Micron Technology (NASDAQ: MU | MU Price Prediction) has become one of the most volatile stocks in the AI trade, and the debate over where shares go next is getting louder.
The stock has pulled back about 12% over the past month and sits well below its late-June peak above $1,200. Yet Micron is still up roughly 200% in 2026, and Wall Street’s biggest bulls have not backed away. Mizuho currently has a $1,375 target, Bank of America is at $1,550, and both DA Davidson and Susquehanna have gone as high as $2,000.
The reason for that optimism is straightforward: Micron is gaining market share at the same time AI infrastructure demand is keeping the memory market extraordinarily tight.
But investors should not ignore the other side of the argument. Memory remains one of the semiconductor industry’s most cyclical businesses. Today’s profit margins are exceptional by historical standards, and Michael Burry has taken a direct short position against Micron.
So is the recent pullback an opportunity, or are investors looking at peak earnings in one of the hottest stocks on the market?

Micron Is Gaining Ground in the Memory Market
The bull case starts with market share.
Recent DRAM market data cited by Barron’s showed Micron increasing its share to approximately 25% in the second quarter, up from 22% in the first quarter. That puts the company closer to SK Hynix while it continues competing with industry leader Samsung.
More importantly, Micron’s financial results show just how powerful the current memory cycle has become.
Fiscal third-quarter revenue reached a record $41.46 billion, up from $23.86 billion in the prior quarter. Non-GAAP earnings reached $25.11 per share, while operating cash flow came in at $25.39 billion.
And management expects the numbers to get even larger.
Micron’s fiscal fourth-quarter guidance calls for approximately $50 billion in revenue, an 86% gross margin, and non-GAAP earnings of roughly $31 per share. Those are extraordinary numbers for a company that has historically been known for boom-and-bust memory cycles.
The AI opportunity is also moving forward technologically. Micron says HBM4 is already in high-volume shipments for its lead customer’s platform, while HBM4E development is underway with volume production expected in calendar 2027.

Wall Street Still Sees Huge Upside
What makes Micron especially interesting is the disconnect between its earnings expectations and its valuation.
Mizuho analyst Vijay Rakesh currently argues that Micron trades at a forward P/E below 6x, while maintaining a $1,375 price target. He expects DRAM and NAND supply to remain tight through 2027, with more meaningful supply additions not arriving until 2028.
That distinction matters. The original version of this story described Micron as trading at roughly 5.6 times trailing earnings. That is not the right way to characterize the current valuation. The eye-catching single-digit multiple comes from Wall Street’s enormous forward earnings estimates.
At a recent share price around $853, Mizuho’s $1,375 target implies roughly 61% upside. Bank of America’s $1,550 target implies approximately 82%, while the $2,000 targets from DA Davidson and Susquehanna would represent well over 100% upside from current levels.
That helps explain why the bull camp remains enthusiastic even after Micron’s spectacular run.
But That Cheap P/E Comes With a Catch
The problem with valuing a cyclical company on peak forward earnings is that those earnings can disappear quickly if the cycle changes.
Micron’s projected 86% gross margin is remarkable. It also illustrates just how favorable current supply-and-demand conditions have become.
Memory shortages remain severe, particularly as AI infrastructure consumes increasing amounts of HBM and server memory. TrendForce says supplier inventories are extremely low and that AI-related demand continues to crowd out capacity for other parts of the market.
But investors buying Micron today are effectively betting that those favorable conditions last long enough to justify the company’s enormous earnings expectations.
Michael Burry is betting the other way.
The “Big Short” investor disclosed a direct short position in Micron around $1,051.87 per share in early July, arguing that investors were underestimating the company’s long history of severe cyclical drawdowns.
Micron also faces a growing competitive threat from China’s CXMT. Recent estimates put CXMT at roughly 7% of the DRAM market, although bullish analysts argue its capacity and access to the highest-end AI markets remain limited.

Could More Efficient AI Hurt Memory Demand?
There is another risk worth watching, but it needs to be framed carefully.
Google Research introduced TurboQuant earlier this year, a compression technology designed to significantly reduce the memory required for AI key-value caches. The technology demonstrates that software improvements can make AI inference considerably more memory-efficient.
That could eventually reduce the amount of memory required for a given AI workload.
What we cannot conclude is that these breakthroughs will necessarily reduce total demand for Micron’s products.
More efficient AI could simply make it cheaper to run more AI workloads, resulting in dramatically greater overall usage. That is essentially the Jevons Paradox argument frequently made by AI bulls.
There is already a good example of how counterintuitive this can become. Nvidia is reportedly considering configurations of Rubin Ultra with less memory per accelerator. Yet UBS analyst Timothy Arcuri argues the change could actually increase overall HBM consumption because more accelerators could ultimately be produced and deployed.
So software efficiency is a legitimate long-term risk to monitor, but it is not yet evidence that the memory boom is about to end.
What Micron Investors Should Watch
For investors, the Micron debate really comes down to one question: How long can the current memory shortage and pricing environment last?
If AI infrastructure spending continues to grow, memory remains constrained through 2027, and Micron continues taking market share, today’s forward valuation could prove surprisingly inexpensive. That is the scenario behind Wall Street’s $1,375, $1,550, and even $2,000 targets.
But the enormous earnings forecasts are also exactly why investors should be careful about treating a sub-6x forward P/E like an ordinary value-stock multiple. If memory pricing normalizes, new supply arrives faster than expected, or competition intensifies, those forward earnings estimates could come down quickly.
Micron’s fundamentals today are exceptionally strong. The harder question is how much of that strength is permanent and how much belongs to an extraordinary point in the memory cycle.
For now, the bulls have plenty of evidence on their side. But after a stock has risen roughly 200% in a year, investors should probably pay as much attention to the durability of those earnings as they do to how cheap the headline P/E looks.