Micron Technology’s stock just went on an absolute tear, surging 91% in 2025. The stock jumped over 41% in just one month.
But here is the thing. While everyone seems to be obsessing over Nvidia and the GPUs , Micron has been positioning itself as the unsung hero of the AI revolution. And Wall Street is actually taking notice.
So what is going on?
The Memory Game
Let’s start with the basics.
When you hear “AI,” you probably picture super-powerful processors doing complex calculations. And that’s true — they are the brains.
But here’s what most people miss: brains alone aren’t enough. AI systems also need memory — a lot of it. And not just any memory, but ultra-fast memory that can keep up with the processors. Without it, even the smartest chip can get stuck waiting for data.
That’s where Micron comes in. They make High Bandwidth Memory (HBM) — think of it like a super-fast express lane for data. Regular memory is more like a narrow country road, where traffic moves slowly and jams easily. HBM lets the processors get the data they need instantly, which is exactly what makes today’s AI systems run so fast.
And business appears to be booming. Micron’s HBM revenue hit nearly $2 billion in Q4 alone and that’s an annualized run rate of $8 billion. To put that in perspective, their memory products for AI contributed $10 billion in revenue for fiscal 2025, representing a more than fivefold increase year-over-year.
The Numbers Tell a Story
Micron just wrapped up fiscal year 2025 (ended August 28, 2025), and the results were impressive:
- Revenue: $37.4 billion (up 50% year-over-year)
- Gross margin: 41% (expanded by 17 percentage points)
- Earnings per share: $8.29 (soared 538% year-over-year)
- Data center revenue: 56% of total revenue with 52% gross margin
But here’s where it gets interesting. When you think of AI or computers, you probably imagine super-powerful processors doing all the heavy lifting. But even the smartest chip is useless if it can’t get data fast enough. That’s where DRAM, or Dynamic Random-Access Memory, comes in. Think of it as the short-term memory of your device — it temporarily holds all the data the processor needs right now, whether it’s open apps, games, or AI calculations.
Unlike storage drives, DRAM is fast but forgetful: turn off the power, and everything disappears. Over the years, DRAM has evolved into different types: standard DDR DRAM powers most PCs, LPDDR is optimized for smartphones, and High-Bandwidth Memory (HBM) feeds data-hungry AI servers at lightning speed. With more PCs, phones, cars, and AI servers demanding memory, and supply struggling to keep up, prices have soared — which is why Micron is seeing record DRAM revenues and margins.
The Supply Crunch & Micron’s Moat
Now, you might be wondering: why can’t competitors just flood the market with memory chips and kill Micron’s margins?
Great question. Here’s the thing: building advanced memory chips isn’t like opening a lemonade stand.
First, there are three major players in the global memory market: Micron, SK Hynix, and Samsung. That’s it. The barriers to entry are astronomical we are estimating about $18 billion in annual capital expenditure just to stay competitive. New fabs take years to build and cost billions. You can’t just decide to become a memory maker overnight.
Second, HBM is exceptionally difficult to manufacture. You’re literally stacking memory chips vertically as it requires cutting-edge packaging technology, precise engineering, and years of R&D. Micron has already moved through HBM3E and is sampling HBM4 to customers, while Samsung is still playing catch-up trying to get its HBM products validated by Nvidia and other hyper scalers.
Third, there seems to be a massive supply-demand mismatch happening right now. Micron expects DRAM bit demand to grow in the high teens and NAND demand to expand in the low to mid-teens during calendar 2025. But their supply growth? It’ll stay below industry demand for both products.
Why? Because all three major players learned from past cycles. They’re seemingly being disciplined about capacity additions, avoiding the overbuilding that destroyed margins in previous downturns. Management explicitly highlighted this disciplined buildout strategy on their earnings calls.
| Thing | DRAM | NAND |
| What it is | Volatile memory → forgets everything when power is cut. | Non-volatile → remembers even when the device is off. |
| Role | Short-term working memory. Let’s run apps on the CPU/GPU and multitask. | Long-term storage. Holds files, OS, photos, and videos. |
| Speed | Super fast (nanoseconds). | Slower than DRAM but still much faster than HDD. |
| Size | Smaller (GBs, some TB in servers). | Bigger (hundreds of GBs to multi-TB). |
| Where | PCs, laptops, servers, GPUs, phones, and gaming consoles. | SSDs, phones, USB drives, memory cards. |
| Use cases | AI training, multitasking, gaming, and cloud workloads. | Data centres, phone storage, IoT, and autonomous cars. |
| Analogy | Like your brain’s short-term memory – it helps you think quickly but forgets fast. | Like long-term memory – stores things permanently. |
| Thing | DRAM | NAND |
| What it is | Volatile memory → forgets everything when power is cut. | Non-volatile → remembers even when device is off. |
| Role | Short-term working memory. Lets CPU/GPU run apps, multitask. | Long-term storage. Holds files, OS, photos, videos. |
| Speed | Super fast (nanoseconds). | Slower than DRAM but still much faster than HDD. |
| Size | Smaller (GBs, some TB in servers). | Bigger (hundreds of GBs to multi-TB). |
| Where | PCs, laptops, servers, GPUs, phones, gaming consoles. | SSDs, phones, USB drives, memory cards. |
| Use cases | AI training, multitasking, gaming, cloud workloads. | Data centers, phone storage, IoT, autonomous cars. |
| Analogy | Like your brain’s short-term memory – helps you think quickly but forgets fast. | Like long-term memory – stores things permanently. |
Translation? High barriers to entry, oligopolistic market structure, technical complexity, disciplined capacity management, and tight supply equal strong pricing power and fat margins.
