Why Micron’s AI Windfall Isn’t Convincing Investors

Why Micron's AI Windfall Isn't Convincing Investors

Micron (NASDAQ:MU) just tripled its revenue over its most recent quarter, delivered some of the highest margins in memory history, and guided for continued strength, projecting $33.5 billion in revenue and gross margins expanding to 81% in the current quarter. Yet its stock remains down by 13% since the earnings release.

So why did the stock fall? It isn't even expensive, trading at roughly 7x forward earnings and closer to 4x FY '27 estimates.

Despite explosive growth and record margins, the market sold the news. This signals a shift in the AI trade. Investors are no longer rewarding peak results. They are questioning how sustainable they are. The focus has moved to what it takes to maintain leadership. Heavy capital spending and a re-accelerating Samsung are changing the narrative. For the first time this cycle, durability apparently matters as much as growth.

1. The Capex Conundrum

Micron has lifted its FY26 capex outlook to over $25 billion, up sharply from roughly $13.8 billion in FY25, as it ramps to meet AI-driven demand. High-bandwidth memory is already capital intensive, and the shift to HBM4 adds further complexity. HBM4 introduces a logic base die built on advanced nodes (5 nm and below), which means that Micron will likely need to rely on external foundries like TSMC.

Micron still handles the most complex step, stacking 12 to 16 memory layers on the base die, which requires new advanced packaging fabs being built in the U.S. At the same time, the company is expanding capacity with large-scale megafabs to meet demand that is already sold out through 2026. HBM4 might also force tighter integration with TSMC's logic die, driving upgrades across Micron's existing fabs to ensure precise alignment and yield.

The spending required to meet cutting-edge AI demand also increases future supply, which can put pressure on pricing over time. Investors are focusing on that trade-off. The spending is essential to stay competitive, but it weighs on near-term free cash flow and raises questions about how durable today's lofty margins will be.

2. The Samsung Threat

Another headwind to Micron's valuation is the resurgence of Samsung. While Micron has enjoyed an efficiency edge in the HBM space, Samsung could be executing a vertical integration play that Micron may not be able to replicate.

Samsung's 'all-in-one' model allows it to design the logic, fab the base die on its own 4 nm lines, and stack its own HBM - all in-house. This vertical integration gives Samsung a structural margin advantage and the ability to offer turnkey AI memory solutions for custom silicon designers. Furthermore, Samsung's foundry business is gaining critical momentum in the AI space, having recently secured the contract to manufacture the SRAM-based Groq 3 LPU for NVIDIA's new LPX inference racks. While the Groq 3 relies on SRAM rather than HBM, Samsung's ability to win such top-tier AI manufacturing contracts demonstrates its growing influence across the compute stack.

Even without full integration, sourcing both compute and memory from a single vendor simplifies procurement and optimization for customers. This creates a more turnkey solution for AI data centers and potentially allows Samsung to capture a larger share of the AI spend. While SK hynix remains the biggest HBM provider by far, Micron has had an edge over Samsung with its early HBM3E qualification and U.S.-based fabs, but these advancements could narrow that gap.

3. Cyclical Peak

The market's reaction suggests a fear that we are approaching "Peak HBM." While revenue growth over the last quarter came in at 196%, the law of large numbers and the entry of Samsung's massive capacity suggest that the rate of margin expansion may be approaching a ceiling. Guided gross margins of roughly 81% for the ongoing quarter reflect a severe supply shortage. Historically, such high margins attract rapid capacity expansion from all players. SK Hynix and Samsung are both adding production lines. The market likely expects supply to catch up with demand by 2027. This would lead to lower average selling prices. Micron's 13% stock drop suggests the market is pricing in this eventual normalization.

Demand for AI memory is coming largely from hyperscalers. However, the longer-term upside for NVIDIA comes from life sciences and drug discovery.

Could There Still Be an Upside?

At a mere 7x forward earnings, Micron is being priced like a peak-cycle commodity business. That's not irrational.

Semiconductor earnings can collapse quickly when supply catches up. But this cycle has some differences that matter.

HBM is structurally supply-constrained. It consumes far more wafer capacity, has complex packaging bottlenecks, and comes with long qualification cycles. This slows down supply responses and supports pricing durability. On top of that, multi-year contracts and prepayments from hyperscalers add visibility that memory has historically lacked. If margins don't collapse as sharply as in past cycles, even a modest re-rating could drive meaningful upside.

Micron may not escape cyclicality, but AI is changing the depth of the downcycle, and at 7x earnings, the market may be over-discounting that risk.

While Micron may have upside if the memory upcycle holds, its stock remains highly volatile, and earnings can swing sharply with changes in supply and pricing. When you want to survive market swings to protect wealth and grow your money in the long run, portfolios are the right choice. Trefis High Quality Portfolio can help you stay invested, capture upside, and mitigate the downside associated with any individual stock.