Micron’s climb into the $1 trillion club was not built on a blowout quarter or a clean earnings surprise. Seeking Alpha reported late last week that UBS tripled its price target, and Asia Business Daily reported that the stock jumped 19% in a single day after that call. That is a lot of market value to hang on one analyst’s math.

The math itself was aggressive. Asia Business Daily reported that UBS used a 15 times forward earnings multiple on projected 2029 EPS of $117, then raised Micron’s target to $1,625. That is the sort of framework usually reserved for companies the market thinks can compound cleanly for years, not for a memory maker that has spent most of its life getting knocked around by supply cycles.

The pushback is just as blunt. Seeking Alpha said this week that Micron’s fundamental value is closer to $300, with the stock near $970. The gap is wide enough that this is no longer a tidy valuation disagreement. It is a fight over what kind of business Micron is allowed to be.

The bullish argument starts with a real shift. Seeking Alpha said this week that Micron’s current cycle is being driven by AI-fueled demand, supply tightness and record margins. That is a better setup than the old version of the company, which lived and died by PC upgrades and the memory market’s mood swings.

UBS’s more daring claim is that long-term supply agreements are making that cycle look less violent. Asia Business Daily reported late last week that UBS pointed to LTAs as the main reason for its higher target, arguing that as much as 30% of industry-wide DRAM volume could soon sit under long-term contracts. The idea is simple enough. If customers are locking up supply for years instead of buying at the spot market’s mercy, earnings should become easier to model and less prone to whiplash.

There is something to that. A business with better visibility deserves a better multiple than one trapped in permanent panic about the next downcycle. That is the heart of the bull case, and it is not silly.

UBS also raised all of its future earnings estimates for Micron, Asia Business Daily reported, which gives the bullish case more substance than a simple “AI good, memory good” slogan. If the contracts really do blunt short-term volatility, and if AI demand stays tight enough for Micron to earn its way into those 2029 numbers, then the current valuation starts to look less like fantasy and more like a forward bet on a different kind of memory business.

Still, the bull case depends on things that are not settled. The research here does not show whether those LTAs fix prices, merely improve visibility, or come with exit clauses that soften their value in a downturn. That is not a small detail. It is the hinge the whole story swings on.

The skeptical view is less glamorous and easier to understand. Seeking Alpha said this week that memory remains a highly cyclical, capital-intensive oligopoly. That structure has not stopped the industry from cycling through shortages, price spikes and ugly reversals. It may be a better business than it used to be, but it has not become a different species.

A big reason bears are uneasy is that supply is coming. Seeking Alpha reported late last week that major DRAM producers are expanding capacity, with mega-fab spending across the industry expected to resolve shortages and drive down prices by 2028 to 2029. That is the old memory playbook: high prices attract capital, capital creates supply, and supply eventually takes the shine off the party.

Micron’s management guides for tightness through 2026, Seeking Alpha reported this week, but bearish analysts read the same setup differently. They expect margin normalization to begin in the second half of 2027. That distinction matters. One is a company view of the next 12 months. The other is a valuation warning about what happens after the current shortage has had its run.

There are signs the air is already coming out elsewhere. Seeking Alpha reported late last week that DRAM spot prices are already moderating and that non-AI demand from PCs and smartphones is still weak. The same piece said the non-AI side of Micron’s business, which has been helping earnings beat expectations, could hit a moderation wall in the second half of 2026 and show up in the company’s Q3 print.

That is the part the market may be glossing over. AI is not the whole business, and if the non-AI side slows while new supply is still ramping, the current margin story gets harder to defend. Seeking Alpha said Micron’s 5% free cash flow yield, even with support from one- and five-year contracts, reflects temporary pricing power rather than a permanent reset. A decent yield can still be a trap if investors use it to justify a multiple that assumes the cycle has been cured.

There is one escape hatch, and the bears know it. Seeking Alpha said a breakthrough in AI demand, something closer to AGI, could sustain shortages and high prices longer. But that is not the base case. It is the caveat that gets mentioned when the main argument starts to wobble.

Sandisk deserves a mention here, but not a fake precision that the evidence cannot support. The two names sit in related memory territory, yet Micron is primarily a DRAM and HBM story, while Sandisk is centered on NAND flash storage. That is not a footnote. It changes the supply story, the demand drivers and the margin structure.

The temptation is to mash them into one clean “AI memory bubble” trade. Convenient, yes. Accurate, no. Micron’s current rerating is tied to HBM shortages, long-term contracts and the market’s willingness to pay growth-stock multiples for a once-cyclical supplier. Sandisk’s valuation needs its own earnings framework and its own supply-demand context. The material here does not provide that.

So the honest answer is uneven. Micron can be analyzed on the evidence at hand. Sandisk cannot, at least not with the same confidence, and pretending otherwise would just be dressing up a narrative as research. Investors may want a tidy pair trade. The data does not oblige.

This is why the bubble debate is more interesting than a simple yes-or-no label. Micron is not a company with no earnings and no product. Seeking Alpha said this week that it is benefiting from AI-fueled demand, supply tightness and record margins. That is real business performance, not pure story time.

But the market price is not paying for what Micron is today. It is paying for what Micron might become if memory stopped behaving like memory. That is a bigger claim. It assumes LTAs will tame volatility, AI demand will stay stubbornly tight, hyperscaler spending will keep carrying the load, and the industry’s coming capacity expansion will somehow arrive without breaking the story. That is a lot of faith packed into one stock chart.

The most telling part may be how fast the market moved. Seeking Alpha reported late last week that Micron went from $500 billion to $1 trillion in 48 days, with little new fundamental information beyond the UBS call. That does not prove a bubble. It does suggest a market that is more than willing to run ahead of the evidence.

Micron is not a bubble in the cartoon sense. It is not a company with no earnings, no demand and no serious business case. The AI cycle is real, the HBM shortage is real, and the long-term contract story is at least plausible. That is why this debate is worth having.

The risk is subtler. At roughly $970 against a fundamental estimate near $300, Seeking Alpha said this week that Micron is being priced as if the memory downcycle has been demoted to a historical curiosity. That is the bet. Not on AI alone, which is easy enough to justify, but on permanence.

The next test is not far off. Seeking Alpha expects Micron’s non-AI business to show strain in the second half of 2026, while Seeking Alpha points to a correction likely beginning in the second half of 2027. In between sits the real question for investors: whether today’s valuation is a sensible read on a changing business, or a generous price for a cycle the market has decided not to believe in anymore.