much more right now, which is crazy to say for a stock that's gone up 500% over the past year.

So here is Micron stock over the past year. And you'll notice something interesting. This stock has kept getting cheaper. It's trading at a lower P ratio even though the stock price has gone up so much throughout the year.

And that's due to the rapid earnings per share growth. So this current 14.6 trailing P ratio, it's actually one of the cheapest of the past year. And if we expand this to the past 2 1/2 years or so, you can see that even though Micron stock price has gone up so much in that time, the actual earnings power of the business has grown so much, it's actually cheaper today based on its earnings.

And if we look at the stock price really since June, the stock hasn't gone anywhere. The price has basically been flat while the business in that time continued rapidly growing.

And it's not just on paper revenue and earnings per share. Micron's actually bringing in a ton of free cash flow right now.

And that's despite them investing in expanding their capacity of production, which is expensive. So, over the trailing 12 months, they've earned 58.96 billion of free cash flow, which is up 3,400% year-over-year, which is insane.

They earned $ 32.86 billion of free cash flow just in the latest quarter. And for contacts, operating cash flow in that quarter was $43.97 billion, up 667% year-over-year.

So, their capex, while it's been growing and was just 11.11 billion in the latest quarter, it's not crazy growth. It's only up 96% year-over-year.

So, for the purposes of shareholders right now, Micron is just raking in money. You can see their cash and cash equivalents and short-term investments are at an all-time high right now. Just hit 43.4 billion.

And that's just going to keep stacking. And if this lasts a few years, Micron's going to have one of the best balance sheets in the entire world. Their net debt currently sits at $38 billion.

So, they have far more cash than they have debt. And this is the best in the company's history by far.

Now what's interesting is that Micron share count has actually continually gone up over time. They haven't been a big share repurchaser. So this could be an interesting thing if in the next year they announce a big trunch of share repurchases. That could be possible.

So now I want to explain like the most important thing for forecasting how Micron will do in the next few years. And this really ties back entirely to the AI data center buildout and Micron's role in it.

Afterwards, I'm going to give the latest price targets from Wall Street analysts for Micron stock and I'm going to give my thoughts on Micron's valuation and where the stock could end up being if they hit their future earnings estimates.

So, how does memory and what Micron makes actually relate to the AI data center infrastructure?

Now what's interesting is that when we started building out these AI data centers and we were in the early steps of training and serving these models like 2023 2024 time period we didn't necessarily see this kind of shortage happening.

Memory was used a lot in training the AI models but not that much in the one-time use of it. But there's been two basically huge things that have caused this explosion in need and usage around memory and that's the idea of AI reasoning and then recently AI agents.

So reasoning was an innovation where if you found you just let the AI model talk to itself and continually prompt itself, generate tons of tokens and continue thinking about something for a long period of time, you actually could get significantly better results.

And now that kind of intuitively makes sense, but people weren't really thinking about that in 2023.

And this basically increases the model's working memory. It has to do with the KV cache if you're into the technical part of it. And basically, as this context is built up, the model reads and writes it into memory.

And that way, you don't have to continually recmp compute tokens as well.

So the idea of reasoning was the first thing that started to really cause this increase in memory usage. And then this year, you have AI agents that are starting to become a very huge growth driver.

So you have way more things that are continually running, sometimes not even being prompted by humans.

And it's this new demand signal that basically has allowed Micron and all these memory companies to raise their prices significantly.

And that's why they're having this rapid increase in earnings and are now each trillion dollar companies. But is this going to continue? Because that's really the big question if you're analyzing Micron stock.

And I'll give you kind of two different scenarios here. I'm not necessarily going to give you the AI is going to explode next year and it's all over. Data centers are important.

They're not going to be built frankly because I don't think that's actually reality. But I guess some people kind of believe in that.

But there are two scenarios that I kind of see here with Micron stock.

The first one's probably the most bullish one and that's really the idea that we're dealing with exponentials here. The models exponentially improving. So intelligence on the frontier becoming exponentially more capable and then also over time bring that capability levels to exponentially cheaper, more affordable, easier to run models as well.

And you're already seeing this with some openweight models being fantastic, especially compared to what we had a year or two ago and they're able to be served for very cheap. But here's the thing, when that kind of thing happens, if you're in the space, you probably heard this a thousand times, but the idea of Jeban's paradox, the cheaper and more capable something becomes, the more it actually gets used.

So in my opinion the most likely thing is that there is a dramatic and exponential continued growth of AI usage and even right now this is a demanddriven phenomena.

