Micron Investors Are Partying Like It's 1987, And Analysts Say the Stock Still Looks Pretty Cheap
It's December 1987. A gallon of gas costs $0.95. The Simpsons just made its debut. And Micron Technology's stock is soaring 78.8% in a single month, the kind of move that makes even seasoned investors do a double take.
Fast-forward to May 2026, and here we are again. Micron shares are up 73.7% this month alone. The company just crashed through the $1 trillion market-cap ceiling. And a UBS analyst dropped a price target so eye-popping, $1,625, up from $535, that the financial internet collectively asked: "Is this real?"
Here's the kicker: despite all of that, multiple analysts are using the word "cheap."
Yeah, I had the same reaction.
So let's unpack what's actually happening, the numbers, the technology, the risks, and figure out whether this party has more songs left on the playlist, or if someone's about to turn on the lights.
Wait, Why 1987? A Quick History Detour
The 1987 comparison isn't just a catchy headline. It's a genuine historical parallel.
Back in December 1987, DRAM prices were rebounding after a brutal downturn. The U.S.-Japan semiconductor trade agreement had just kicked in, and memory-chip makers, Micron among them, saw their stocks rip higher as demand returned faster than anyone expected.
Sound familiar?
The memory industry has always been cyclical, boom, bust, repeat. McKinsey once calculated that memory-chip manufacturers collectively destroyed $9.5 billion of investor wealth from 1997 to 2012 across multiple boom-bust cycles. It's a brutal business.
But here's where 2026 might actually be different from 1987, and we'll get to that in a minute.
The Numbers That Explain the Euphoria
Q2 FY2026: A Quarter for the Record Books
Let's start with the raw financials, because they're honestly hard to believe.
Micron reported fiscal Q2 2026 revenue of $23.9 billion , that's up 196% year-over-year from $8.1 billion in the same quarter last year. Analysts were expecting $20.1 billion. Micron blew past that by nearly $4 billion.
Non-GAAP earnings per share came in at $12.20, versus the consensus estimate of $9.31. Gross margins hit 74.4% , up from 36.8% a year ago. Net income? $13.8 billion. A year ago, that number was $1.6 billion.
For context: in 2023, Micron's gross margin sat below 20%. This turnaround is genuinely historic.
Oh, and free cash flow for the quarter reached $6.9 billion , a number that would have been unthinkable even two years ago, when Micron generated $15.5 billion in revenue for the entire year.
Guidance That Made Analysts Scramble
If Q2 was jaw-dropping, the Q3 guidance was the moment analysts started frantically revising their spreadsheets.
Micron guided for $33.5 billion in quarterly revenue , more than triple the year-ago period, with gross margins expected to hit 81% and EPS around $19.15.
81% gross margins. Let that sink in. Most software companies would envy those numbers. This is a hardware company that manufactures physical chips in multi-billion-dollar fabrication plants.
David Miller, CIO at Catalyst Funds, captured the mood well: even after this "large tear," the stock trades at under 10 times forward earnings. "It's interesting that even after such a strong move, it could still potentially be viewed as a value opportunity," he told MarketWatch.
"Stupid Cheap", The Valuation Case That's Hard to Ignore
Forward P/E: The Metric That Changes Everything
Here's where things get counterintuitive, and where the "still pretty cheap" part of our headline comes from.
If you pull up Micron's trailing P/E ratio right now, you'll see something around 42-43x. That looks expensive. Market bears point to it as evidence the stock has run too far, too fast.
But trailing P/E is a rearview mirror. It looks at what the company earned last year , back when AI memory demand was a fraction of what it is now.
The forward P/E , based on expected earnings over the next 12 months, tells a completely different story. That number has compressed to roughly 7x to 9x, depending on whose estimates you use.
