Semiconductor stocks experienced a dramatic surge in April, with several names posting 70%+ one-month gains that would outperform one-year returns for most portfolios. According to CNBC's Halftime Report episode "Can the Market Move Higher in May?" (5/1/26), the rally has created a clear split among market participants: whether to keep buying after such large moves or sell on the risk that momentum has gone too far. The bullish argument centers on prices finally catching up with improving underlying fundamentals, with Q1 earnings showing overall growth near 29% and tech earnings exceeding 40%, both well above expectations. This growth is being driven by roughly $700 billion in hyperscaler data center spending, fueling sustained demand for chips.
The semiconductor sector has experienced a dramatic leadership change over the past month, with Nvidia (NASDAQ:NVDA) falling to third place among major chip stocks. According to reports from Investing.com India, Nvidia has gained only 7% over the first four months of 2026, significantly underperforming the broader CPU rotation. Advanced Micro Devices (AMD) has surged 70.4%, Micron Technology (MU) has added 44.7%, and even Intel (INTC) has outpaced Nvidia by a wide margin. The divergence has widened further over the past week, with Nvidia falling 4.3% while AMD posted another 3% gain and Micron surged 7.3%. Nvidia is now the laggard within its own sector, an unfamiliar position for a stock that has effectively been the market for the last 24 months.
The shift from general-purpose to custom AI chips represents a fundamental change in how major data center operators are spending on AI infrastructure. As reported by The Motley Fool, hyperscalers are investing heavily in custom application-specific integrated circuits (ASICs) designed for highly specific workloads. Counterpoint Research projects that shipments of custom ASICs are going to triple by 2027 compared to 2024 levels, with makers of custom chips including Marvell and Broadcom capturing a sizable chunk of that growth. Alphabet has been partnering with Broadcom since 2016 to design custom Tensor Processing Units, recently extending their agreement through 2031. Amazon has a partnership deal with Marvell running through 2029 for its Trainium chips, while Marvell calls its custom AI accelerators XPUs with sales climbing rapidly. Even Nvidia has joined the trend, announcing a $2 billion investment in Marvell in March and forming a strategic partnership through NVLink Fusion.
Much of the Nvidia to CPU rotation stems from positioning dynamics, as reported by Investing.com India. Nvidia entered 2026 as the most-owned, most-crowded, and most-hedged name on the Street. When relative performance flips, the marginal dollar chases laggards rather than leaders, and CPU names with depressed valuations become the obvious release valve. The shift reflects the market beginning to discount a fundamental change in AI infrastructure requirements, where hyperscalers building out agentic infrastructure cannot rely solely on Nvidia silicon. This positioning factor explains why crowded trades exhaust themselves long before fundamentals roll over, making the current rotation a natural market correction rather than a fundamental shift. The pattern seen in April's rally is typical of late-stage momentum moves, where the most speculative names lead while largest players lag.
The Nvidia to CPU rotation represents a longer-term thesis worth respecting but not a trade worth chasing into earnings, as noted by Investing.com India. AMD reports earnings tomorrow, and Nvidia reports later this month. If AMD guides cautiously on data-center momentum or Nvidia delivers another beat-and-raise quarter with strong commentary, a large portion of this rotation could reverse in a matter of days. The upcoming earnings reports will be crucial in confirming whether the current market rotation reflects fundamental shifts in AI infrastructure requirements or temporary positioning adjustments. Intel has already demonstrated strong performance with non-GAAP EPS of $0.29 against a $0.0127 consensus and revenue of $13.577 billion, with Data Center and AI revenue up 22% YoY. CEO Lip-Bu Tan's commentary tied the rally to agentic AI driving demand for Intel's CPUs and packaging, while Wolfe Research has called chips "overbought, extended, overhyped" and questioned whether 60 to 70% monthly gains are sustainable.
