Micron's AI Hype: A Data Analyst's Reality Check
The HBM Mirage
Micron (MU) has been on a wild ride, fueled by the AI narrative. A 169% jump this year is eye-catching, no doubt, though a recent 10% pullback from its all-time high should give investors pause. The question isn't whether Micron is benefiting from the AI boom – it clearly is. The question is whether the market's expectations are outpacing reality.
The latest wobble seems to stem from CFO Mark Murphy's comments about potentially increased capital expenditures, or CapEx, at a recent Global Technology Conference. He suggested that the current $18 billion annual spending pace is facing "pressure" and will likely need to increase beyond 2026. Wall Street hates uncertainty and future spending commitments, hence the sell-off.
But let's dig into what's actually driving this potential CapEx increase. The narrative is all about High Bandwidth Memory (HBM), the specialized memory chips crucial for AI accelerators like those from Nvidia (NVDA) and AMD (AMD). UBS analyst Timothy Arcuri, ranked 7th among stock experts (according to TipRanks), is particularly bullish, arguing that HBM will "crowd out" the traditional memory market. He believes nearly all industry capacity additions through 2027 will be directed toward HBM.
That's a bold claim.
Arcuri also points out that Micron expects supply tightness to last through the end of 2026, longer than previously projected. Their HBM supply is supposedly fully booked through 2026. Furthermore, DDR5 contract negotiations are reportedly approaching +20% quarter-over-quarter increases, with some mobile DRAM deals seeing ASPs up nearly 40% sequentially. According to one report, UBS is very positive on Micron stock.
Here's where the data needs a closer look. While the ASP increases are real, the "crowding out" thesis hinges on a few key assumptions. First, that HBM demand will continue its exponential growth. Second, that Micron can actually execute on its HBM production plans. And third, that traditional DRAM demand won't see a resurgence of its own.
Cracks in the Foundation
Let's consider the capacity limitations Arcuri mentions. He acknowledges that Micron might face a "slightly shallower ramp in HBM revenue" compared to previous quarters due to these limitations. But what if those limitations are more significant than he suggests? What if Samsung and SK Hynix are better positioned to capitalize on the HBM boom?
And this is the part of the report that I find genuinely puzzling. The market seems to be extrapolating current trends indefinitely. But the memory market is notoriously cyclical. As the Seeking Alpha analyst points out, it's characterized by regular boom and bust cycles. To assume that HBM demand will only go up, and that Micron will flawlessly execute its HBM strategy, strikes me as overly optimistic.
The "AI-powered stock more than quadrupled in three years" narrative, while technically true, also needs context. Micron's stock has indeed surged 319% over the past three years (as of November 18, 2025). But as that same article notes, a similar gain over the next three years is a "stretch." The author rightly points out that the stock is valued at 7.0 times trailing sales today, miles ahead of the 3.9 average over the past five years. One article questions if this AI stock can surge 319% in 3 years.
Another point: While everyone is focused on HBM, let's not forget the bread and butter: traditional DRAM. Arcuri believes "profitability in core DRAM strengthening even further – leading to DDR gross margin to surpass HBM for the first time in early C2026." If that's true, then the entire HBM-centric narrative starts to crumble. Why invest so heavily in CapEx for HBM if traditional DRAM is going to be more profitable? (A question Micron's management needs to address more directly.)
I've looked at hundreds of these filings, and the lack of clarity around the specific allocation of CapEx between HBM and traditional DRAM is troubling.
Furthermore, the current "chip shortage in the memory market," as one article puts it, is likely a temporary phenomenon. Samsung is already raising its chip prices by 60% in two months, a move that Micron and SK Hynix will likely follow. But these price increases will eventually incentivize increased production, leading to oversupply and, inevitably, a price correction.
The Hype Has Outrun the Fundamentals
Micron is not a bad company. It's one of the top three DRAM manufacturers in the world. But the current valuation is pricing in near-perfect execution and sustained, exponential growth in HBM demand. My analysis suggests that the market is underestimating the cyclical nature of the memory market, the competitive pressures from Samsung and SK Hynix, and the potential for traditional DRAM to remain a significant profit driver. The average MU price target of $227.14 implies 0.55% downside from current levels – a discrepancy between analyst sentiment and cold, hard numbers.