From $103 to $365: Why Micron’s AI-Fueled Rally Still Has Room to Run

Mid-January 2025, we tagged Micron Technology (NASDAQ: MU) as one of the best AI stocks to hold in 2025. At the time, MU stock was priced at $103.21 per share against its current price of $365.00, representing an impressive 251% profit-taking opportunity.

Despite a trailing price-to-earnings (P/E) ratio of 34.48, Micron trades at a forward P/E of just 11.45. This is well below the semiconductor industry’s average forward P/E of 37.29.

At its current price level of $362.75, MU stock exceeds its average price target of $350.46, according to aggregated forecasts of 45 analysts collected by the Wall Street Journal. On the low end outlook of $107, MU shares could end up just slightly higher from a year-ago price, while the ceiling price target is $500.Let’s examine if the latter scenario – getting closer to $500 or even higher – is more likely.

The Role of Memory in the AI-Semiconductor Ecosystem

The underlying assumption of Micron’s continued rise, as Nvidia’s key memory supplier, relies on the AI surge transforming into a durable economic layer instead of popping up as an overvalued bubble. Over the last two years, we have explained multiple times that the bubble scenario is less likely, with some caveats.

With that in mind, let’s recap what is so special about Micron’s memory. It is well known that all AI models and transformers need both powerful CPUs – for overall data handling – and GPUs for heavy data crunching. Yet, this compute power is rendered idle, resulting in stalled throughput, if there is insufficient near memory for large blocks of data to be read and written back.

AI models are especially memory hungry as they are built from matrix multiplications, vector operations and tensor contractions. And at every compute step, AI accelerators – GPUs, TPUs and NPUs – must be fed data to execute tens of thousands of operations simultaneously.

Before the AI era, it was common to say that a process is compute-bound (CPU), but now it is the case that AI chips are memory-bound.

Join our Telegram group and never miss a breaking digital asset story.

Micron’s Position in the AI-Semiconductor Ecosystem

Unlike Nvidia and AMD, which rely on TSMC to fabricate their chip designs, Micron has its own fabrication facilities as an Integrated Device Manufacturer (IDM). In the classical arena of cloud computing – server clusters – the company supplies DDR5 DRAM as the main workhorse for CPUs and GPUs.

For AI accelerators, Micron supplies high-bandwidth memory (HBM), specifically HBM3, the latest HBM4 and customized HBM4E, which has around 2.5x higher bandwidth. In December’s Counterpoint report for Q3 2025, Micron ranked third in the global DRAM/HBM market share, at 26%.

As evidenced by AMD’s presentation at CES 2026, missing the AI train is not an option. Accordingly, the focus is now, and for the foreseeable future on HBM. The problem is, HBM modules are more complex to fabricate, resulting in the need for more silicon (wafer) to produce chips than DRAM.

Consequently, memory companies have to reallocate resources from DRAM to HBM. This is causing enormous DRAM price spikes since October, as evidenced by PC Part Picker price charts. Most recently, Micron’s Executive Vice President of Operations Manish Bhatia noted that “the shortage we are seeing is really unprecedented,” per Bloomberg.

Although the obvious solution would be to expand production capacity, setting up new memory fabs – to go from tool installation to wafer yield ramping – is an extremely complicated and time-consuming process. Additionally, HBM memory chips have lower yields than standard DRAM.

On top of that, memory companies like Micron are incentivized to reap scarcity-induced profits. In numbers, per Microns’ latest earnings, that translates to $10.8 billion revenue from DRAM, which is up 69% year-over-year.

The Bottom Line

In the latest fiscal Q1 2026 earnings report, Micron delivered gross margin and earnings-per-share well above its top guidance. Expectedly, Micron’s Cloud Memory (CMBU) division was the culprit. While CMBU’s gross margin increased to 66% from 51% in the year-ago quarter, the revenue increased nearly 100% to $5.28 billion.

In other words, the AI boom is an enormous boon for Micron and is likely to continue to be so. Additionally, Micron’s Automotive and Embedded (AEBU) division achieved record $1.7 billion revenue, now representing 13% of the company’s total revenue of $13.6 billion (79% from DRAM).

More than DRAM scarcity and HBM demand itself, it is important to note that the robotaxi economy is just starting to launch, as we explored in Nvidia’s grip on autonomous vehicles stack. Micron will be an indispensable cog in this evolution.

Seeing this future, Micron signed a $1.8 billion letter of intent on Saturday to purchase Powerchip Semiconductor Manufacturing Corporation (PSMC)’s P5 fab site in Taiwan. As we mentioned previously, not only is setting up new fabs time-consuming, but integration as well. In turn, extra DRAM output from PSMC is not expected until H2 2027.

Notwithstanding any bubble popping, this is broadly the timeline within which Micron will continue to reap scarcity-driven profits. In conclusion, the scenario for a $500 MU stock price milestone is likelier than ever by the end of 2026.

Disclaimer: The author does not hold or have a position in any securities discussed in the article. All stock prices were quoted at the time of writing.