To put that in perspective:
- The S&P 500 trades at roughly 22x forward earnings
- Nvidia trades at roughly 21-25x forward earnings
- The PHLX Semiconductor Index (SOX) trades at roughly 26.4x
Micron, a company at the literal center of the AI infrastructure buildout, whose HBM chips sit directly on top of Nvidia's most advanced GPUs, trades at less than one-third of the semiconductor index multiple.
D.A. Davidson analyst Gil Luria put it bluntly: memory is the "only sector of the AI trade" trading at a reasonable multiple right now.
Micron vs. Nvidia vs. the S&P 500, A Valuation
Here's the thing about value in plain sight: sometimes it's hiding in the one place nobody's looking. As one analyst put it, the only reason Micron trades at 7x forward earnings while sitting at the heart of an AI megatrend is that "your advisor still thinks 'memory' means 2018".
HBM: The Not-So-Secret Weapon Behind the Boom
What Is High-Bandwidth Memory, and Why Should You Care?
If you've scrolled past "HBM" in financial headlines and nodded along pretending to know what it means, this section is for you. No judgment.
High-Bandwidth Memory (HBM) is a specialized type of DRAM that stacks memory chips vertically, think of a skyscraper instead of a strip mall. This vertical stacking lets data travel shorter distances, which means much faster speeds and much lower power consumption.
Why does that matter? Because modern AI, the kind that powers ChatGPT, Claude, and every other large language model, is insanely memory-hungry. An AI server consumes roughly eight times more DRAM than a traditional server. And HBM specifically uses over three times more wafer capacity per bit than standard DDR5 memory.
Here's a simple analogy: if regular DRAM is a garden hose, HBM is a fire hydrant. AI models are the burning building. You can guess which one the fire department actually needs.
Only three companies on Earth can manufacture HBM at commercial scale: Micron, Samsung, and SK Hynix. That's the entire supplier list for a product that every hyperscaler on the planet is desperate to get their hands on.
Sold Out Through 2026, and Customers Are Begging for More
Micron's CEO Sanjay Mehrotra confirmed something remarkable: the company's entire HBM production for 2026 is already spoken for. Not "selling well." Not "strong demand." Sold out.
And here's the genuinely wild part: Micron estimates it can only fulfill somewhere between 50% and 67% of what its largest AI customers are asking for. Customers are effectively showing up with a shopping list and being told, "We can give you half of that. Maybe."
This is not a temporary blip. With no significant new HBM capacity expected before 2028, the supply-demand imbalance has years to run.
Mizuho analyst Vijay Rakesh, who ranks 5th out of more than 12,000 analysts tracked by TipRanks with a 92% success rate on MU stock, met with Micron's leadership and came away convinced: supply is unlikely to catch up anytime soon.
"This Time Is Different", The Structural Shift Argument
Long-Term Agreements: Goodbye Commodity Cycles?
Remember how I said this might actually be different from 1987? Here's why.
Historically, memory chips were commodities. Buyers would shop around, play suppliers off each other, and prices would swing wildly with supply and demand. When times were good, manufacturers built more factories. When those factories came online, supply flooded the market, prices crashed, and everyone lost money. Rinse, repeat. For decades.
But something structural has changed. The new long-term agreements (LTAs) being signed across the industry include fixed-volume commitments, extended durations (three to five years), and, crucially, partially fixed pricing.
UBS analyst Timothy Arcuri, who ranks No. 2 among more than 12,000 analysts, estimates that hyperscalers have already locked up roughly 60% to 70% of the industry's server DDR5 memory supply through these agreements.
What does that mean in plain English? Micron's revenue and earnings are becoming more predictable. The wild swings that defined memory investing for decades may be dampened. And if the market starts pricing Micron like a stable infrastructure play rather than a boom-bust commodity bet, the valuation multiple could expand dramatically, what analysts call a "re-rating".
Gil Luria adds another dimension: "de-commoditization." Some HBM products are now co-designed with Nvidia to fit specific GPU architectures, meaning they're "no longer interchangeable" between suppliers. When your product is custom-built for a specific customer's needs, you have pricing power. When it's interchangeable with competitors, you don't.
