By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
The panelists agree that Micron's high gross margins are driven by HBM, but disagree on whether this is cyclical or structural. They also highlight the risk of DRAM cyclicality and the potential impact of a shift in AI workloads from training to inference.
Risk: DRAM cyclicality and a potential shift in AI workloads
Opportunity: Sustained demand for HBM due to AI growth
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
Micron Technology(NASDAQ: MU) is having a moment. The maker of memory chips just posted second-quarter fiscal 2026 revenue of $23.9 billion, up 196% year over year. Gross margins have scaled up past 70%, a figure that would have seemed absurd for a memory company just three years ago. The stock trades at a forward P/E below 10, while the semiconductor industry median is around 30.
If you squint, it looks a lot like Nvidia(NASDAQ: NVDA) before everyone figured out that the chip designer would dominate the artificial intelligence (AI) hardware bonanza. So naturally, the question is circulating: Is Micron the next Nvidia? I've spent some time with this question.
Will AI create the world's first trillionaire? Our team just released a report on a little-known company, called an "Indispensable Monopoly," providing the critical technology Nvidia and Intel both need.
But the reasoning should tell you something useful about both companies.
Why the Nvidia comparison is tempting
The Micron-as-Nvidia argument makes sense at first glance.
Micron makes high-bandwidth memory, and AI accelerators need high-bandwidth memory the way cars need wheels. There are only three companies on the planet that can make it at scale: Micron, SK Hynix, and Samsung(OTC: SSNLF). The company has confirmed its high-bandwidth memory capacity is 100% sold out through calendar year 2026, with shortages expected next year as well.
Return on equity is 40%. The addressable market could hit $100 billion by 2028. If you're looking for pick-and-shovel AI plays, Micron is a pretty compelling shovel. Just like Nvidia, right?
Tollbooth vs. destination
Well, not exactly. Micron may be a great AI investment, but it's fundamentally different from Nvidia.
Nvidia designs the AI accelerators, controls the CUDA software ecosystem, and defines the architectural standards that the rest of the industry builds around. Developers don't just tolerate Nvidia's platform; they actively seek it out.
Micron's memory is essential, but it's essential in the way that a highway is essential to a theme park. Customers pass through because they have to, not because they want to. It's fair to assume that Micron's customers would jump to Samsung or SK Hynix in a heartbeat if the other supplier offered lower prices (or any other advantage).
This distinction matters for long-term pricing power. Platform owners like Nvidia can expand their moats over time through ecosystem lock-in. Component suppliers, even critical ones such as Micron, remain vulnerable to shifts in the supply-and-demand balance.
Both can be profitable. Only one company has built a platform monopoly, and it's not Micron.
What the market is actually saying
Micron stock trades at a forward price-to-earnings ratio of 7.4 in a sector where 30x multiples are the norm. That looks like a screaming buy. Or perhaps the market knows something you don't, applying a risk discount whether you like it or not.
Memory companies have been through this before. Margins spike, everyone declares a new paradigm, the majors build more manufacturing facilities, and then the whole thing collapses into a price war.
This cyclical sector is in an upswing right now. Micron's 74% gross margin is about double what it was three or four years ago. That's probably not sustainable in the long run, unless the AI boom continues forever. Even if things really are "different this time," the red-hot demand should eventually balance out because of rising supply-side capacity.
Micron is a different kind of winner
So Micron just isn't the next Nvidia.
It is a supplier to the AI ecosystem, not its architect. Its role further down the supply ladder caps Micron's long-term pricing power and justifies at least some of the valuation discount the market currently applies.
None of this makes Micron a bad investment. The balance sheet is pristine, with a debt-to-equity ratio of 0.14 and $14.6 billion in cash. Capacity is sold out for months, maybe even years.
But executing well as a supplier is different from owning a platform, even in the massive AI surge. Investors should evaluate Micron on its own merits rather than as a proxy for Nvidia's trajectory.
Tolls can be lucrative. Just don't mistake Micron's indispensable toll road for Nvidia's desirable destination.
Should you buy stock in Micron Technology right now?
Before you buy stock in Micron Technology, consider this:
The Motley Fool Stock Advisor analyst team just identified what they believe are the 10 best stocks for investors to buy now… and Micron Technology wasn’t one of them. The 10 stocks that made the cut could produce monster returns in the coming years.
Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you’d have $469,293!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,381,332!*
Now, it’s worth noting Stock Advisor’s total average return is 993% — a market-crushing outperformance compared to 207% for the S&P 500. Don't miss the latest top 10 list, available with Stock Advisor, and join an investing community built by individual investors for individual investors.
Anders Bylund has positions in Micron Technology and Nvidia. The Motley Fool has positions in and recommends Micron Technology and Nvidia. The Motley Fool has a disclosure policy.
Four leading AI models discuss this article
"Micron's integration into the AI hardware stack via HBM creates a technical barrier to entry that justifies a higher valuation multiple than traditional cyclical DRAM suppliers."
The article correctly highlights the 'tollbooth vs. platform' distinction, but it underestimates the structural shift in High Bandwidth Memory (HBM). Unlike legacy DRAM, HBM is custom-integrated into the AI chip package, creating a technical 'stickiness' that mimics a moat. With a forward P/E of 7.4x, the market is pricing Micron for a cyclical cliff that may not arrive as quickly as feared. If Micron maintains its HBM3E market share, the current 70% gross margin isn't just a cyclical peak; it's a new floor for a high-value, supply-constrained specialized component. The valuation gap relative to the broader semi-cap index is too wide to ignore.
