Micron Is Up 187% and Still Trades at 6x Earnings. Is The Market Wrong?

Micron (MU) stock has been on a tear so far this year, up 187% despite recent jitters. With AI infrastructure spending still on an upward trend, the memory chip shortage continues to provide a great investing opportunity. Yet things aren’t as simple as the relationship between demand and supply. Investors today continue to fear the cyclical nature of the industry, creating a mismatch between the stock price direction, analyst sentiment, and memory demand. To alleviate these fears and understand why the earnings multiple presents an opportunity rather than a warning, it is important to understand why this shortage exists in the first place.

Industry experts measure demand in bits rather than the number of chips. Bits show how many gigabytes or terabytes of data customers want across all chip types. The distinction is important because the current memory cycle is unlike those of the past. The production mix has shifted to high-end chips, precisely the high-bandwidth memory (HBM) chips for AI accelerators. Micron and its peers have been unable to convert production lines to this type because the demand simply never existed at this scale before. Now they’re trying to catch up.

HBM Creates A New Bottleneck

What makes matters worse is that HBM takes up a lot of wafer capacity. According to research firm TrendForce, high-bandwidth memory is set to account for just over 10% of the total DRAM bit supply between 2025 and 2027. Yet it is expected to consume over 23% of the total DRAM wafer input. This tightens the market across other types of memory chips used in everyday laptops and smartphones. As a result, memory makers are forced to cut down on other types of chips to service this increasing wafer demand, resulting in shortages for chips that aren’t needed for AI workloads.

Micron’s recent capital allocation trend also supports the wafer shortage thesis. The company has entered into a 10-year supply agreement with GlobalWafers, a Taiwanese company that is the third-largest silicon wafer manufacturer in the world. The agreement also includes $500 million in strategic financing support, which is another way for the company to say we will fund you, just ensure that we have reliable access to silicon wafers for our memory chip production. 

When Will The Supply Shortage End?

Micron’s management has put it as bluntly as they could: “We don’t really see when the supply is going to be able to meet demand”. These words, coming from the Chief Business Officer Sumit Sadana, show that it is not just the demand, but also the inability of memory makers to convert existing production lines in time to service this massive AI demand.

On the earnings call on June 24, the company also said it was increasing its Capex for fiscal year 2026 to $27 billion. For the next year, it expects Capex in the mid-40s, but clarified that it does not foresee any scenario where the Capex could go past $50 billion. This suggests there is some discipline in spending, and even though the company doesn’t explicitly state so, it is indeed wary of the eventual scenario of supply catching up to demand. 

Wall Street Analysts Reinforce The Supply Crunch

Analysts are often able to talk more openly about supply issues than the company management, and by tracking the Strategic Customer Agreements (SCAs), they do not foresee any let-up in demand until at least the end of 2028. 

Deutsche Bank’s Melissa Weathers is one such analyst who believes memory supply tightness is likely to persist well into 2028. Other analysts have warned that things could ease out by 2029 and 2030, but confirm that the shortage will persist.

Wedbush analyst Matt Bryson not only sees the GlobalWafers investment as a confirmation of the memory supply crunch persisting but was quick to point out that Silicon wafers themselves could become a bottleneck due to the increasing memory demand:

“We see MU's investment in GlobalWafers as potentially signaling MU views wafer supply as another possible hardware bottleneck, a result that makes sense given the likely sharp increase in wafer requirements as memory and logic investments scale into 2028, 2029 and 2030” 

Wall Street commentary confirms the memory shortage thesis, and the average analyst price target of $1538 points to an 80% upside.

What Explains The Low Valuation?

Those who have invested in memory stocks in the past would know how brutal the downturns can be. The industry is extremely cyclical, and that is one reason why companies didn’t invest in major additional capacity during previous downturns. The big three of memory chips manufacturing, Samsung, SK Hynix, and Micron, together control over 90% of the DRAM market. No new player is likely to enter this business. In fact, Intel shut down its DRAM business over 40 years ago after inventing the technology, simply because the margin destruction due to fluctuating demand was brutal. In 2020, it also sold its NAND business to SK Hynix. A competitor is unlikely to emerge at such short notice, except, of course, one threat from China that is weighing down on investor sentiment.

CXMT is China’s largest DRAM maker. It is in the news because Apple is seeking permission to buy DRAM chips from the company. However, CXMT hardly has any HBM production at scale and therefore cannot help alleviate the HBM-induced supply crunch. It can only cover the supply shortage outside AI, which, at best, would hurt the margins of the major memory makers. Even that is unlikely considering the artificial intelligence race that the US and China find themselves in.

Investors are also unsettled by the additional capacity expected to come online in 2027 from all three major manufacturers. However, analyst reports have already confirmed that this won’t solve the supply crunch, only make the problem smaller. The companies should therefore continue to enjoy higher gross margins. Also, the current valuation already prices in this capacity risk.

The stock continues to trade at a 2027 P/E of 6x and a 2028 P/E of just 5.5x, partly due to these risks. The nature of HBM, the continuation of the AI infrastructure buildout, and the emerging silicon wafers bottleneck all point to one thing: the market is pricing Micron wrong and treating this as just another memory cycle when things have shifted structurally due to AI. The low P/E is an opportunity rather than a warning sign.

Want to explore opportunities similar to Amazon? Quantli’s Strategy Lab lets you build and backtest custom strategies using historical market data, helping you see how an approach would have performed across different market conditions.