Only Three Companies on Earth Can Do This
The supply side is just as compelling. Building an HBM-capable fabrication plant takes years and billions of dollars. Micron is constructing a $100 billion campus in New York that broke ground in January 2026, wafer output isn't expected until the second half of 2028.
This isn't like software, where you can spin up more capacity with a few clicks. It's more like building a dam, years of planning, permitting, and construction before a single drop of water flows through.
The Bear Case: What Could Go Wrong
Alright, let's take off the rose-colored glasses for a moment. If everything were perfect, there'd be no article to write, the stock would already be fairly priced and nobody would be debating it.
Memory Is Still Cyclical, Anyone Who Says Otherwise Is Selling Something
Morningstar analyst William Kerwin offers the sober counterpoint: "We see AI driving a long and enduring upcycle, but a cycle nonetheless".
The memory industry has humbled investors who believed "this time is different" many, many times before. The McKinsey data, $9.5 billion in cumulative wealth destruction from 1997 to 2012, is a graveyard of optimistic forecasts.
Even UBS's own bull case comes with a sharp warning: Arcuri noted that if HBM demand weakens, Micron could fall to $250, representing roughly 66% downside from recent levels. That's a sobering range: $1,625 on the upside, $250 on the downside.
There's also the demand-destruction angle. Gartner and IDC forecast that worldwide PC and smartphone shipments could decline sharply in 2026, PCs by 10-11%, smartphones by 8-13%, partly because memory prices have surged 130%, making devices more expensive for consumers. High prices cure high prices, as the commodity traders say.
Insider Selling, Technical Stretch, and the Capex Question
Three additional yellow flags worth noting:
- Insider selling: Micron insiders sold approximately $54 million worth of shares over the last three months. Insider selling doesn't always signal trouble, executives diversify their holdings for all kinds of reasons, but it's worth acknowledging.
- Technical stretch: The stock hit levels 147% above its 200-day moving average earlier this year, exceeding even the extremes seen during the dot-com bubble and the Windows 95 era. Parabolic moves eventually correct, the question is when, not if.
- Capex risk: Micron boosted its fiscal 2026 capital spending plans to over $25 billion to fast-track new fabrication plants. That's necessary for long-term growth, but heavy spending can pressure free cash flow in the near term.
What the Analysts Are Actually Saying
Wall Street is overwhelmingly bullish, but the conviction level varies dramatically. Here's the landscape as of May 27, 2026:
Overall, Wall Street maintains a Strong Buy consensus on MU, with 27 Buys and 3 Holds over the past three months. The average analyst price target sits around $697.78, though that figure is being rapidly revised upward, and has already been surpassed by the stock's current trading price.
So... Is Micron Stock Still a Buy?
Here's the honest answer, and it depends entirely on who you are as an investor.
If you're a long-term believer in the AI infrastructure story... there's a genuine case that Micron is one of the most compelling values in the semiconductor sector. A forward P/E of 7-9x for a company whose products are literally sold out through 2026, with customers receiving only 50-67% of what they want, is unusual by any historical standard. The structural shift toward long-term agreements with fixed pricing could fundamentally change how this business is valued.
If you're a cautious investor who remembers 2018 (or 2000, or 2008)... the parabolic chart, the insider selling, and the immutable reality of memory cycles give ample reason to stay on the sidelines. "This time is different" are the four most expensive words in investing, and they've burned plenty of smart people before.
If you're somewhere in the middle... (which is most of us), consider position sizing. The bull case is compelling, but the downside risk is real. A small, carefully monitored position, one you'd be comfortable holding through a 30-40% drawdown, might be the sweet spot. Because if the structural thesis is right, the re-rating opportunity is substantial. And if it's wrong, you'll want to live to fight another day.
One thing is clear: Micron in 2026 is not your father's memory stock. Whether that justifies the current price, or leaves room for more, is the multi-trillion-dollar question.