There's a compute shortage in general they don't have enough infrastructure built to serve the demand and as new capabilities continue to be unlocked and they get made extremely affordable.

People are going to want to use intelligence in far more applications and memory is going to be a part of that equation.

But then I'll give you some potential downsides. Micron is currently a constraint on some of the smartest companies in the world. Companies like Nvidia.

So AI infrastructure is getting more expensive. That impacts the model companies. That impacts OpenAI. That impacts Enthropic. That impacts XAI. It impacts companies like Google.

So the biggest thing going against Micron and these memory makers right now is that all of the smartest, most technical companies that are actually driving this like whole super intelligence revolution, they don't want this cost to continually grow and get insane.

So there are continually new innovations into try and make KV cache more efficient, try and make memory less of a problem. So there could be continued model improvements that aim directly at that goal.

So Micron, no matter what way you square it, that's going to be one of the companies and they are investing CFX in building out more infrastructure. But this takes years and years to bring online and they're bringing it online much slower than companies like Nvidia and all these AI companies actually want them to because they realize all the demand signals right now are pointing to continued exponential increases.

So when we talk 5 years out with Micron, SKH Highix and Samsung expanding their production, there definitely will be far more memory produced 5 years from now than is today.

But the entire question is whether demand continues growing massively in that time. And there is a potential reality and scenario where despite these companies investing and increasing the supply, this just might be a big super cycle and they still are behind demand.

So, they're just able to sell way more product at still very high margins. That is a thing that could happen that I don't think the market is properly pricing in.

So there is a possibility over a 10-year view, maybe 5 to 10 years, that Terrafab with the combination of increased supply from these suppliers, they could end up growing memory supply so much that they crush the price back down.

It becomes once again a low margin commodity product. And that is a real thing that could happen.

And since I'm more in the AI bowl on the demand side, that's more the scenario where I could potentially see downside that the whole question basically is supply and demand over the long run.

But that really is the whole downside argument for Micron stock and why it's trading at such a cheap multiple. Historically, it's been a boom or bust commodity.

And even though Micron might tell you their memory is great, there's nothing necessarily special about it. Even though Micron has rebranded as being a AI company now because that's where most of the money is being made for them.

They are in my opinion not necessarily leading any part of this revolution other than from the infrastructure side. The innovation in my opinion is not happening at Micron. They are not the ones leading the AI revolution.

Then of course all the model companies, even the hyperscalers, private and public model companies, OpenAI, Anthropic, Micron, and these memory makers, they just make a component piece of the whole AI data center infrastructure.

But given that they're not actually leading that whole movement, I don't think they have very many competitive advantages, especially when you mentally just compare it to a company like Nvidia, which is really driving the whole industrial revolution here.

It's being priced right now like these earnings are temporary and any year could collapse completely and they go back to that 2025 maybe half of 2026 levels or potentially even negative if there's a massive overupp.

We just had a flurry of it post earnings. We had a $1,400 price target, a $1,500, $20 price target, $1,300, $1,400, $1,300, $1,250, and one analyst even with a $2,100 price target.

These are 12 month price targets, by the way. So, what they're expecting the stock price to be 12 months from now, one year from now. And for context, the current stock price is $1,92.

Now, investors will often talk about multiple compression, and that's kind of what's happened here with Micron stock. It's trading at a very low multiple right now and that can happen for a variety of reasons but one of them would be some kind of cyclical business which Micron historically has been and they're in a period of outear earning.

So these are like temporary high earnings and in those scenarios the stock might trade at a cheap multiple but if earnings collapse on the other end then they're not necessarily trying to overvalue the price and pay 20 times that peak earnings that's not sustainable.

But right now, forward-looking earnings is 6.8 P ratio for Micron. And if they just hit these earnings per share estimates among Wall Street analysts, even if they stay at a 6.8 P ratio, I'd be 77.2% upside from here through 2030, 15.8% annualized.

And again, that's keeping this ridiculously low multiple just having the actual fundamental earnings growth. And these earnings numbers might be conservative. Let's say that they even trade at 10 times and that would give a projected 2030 price of $2,821. 160% upside from here, 27.8% annualized.

Let's say it's 15p ratio. Let's say the market gets crazy, starts to get a little bullish. That'd be a projected price of $4,200 by 2030. 290.8% upside from here. 41.8% growth annualized. And a 20p ratio gets even crazier.

And frankly, I have no idea. All I know is in 2027, they're going to make a crap ton of money. That is also likely to happen in fiscal 2028. Beyond that with memory, I really don't have any idea.