The recent decline in semiconductor stocks represents a bear trap rather than a structural shift, according to Bank of America analyst Vivek Arya. The PHLX Semiconductor Index, tracked by the iShares Semiconductor ETF (NASDAQ: SOXX), fell 11% since the start of the third quarter following an 88% surge in the second quarter. This pullback occurred during the sector's historically weakest seasonal window, but the underlying driver of artificial intelligence infrastructure spending continues to accelerate.

Global cloud and AI infrastructure spending is on track to approach $1.5 trillion by 2027, marking a 40% to 50% increase year over year. Arya expects leadership to shift back toward companies directly tied to AI capital expenditures as visibility into 2027 spending improves during the second half of the year. Hyperscalers remain focused on maximizing AI utilization rather than cutting infrastructure spending, which should sustain demand for chips powering AI data centers.

Top AI Capital Expenditure Plays

Bank of America has identified seven companies positioned to win the AI buildout. These firms are leveraged to the ongoing capital spending cycle by hyperscalers and cloud providers.

Micron Technology: Top Pick

Micron Technology stands out as Bank of America's top pick, described by Arya as one of the market's biggest AI mispricings. Memory now accounts for roughly 35% to 40% of AI cloud capital spending, more than double historical levels. Despite this, memory stocks trade at modest valuation multiples due to investor concerns about traditional boom-and-bust pricing cycles.

Arya argues that memory is transitioning "from a cyclical commodity to a strategic AI enabler." Long-term supply agreements between memory suppliers and hyperscale customers are fundamentally changing industry economics by making pricing more durable and revenue streams more predictable. This shift could allow memory companies to command higher valuation multiples. The firm reiterated its Buy rating on Micron and maintained a $1,550 price target, implying roughly 59% upside from current levels.

Open-Source AI Impact

Concerns that increasingly capable Chinese open-weight AI models could reduce semiconductor demand are unfounded, according to the report. While lower-cost AI models may pressure software economics, Bank of America believes they will expand AI adoption by making inference cheaper and accelerating deployment. The report notes that broader AI usage will require more compute, memory, networking, and power infrastructure over time, posing a risk to model economics rather than semiconductor demand.

Micron Technology Inc has overtaken Nvidia Corp as the top-performing stock in the artificial intelligence sector over a five-year horizon. According to TradingView data through June 30, Micron's stock has risen roughly 1,320%, significantly outperforming Nvidia's 859% return. The shift in leadership occurred primarily over the past two months, marked by a near-vertical ascent in Micron's share price and a sideways consolidation in Nvidia's stock. Currently, Micron Technology has a market capitalization of $1.10 trillion.

The divergence in performance is underpinned by fundamental financial results. In its most recent quarter, Micron reported revenue of about $41.5 billion, representing a 346% year-over-year increase. This growth rate substantially exceeded Nvidia's quarterly revenue growth of 85% over the same period. The disparity is even more pronounced on the bottom line, with Micron posting adjusted earnings of $25.11 per share compared to Nvidia's $1.87.

Financial Performance Comparison

Valuation and Market Sentiment

Despite the significant rally, valuation metrics suggest both stocks remain relatively inexpensive compared to historical averages. Nvidia trades near 19.9 times next-twelve-month earnings estimates, below its 31.3x historical average. Micron trades at approximately 8.0x forward earnings, which is under its own 11.6x mean and well below Nvidia's multiple. This discount reflects market skepticism regarding the memory sector, which continues to be priced as a cyclical commodity rather than core AI infrastructure.

Future Outlook

The surge in Micron's performance is largely attributed to high-bandwidth memory (HBM), a critical component for AI accelerators. Tight supply and surging demand for HBM have driven prices, margins, and earnings higher. Analysts at 22V Research describe the transition as a shift from a "compute race" to a "memory race." Micron has reinforced this thesis by pointing to future demand from AI agents, robotics, and vehicles, citing long-term customer agreements worth roughly $100 billion through 2030.