A technical indicator with a nearly perfect track record is flashing red for semiconductor stocks, suggesting the sector’s historic rally could be on borrowed time.
A Bank of America report warns that the semiconductor sector is now more overbought than at any point in the last three decades, including the dot-com bubble. The analysis highlights seven similar historical signals since 1995 which were followed by an average market correction of 44 percent. The warning comes as investors have piled into chip stocks, chasing a rally fueled by optimism over artificial intelligence.
“Semiconductors remain hot, but risks are rising,” Bank of America technical strategist Paul Ciana said in the May 18 report. The note points to the VanEck Semiconductor ETF (SMH), a proxy for the sector, which has surged roughly 50 percent this year and is now trading at a record 150 percent above its 200-week moving average, a long-term trend anchor.
This extreme deviation from trend surpasses the 100 to 108 percent premiums seen at the peaks of 2021 and 2024. The sector’s 14-week relative strength index (RSI), a measure of momentum, has also pushed above 80, a level considered historically overbought. The alert is amplified by broader market froth, with a separate BofA survey showing fund manager cash levels have fallen below 4.0 percent, a contrarian sell signal, while the 30-year Treasury yield has climbed above 5.18 percent.
The signal does not predict an immediate crash but rather the start of a “top process” that historically takes an average of 21 weeks to play out before a significant peak is reached. This puts the onus on investors to weigh the euphoric momentum against a grim historical precedent that suggests the next major move could be down.
The bearish technical warning for the broader sector is running directly into a bullish fundamental case for specific companies within it. In a note published just days apart, Bank of America’s own fundamental analysts nearly doubled their price target on memory chip maker Micron Technology Inc. to $950 from $500, citing a massive upgrade to the total addressable market for AI data centers.
This apparent contradiction highlights the key debate for investors: are the sky-high valuations justified by a new paradigm in AI-driven earnings, or are they a classic symptom of a speculative bubble? The Micron upgrade is based on a structural belief that the AI memory supercycle is reshaping the company's long-term earnings power. BofA’s fundamental team raised its 2030 AI data center market estimate to $1.7 trillion, arguing that every AI accelerator requires high-bandwidth memory, growing Micron's slice of a much larger pie.
The technical signal, however, is indifferent to fundamentals. It simply measures price and momentum, warning that investor sentiment has reached a level of euphoria that has historically been unsustainable, regardless of the underlying story.
History shows that extreme overbought signals are not an immediate sell button but the beginning of a more volatile and dangerous phase. The most recent parallel from early 2024 saw the SMH ETF’s RSI break above 80, followed by a brief 14 percent pullback. The ETF then surged to a new high, but the RSI did not, creating a bearish divergence that preceded a much larger 29 percent correction. The entire process from the initial signal to the eventual low took 57 weeks.
The most extreme example remains the 2000 internet bubble, where the Philadelphia Semiconductor Index (SOX) ultimately crashed 85 percent after a similar signal appeared. While BofA does not present this as a base case, it serves as a reminder of the potential tail risk in the market.
For now, investors are left watching for key confirmation signals that would validate the warning. These include a bearish divergence where price makes a new high but the RSI fails to, the appearance of weekly reversal candles, and, most critically, a break below the 50-week moving average. History suggests that once this key support level is lost, a much deeper correction toward the 200-week average often follows.
This article is for informational purposes only and does not constitute investment advice.