The bear case is that HBM capacity will eventually face a supply glut once Samsung and SK Hynix finish their massive capex cycles, leading to a brutal commodity-style price war that crushes Micron’s current margin expansion.
"Micron is a cyclical compounder, not a platform, but the current cycle's length and depth may justify 7.4x multiples if capacity constraints persist through 2026-2027."
The article's tollbooth-vs-destination framework is intellectually honest but incomplete. Yes, Micron lacks Nvidia's moat—but that doesn't mean it's a bad investment at 7.4x forward P/E. The real question is whether 74% gross margins are cyclical or structural. If AI demand genuinely sustains (not forever, just 5-7 years), Micron's sold-out capacity through 2026 plus $14.6B cash and 0.14x debt-to-equity creates a cash machine. The article correctly warns against confusing Micron with Nvidia, but conflates 'not a platform' with 'not worth owning.' Memory suppliers have been cyclical—but this cycle may be longer and deeper than historical precedent.
If Samsung and SK Hynix ramp capacity aggressively in 2025-2026, Micron's pricing power collapses faster than the article suggests, and a 7.4x multiple suddenly looks expensive for a cyclical compounder facing margin compression.
"Micron could benefit from AI-driven memory demand, but without a platform moat similar to Nvidia, any multiple expansion hinges on sustained demand and constrained supply rather than a durable pricing advantage."
Micron trades at a sub-10x forward P/E while AI demand could lift HBM sales and margins, but the stock is not Nvidia: Micron is a supplier, not a platform architect, so structural pricing power is limited and vulnerable to memory cycles. The bull case hinges on 1) sustained AI capex and 2) entrenched HBM shortages, which could push margins higher and extend the sell-out. Yet supply expansion by Samsung and SK Hynix could pressure pricing, and a cyclicality-driven downturn could crush multiples quickly. Investors should stress-test the moat, not just the ballast of cash and growth.
The strongest counter is that memory pricing is highly cyclical and capacity can ramp at scale, so pricing power may erode faster than the market expects—unlike Nvidia's platform moat, Micron has no enduring network effects to lean on.
"Micron's AI-driven margins face structural limits from its component-supplier position that Nvidia's software moat avoids."
The article correctly flags Micron's supplier status versus Nvidia's platform control, but underplays how concentrated HBM supply among three players could extend elevated margins beyond typical cycles. With capacity sold out through 2026 and ROE at 40%, MU's forward P/E of 7.4 embeds excessive cyclical fear. Geopolitical risks around Taiwan and Korea fabs, plus potential AI capex slowdowns at hyperscalers, remain under-discussed. Execution on new nodes will determine if this upcycle differs from prior memory booms.
Samsung and SK Hynix are already expanding HBM lines aggressively; any demand pause could trigger the same price collapse seen in 2018-2019 DRAM cycles, erasing Micron's margin gains within 18 months.
"Micron's HBM success cannot fully insulate the company from the cyclical volatility of its core legacy DRAM business."
Claude and Gemini are over-indexing on HBM 'stickiness' while ignoring the commodity reality of the broader DRAM portfolio. Even if HBM3E remains a premium niche, it still represents a minority of Micron's total bit supply. If hyperscalers pivot from training to inference, the demand profile for HBM shifts toward lower-power, higher-density requirements that may favor different architectural trade-offs. Micron isn't just fighting Samsung; they are fighting the inherent volatility of a portfolio still tethered to legacy PC/mobile cycles.
"HBM margin sustainability is irrelevant if broader DRAM demand collapses or shifts to lower-bandwidth inference workloads, eroding Micron's total utilization and pricing power."
Gemini's portfolio pivot is the real tell. HBM3E margins mean nothing if DRAM—still 60%+ of Micron's revenue—faces cyclical collapse. Training-to-inference shift is real, but it's a demand *composition* problem, not a HBM stickiness problem. If hyperscalers need less total memory bandwidth per workload, Micron's utilization drops even if HBM pricing holds. The article missed this entirely: Micron's moat isn't HBM; it's capacity. Once Samsung/SK Hynix finish capex, capacity becomes abundant and the portfolio reverts to commodity pricing.
"HBM stickiness is not a floor; DRAM cyclicality and capacity waves can erode margins."
Gemini overplays HBM3E stickiness as a floor. Even if HBM3E commands premiums, MU remains DRAM-heavy (60%+ of revenue) and highly cyclical. A Samsung/SK Hynix capacity wave could flood pricing, dragging overall gross margins toward historical DRAM levels. The real risk is demand normalization, mix-shift away from memory-bandwidth-heavy AI workloads, and hyperscaler capex slowdowns that temper the upcycle. That argues for a wider distribution of outcomes and a tighter stop on MU's multiple.
"HBM manufacturing complexity delays supply glut beyond what capex announcements suggest."
Claude's view that capex completion makes capacity abundant ignores HBM's manufacturing hurdles. Stacking and TSV processes create yield bottlenecks that slow effective supply growth beyond announced investments. This technical friction, combined with Gemini's noted DRAM cyclicality, suggests margins could hold longer than a simple commodity reversion implies, particularly if AI training demand persists through 2026.
The panelists agree that Micron's high gross margins are driven by HBM, but disagree on whether this is cyclical or structural. They also highlight the risk of DRAM cyclicality and the potential impact of a shift in AI workloads from training to inference.
Sustained demand for HBM due to AI growth
DRAM cyclicality and a potential shift in AI workloads