The company has already locked in pricing agreements for most of its 2026 HBM3E capacity through fixed contracts. They expect to sell out their entire FY2026 capacity within months. CEO Sanjay Mehrotra confirmed they’re negotiating HBM4 pricing, which is “significantly higher” than HBM3E pricing.
The Tech Edge
But Micron isn’t just riding the wave and they are trying to stay ahead of it. They are already shipping HBM3E while sampling HBM4 to customers. Their HBM market share is expected to align with their overall DRAM market share in Q3 of calendar year 2025.
The technology roadmap is aggressive. They’re transitioning from 1-gamma DRAM nodes (which will contribute the majority of supply in 2026) to 1-beta nodes for HBM production.
In simpler terms, they’re constantly making their memory chips smaller, faster, and more efficient. On top of that, they’re developing HBM4E, the next generation after HBM4, in partnership with TSMC, which provides the base logic die — the part that connects the memory layers and helps them work together seamlessly. These upgrades mean Micron can deliver more powerful, higher-capacity memory for AI servers and other cutting-edge applications, keeping them ahead in the race for faster, smarter technology.
And they’re investing heavily. Micron spent $13.8 billion in capital expenditures in fiscal 2025 and expects to increase that to $18 billion in fiscal 2026, a 30% jump. They’re building new fabs in Idaho, expanding capacity in Japan and Singapore, and racing to meet demand.
They’ve also received a $6.2 billion government subsidy under the CHIPS Act and have already gotten their first disbursement after completing a construction milestone in Idaho.
The Bull Case
So why are investors so excited? The math is compelling:
Hyper scalers are expected to spend $400 billion in Cap Ex for fiscal 2025, with more coming in subsequent years. Some analysts estimate the HBM market alone could reach $100 billion by 2030.
Wall Street analysts are forecasting Micron’s earnings per share at $11.59 for fiscal 2026 and $12.47 for fiscal 2027. At the current stock price of around $165, Micron trades at just 12.4 times forward earnings.
The Bear Case
But not everyone’s convinced this is a sure thing. Here’s what skeptics worry about:
Memory has historically been cyclical. Micron’s free cash flow margins compressed significantly during downcycles over the past decade. What makes this “supercycle” different?
The stock’s price action is also concerning. After the breathtaking rally to $158 and subsequent push toward $160, sellers stepped in. Some technical analysts note the vertical spike resembles the double-topping pattern from late 2021 which preceded a brutal decline.
There’s also the enterprise adoption question. While data center infrastructure spending is robust, are SaaS companies and enterprises seeing enough AI-driven gains to justify this level of investment? Besides Palantir and Microsoft, broad-based enterprise AI adoption remains unclear.
Competition is heating up too. SK Hynix just hit new all-time highs and is ready to move HBM4 into volume production. Samsung, after falling behind, is making a comeback in the HBM market. More competition typically means pricing pressure.
And then there’s execution risk. Despite management’s confidence, HBM4 recently underwent specification changes. Any production delays could hurt their competitive position.
What do you think? Is Micron’s memory super cycle different this time, or are we setting up for another painful downturn?
Disclaimer – This article draws from sources such as the Financial Times, Bloomberg, Seeking Alpha and other reputed media houses. Please note, this blog post is intended for general educational purposes only and does not serve as an offer, recommendation, or solicitation to buy or sell any securities. It may contain forward-looking statements, and actual outcomes can vary due to numerous factors. Past performance of any security does not guarantee future results. This blog is for informational purposes only. Neither the information contained herein, nor any opinion expressed, should be construed or deemed to be construed as solicitation or as offering advice for the purposes of the purchase or sale of any security, investment, or derivatives. The information and opinions contained in the report were considered by VF Securities, Inc.to be valid when published. Any person placing reliance on the blog does so entirely at his or her own risk, and does not accept any liability as a result. Securities markets may be subject to rapid and unexpected price movements, and past performance is not necessarily an indication of future performance. Investors must undertake independent analysis with their own legal, tax, and financial advisors and reach their own conclusions regarding investment in securities markets. Past performance is not a guarantee of future results.
Image Source – Gemini
