A surge in demand for artificial intelligence hardware has created a valuation paradox in the memory chip market, where soaring stock prices for companies like Micron Technology Inc. are making them cheaper for investors. The rally is fueled by an AI infrastructure buildout that has analysts raising earnings forecasts far faster than share prices can keep up, though the industry’s history of boom-and-bust cycles casts a shadow over the rally.
“All high-bandwidth memory can be priced at a ridiculously high price because of supply and demand,” said Kim Forrest, chief investment officer at Bokeh Capital Partners, which holds Micron stock. “As long as that’s the case, I’m in a state of nirvana. It’s a beautiful thing. It’s not that expensive because people are buying it and it’s sold out.”
Micron’s stock has jumped 172% in 2026, making it a top performer in the S&P 500. Yet, its price-to-estimated earnings ratio has fallen to less than nine, down from about 12 in February. This is a steep discount to the S&P 500’s forward earnings multiple of about 21. The dynamic is explained by a massive 768% increase in analysts’ 2027 adjusted earnings per share forecast for Micron over the past year, according to data compiled by Bloomberg. A similar trend was seen in Sandisk Corp., which led the S&P 500 with a 482% gain while its valuation compressed.
The core of the bull case rests on the belief that the AI infrastructure buildout is still in its early stages. Companies are spending billions on data centers and AI hardware, which require vast amounts of high-bandwidth memory (HBM) to function. “Unless there’s a major change in capital spending from the hyperscalers — and right now we’re seeing it go up, not down — the demand for storage and memory is going to continue to grow,” said Rob Thummel, a senior portfolio manager at Tortoise Capital, whose Tortoise AI Infrastructure ETF holds both Micron and Sandisk.
The Nvidia Effect and Broader Market
The market’s focus on AI has been solidified by moves from industry leader Nvidia Corp., which acquired inference-chip maker Groq for $20 billion and has reserved the majority of Taiwan Semiconductor Manufacturing Co.’s leading-edge chip production capacity through 2027. This has created a rising tide for the entire semiconductor ecosystem, from custom chip designers like Broadcom Inc. to connectivity providers like Astera Labs. The newly public Cerebras Systems, a specialist in wafer-scale inference accelerators, priced its IPO at a $56.4 billion valuation, further showing the market's hunger for AI hardware.
However, not all storage-related stocks are experiencing the same valuation compression. Hard drive makers like Seagate Technology and Western Digital Corp. have seen their forward price-to-earnings ratios expand to over 30, up from around 20 at the end of March. This reflects a different business cycle and valuation logic compared to the more volatile memory chip sector.
Cyclical Risks Remain
For investors, the low valuations on memory stocks may be a warning sign that the market is already pricing in peak earnings growth. The semiconductor industry is notoriously cyclical: shortages lead to high prices and expansion, which eventually results in oversupply and a crash in prices and profits. Micron itself saw its stock fall 46% in 2022 as earnings expectations collapsed, only to rebound in the current AI-driven cycle.
“You can’t look at these stocks the way you would a company with steady earnings growth,” said Randy Hare, director of equity strategy at Huntington National Bank, which holds Micron. He believes the “easy money” in the stock has been made and expects more volatility ahead. Jed Ellerbroek, a portfolio manager at Argent Capital Management, avoids memory stocks due to these cyclical concerns. “A shortage eventually begets a glut,” he said, though he concedes it may take years for supply to catch up with the current wave of demand.
This article is for informational purposes only and does not constitute investment advice.