Bank of America has elevated Micron Technology to its preferred name in the semiconductor memory space, citing the growing strategic role of memory in artificial intelligence systems and the potential for multiples expansion as industry dynamics change.
In a note outlining its investment case, the bank asserts that memory has become a structural and long-lasting investment trend. Memory now accounts for roughly 35-40% of cloud AI capital expenditures - a share the bank says is two to three times higher than historical norms. Despite that shift, the bank observes that memory stocks as a group are trading at subdued valuations, around 10 times forward price-to-earnings.
Bank of America reiterates a Buy rating on Micron and sets a price objective of $1,550 per share. The firm contends that market participants may be underestimating the industry's move toward longer-duration commercial arrangements and pricing that is more predictable than in prior cyclical eras.
To arrive at its target, the bank applies a sum-of-parts framework. Under that approach, it assigns $1,040 per share to Micron's conventional cyclical memory operations, valuing that portion at 3 times the CY28E price-to-book ratio. The 3x figure sits at the high end of Micron's historical valuation range of 0.8x to 3.1x and is intended to reflect the possibility of a memory upcycle. The bank values Micron's high bandwidth memory (HBM) business, which services AI compute workloads, at 31 times CY28E price-to-earnings, aligned with the median multiple of the AI compute peer group.
Bank of America argues that as memory's role shifts from commodity to strategic enabler for AI, valuation multiples should expand accordingly. That re-rating is central to the firm's bullish outlook for Micron.
However, the bank also highlights several downside scenarios that could undermine its thesis. These risks include larger-than-expected declines in memory average selling prices, intensified competition from new entrants in China, the possibility of market share losses to larger competitors, and a pullback in demand across key end markets such as data centers, smartphones, and personal computers.
On the corporate development front, Micron has announced plans to boost U.S. investment to more than $250 billion through 2035 and to invest up to $3 billion to strengthen the domestic semiconductor supply chain. After meetings with company management, analysts at TD Cowen reiterated a Buy rating on Micron, noting that physical constraints are keeping supply tight while demand remains robust.
Bank of America's analysis combines an elevated view of memory's importance in AI with a valuation framework that separates legacy cyclical exposure from higher-multiple AI-oriented assets. The firm acknowledges risks to that view even as it makes the case for a material re-rating if the industry continues its transition toward longer-duration contracts and more stable pricing dynamics.
Key points
- Memory now represents approximately 35-40% of cloud AI capital expenditures, two to three times historical levels.
- Bank of America sets a $1,550 price objective for Micron using a sum-of-parts valuation: $1,040 for the cyclical memory business and a 31x CY28E P/E on the AI HBM segment.
- Memory stocks trade near 10 times forward P/E despite the secular shift toward AI-driven demand; sectors impacted include semiconductors, cloud/data centers, and consumer electronics.
Risks and uncertainties
- Downside if memory average selling prices fall more than expected - this would directly affect semiconductor revenues and margins.
- Increased competition from Chinese entrants or share losses to large competitors could erode market position in memory and AI HBM.
- Softening demand in major end markets such as data centers, smartphones, or personal computers would weaken the investment thesis for memory-focused equities.