Micron Stock: 315% Returns in One Year, a $25 Billion Capex Bet, and a Question Nobody Can Answer

Micron Technology’s stock ended Wednesday, April 1, 2026, up 8.9 percent after rising as high as 11.9 percent during the session. The stock increased by about $30 in a single trading day, reaching $367.85 as a result of the move. To put things in perspective, this stock was trading at about $90 a year ago. $61.54 is the 52-week low. Earlier in 2026, it hit an all-time high of $471.34. The real story of MU, a stock that has quickly become one of the most volatile and contentious semiconductor names in the market, is somewhere in between those data points. It has a 315 percent one-year return and is currently about 28 percent below its peak, all the while the underlying company is reporting numbers that most businesses would consider career-defining.

Memory chips are produced by Micron. In particular, it produces DRAM, high-bandwidth memory for AI applications, and NAND flash storage—the kinds of parts found in every server farm processing AI inference at scale, every data center running a large language model, and increasingly, the gadgets that people carry around in their pockets. Founded by a small group of engineers in a dentist’s office in Boise, Idaho in 1978, the company has withstood decades of harsh semiconductor cycles—the kind where memory prices plummet 60 percent in a year, margins evaporate, and half the industry is wiped out—to become one of just three major DRAM producers remaining in the world. Because the current moment is unlike anything Micron has ever experienced, it is important to consider the survivors’ past.

On nearly every metric, the Q2 fiscal 2026 results, which were released on March 18, were exceptional. Revenue increased 196 percent year over year to $23.86 billion. Earnings per share of $12.20 exceeded analyst estimates by $3.41, a margin of outperformance that indicates demand is running much hotter than even optimistic forecasters had predicted. This is not a small miss-to-beat gap. The business then projected Q3 revenue of about $33.5 billion, which would represent a growth of more than 260 percent year over year. At the halfway point, earnings per share were expected to be around $19.15, up from $1.91 in the same quarter the previous year. The gross margin is expected to increase from a 39 percent adjusted figure in the previous year period to about 81 percent in Q3. These figures depict a business functioning in an unprecedented demand environment, largely due to the global deployment of AI infrastructure.

Industry reports state that DRAM prices have increased by 90 to 95 percent, and memory suppliers are reportedly booked for the next two years. A typical chip cycle does not have that kind of supply constraint. It occurs when the industry’s capacity cannot keep up with a structural shift in demand, such as the need for enormous amounts of high-bandwidth memory in AI training clusters. As one of the three businesses with the technical capacity to manufacture the sophisticated memory products needed by AI processors, Micron, Samsung, and SK Hynix are in the middle of that limitation. Wall Street’s consensus is firmly in the “Strong Buy” camp, with 26 buy ratings, 2 holds, and an average price target of $533.40, which implies about 45% upside from current levels. This stance has produced the financial results visible in the Q2 report.

What caused the stock to drop 28% from its peak? Unfortunately, the solution is a concoction of elements that are genuinely ambiguous and somewhat unrelated to one another. The capital expenditure revision was the first and most important: Micron significantly exceeded the previous estimate of $20 billion by raising its fiscal 2026 capital expenditure guidance to over $25 billion. Even if the demand environment warrants the investment, such a spending commitment raises reasonable concerns about the margin trajectory over the medium term. The second was Alphabet’s announcement of TurboQuant, a compiling technology that supposedly lowers AI models’ memory needs without compromising their quality. When that news broke, Micron’s stock fell precipitously. The reasoning behind this was that if AI systems require less memory to function, demand for Micron’s products might decline. Several analysts have contended that the selloff was an overreaction, and the Wednesday rebound suggests the market is reevaluating how serious that threat actually is.

Investors are still figuring out the additional layer of complexity brought about by the insider activity. The Chief Business Officer and Executive Vice President of Micron, Sumit Sadana, sold 25,000 shares on April 2 for about $10.75 million. That appears to be a warning sign immediately following a significant stock increase. In actuality, Sadana still owns more than 248,000 shares valued at over $90 million, so the sale represents a very small portion of his overall holdings. A sale of this magnitude from someone with this much remaining exposure doesn’t tell you much about fundamentals, and executives diversify for a variety of reasons unrelated to their outlook on the company. However, during a time when sentiment can change rapidly, it provides the market with something to chew on.

According to Wolfe Research, Micron and Nvidia are the two semiconductor companies most accountable for the S&P 500’s earnings growth in 2026. The semiconductor group is expected to contribute roughly 39% of the index’s total year-over-year earnings gains. That’s a huge concentration of market performance in a very small number of businesses, which begs the question of what would happen if demand normalized, the memory cycle turned, or the buildout of AI infrastructure slowed. Micron has experienced similar shifts in the past (the late 1990s, the mid-2000s, 2015, and 2019), and each time it emerged with fewer competitors and a stronger position. Whether Micron is a successful company is no longer the question. It most likely is. The question is whether the market has appropriately priced in both the upside of that run and the eventual moderation that always follows, and what the stock is worth at a time when the memory cycle is running hotter than it has ever run.

On paper, Micron’s valuation case is exceptionally strong at about 6.2 times forward earnings, with consensus estimating EPS growth exceeding 651 percent in fiscal 2026 and another 67 percent in fiscal 2027. The recent volatility—a 28 percent drop from an all-time high, a 9 percent one-day recovery, and an insider sale the next morning—makes it evident that the market hasn’t fully settled on a narrative, even though the stock appears cheap by all traditional metrics. Watching Micron’s stock fluctuate gives investors the impression that they are torn between two beliefs: either the AI memory demand cycle is genuine and long-lasting, or the stock has already moved so much that typical uncertainty feels heightened. Neither emotion is incorrect. They will both be around for some time.