Micron Technology shares closed at $958.73 on Monday, Aug. 31, down 21% from their June all-time high, even as the stock remains up roughly 680% over the past year. The pullback comes amid signs that customers are working to manage rising AI infrastructure and memory costs—while Micron’s latest results show the company is still benefiting from a tight high-bandwidth memory market.
Investors are weighing a near-term earnings boom—supported by record revenue and sharply higher profits—against the risk that the current supply-demand imbalance could ease within the next year or two, potentially weakening Micron’s pricing power.
Key takeaways
- Micron stock: Shares closed at $958.73 on Aug. 31, 21% below the June peak, despite a ~680% one-year gain.
- Catalyst: Cost pressure across AI data centers, including reports of higher chip pricing tied in part to HBM costs, has raised questions about the durability of demand.
- Earnings momentum: Micron reported record fiscal 2026 third-quarter revenue of $41.4 billion and a 1,368% jump in quarterly earnings per share.
- Key implication: The market is pricing in a potential normalization of memory pricing, which could compress margins if supply ramps.
What drove Micron’s pullback
According to Nvidia, the largest hyperscale buyers—including Microsoft and Amazon—plan to spend nearly $800 billion combined on AI data center infrastructure in 2026, with spending potentially rising to more than $1.3 trillion in 2027. While that capex outlook supports long-term AI buildouts, the economics of those investments are becoming more difficult as hardware costs climb.
A Bloomberg report cited by the article said Nvidia told customers to expect a 15% price increase, with part of the move attributed to rising high-bandwidth memory prices. As a result, some AI buyers may be reconsidering how much they can spend on full-rate training and inference capacity.
In parallel, the article points to actions taken by AI platform and model providers to adjust usage economics. Microsoft and Anthropic implemented passive price increases for the use of their AI models and software, and the article also notes that Uber Technologies, Walmart, AT&T, and Amazon have adopted AI usage caps for employees to limit budget blowouts. Separately, an UBS Group survey cited in the piece said about 60% of businesses are routing tasks to cheaper, more efficient AI models to reduce usage costs—an approach that can reduce the demand intensity for certain types of compute and, by extension, may influence semiconductor demand patterns.
Finally, the article highlights policy uncertainty. Lawmakers in more than a dozen U.S. states have introduced legislation aimed at temporarily banning new data center construction while they evaluate social, economic, and environmental impacts. That could delay some planned infrastructure and create downstream implications for the supply chain serving AI workloads.
Micron’s financial performance remains strong
Despite the above-headwind narrative, Micron’s recent results show demand has not slowed materially for the company. The article says Micron generated record total revenue of $41.4 billion in its fiscal 2026 third quarter ended May 28, representing a 346% increase from the year-ago period.
It also states that all four business segments delivered triple-digit growth, driven by AI-related memory demand spanning data centers, computers, smartphones, and even vehicles. The company’s pricing power is closely linked to the memory market’s tight supply-demand balance, which the article says is helping expand profit margins as well as boosting revenue.
On earnings, the article reports that Micron’s third-quarter profit surged 1,368% to $24.67 per share. Management’s outlook for the most recently ended fourth quarter (which ended Aug. 31), as cited in the article, calls for sales of $50 billion and earnings of $30.73 per share, with final results expected at the end of September.
Valuation looks compelling—but the cycle risk is central
The article argues that Micron appears inexpensive on traditional valuation measures. Based on Micron’s trailing 12-month earnings of $44.23 per share, it cites a price-to-earnings ratio of 21, which it notes is below the S&P 500 and Nasdaq-100 valuations referenced in the piece. It also says that if Micron’s earnings rise to $155.03 per share in fiscal 2027, as expected by Wall Street per Yahoo Finance, the forward P/E would be about 6.
The central counterpoint is that memory pricing dynamics may not hold. The article’s viewpoint is that Wall Street expects Micron’s earnings to normalize over time—either because AI demand growth slows or because supply expands. It notes that memory manufacturers are racing to add manufacturing capacity, which could reduce Micron’s ability to dictate prices and pressure margins if the market moves from tightness toward balance.
What to watch next
For investors, the next test will be whether Micron can sustain its pricing power into the fourth quarter and beyond as AI buyers adjust to higher infrastructure costs. Key watch items include the company’s fourth-quarter results due at the end of September, any updates on hyperscale capex pace, and regulatory developments around data center construction timelines in U.S. states.
