U.S. chip stocks are severely overbought! Bank of America warns: After seven similar signals in history, the average plunge was 44%.

```

The upward momentum of semiconductor stocks is still continuing, but technical indicators have entered a historically extreme zone. Bank of America warns that while similar overbought signals in the past were not immediately signs of a market peak, they often signaled the beginning of a topping process.

According to Chase Trading Desk, BofA Securities technical strategist Paul Ciana pointed out in a research report on May 18: “Semiconductors remain hot, but risks are rising.” VanEck Semiconductor ETF (SMH) has risen about 50% so far this year, and about 35% since the US-Iran ceasefire news on April 7. As of the week ending May 15, SMH was close to $560, with the 14-week RSI above 80 for the second consecutive week, reaching a historical high.

More extreme is that SMH's current price is about 150% higher than its 200-week simple moving average, surpassing historical highs of about 100% to 108% seen in 2021 and 2024.

This means the market has yet to give a clear reversal signal, but the margin for error is decreasing. Since 1995, the semiconductor sector has experienced 7 similar periods of extreme overbought. Historical samples show that it takes an average of about 21 weeks from RSI rising to around 80 to the eventual peak; about 36 weeks to break key support; about 73 weeks to the local low; with an average drawdown of about 44%.

Technical indicators reach historic extremes

The question in this round of semiconductor rally is not about strength, but about how strong it is.

As of the week ending May 15, SMH's 14-week RSI closed above 80 for the second week in a row. This is the fifth time since 2012 that weekly RSI exceeded 80. In traditional technical analysis, RSI above 70 is often seen as overbought, but in strong trends, 80 is closer to the “overheated” threshold.

Divergence from moving averages better reflects the abnormality of the current market. SMH’s price is about 150% higher than its 200-week simple moving average, a distance that has exceeded the highs of about 100% to 108% in 2021 and 2024. In other words, the current price has significantly deviated from its long-term trend anchor.

The TD Sequential indicator has also issued an upward exhaustion signal. In the week ending May 15, SMH triggered the “red 13” signal, with the risk level corresponding to about $619.67. Calculated resistance levels include around $595, further extending resistance near $619, $642, and $720.

These points themselves cannot define the top. The more crucial question is whether momentum can keep pace if the price continues to move towards these areas. If price hits new highs while RSI does not, the topping process will be closer to the dangerous phases documented in historical samples.

Historical samples show that tops are rarely formed in a single day

BofA’s core judgment is not that “the top is here,” but that risks are rising.

Since 1995, the semiconductor sector has experienced 7 similar extreme overbought phases. Historical paths show it takes an average of about 21 weeks from RSI rising to 80 to the final peak; about 36 weeks to break key support; about 73 weeks to the stage low; with an average drawdown of about 44%.

The most severe sample was the bursting of the internet bubble in 2000. The SOX index took about 137 weeks from the signal to the low, with a peak-to-trough decline of about 85%. The mildest was the early stage after SMH’s launch in 2012, with a drawdown of about 18%.

This means RSI above 80 is not equivalent to a sell button. In strong trends, RSI can remain high for a long time. Exiting too early due to overbought may miss subsequent gains. But historical experience also shows that after extreme overbought, the market often enters a phase of greater volatility.

A common path is: first a short-term pullback, then price hits new highs but momentum begins to diverge. Afterwards, high-level candlesticks weaken, consolidation forms, and finally breaking key moving averages establishes a more destructive top structure.

2024 provides the closest reference

The price movement in 2024 is the closest historical comparison to the current market.

In the week ending March 8, 2024, SMH closed at $224.99 with weekly RSI breaking above 80. The price quickly surged, forming a warning long-legged doji. In the next seven weeks, SMH pulled back about 14%.

But that was not the ultimate top. Over the following three months, prices soared again to new highs, but RSI did not keep pace and a bearish divergence became clear. SMH truly peaked only 18 weeks after the initial RSI signal, then fell towards the 50-week moving average with a drawdown of about 29%.

Afterwards, SMH saw another round of rebound of about 42%, but high-level candlesticks repeatedly showed long upper shadows, indicating insufficient buying momentum. Then SMH entered a roughly 30-week triangle consolidation, eventually gapping down and breaking below the 50-week moving average, leading to a deeper drawdown of about 37%.

It was not until price approached the 200-week average and RSI fell below 30 that SMH formed a V-bottom. From the initial RSI above 80 to the key low, the entire process took about 57 weeks.

The significance of this case is that after extreme overbought, the market may first pull back, then trigger another wave of chasing rally, before entering the real decline phase.

2017 and 2014 show that the longer the top drags out, the more dangerous it becomes

In the week ending October 27, 2017, SMH closed at $50.30 with weekly RSI breaking above 80. Price continued rising for four more weeks, then entered broad oscillation for about 10 months.

This phase saw double-digit gains and declines, ending with a rare diamond top. Structurally, price first exhibited an expanding pattern with higher highs and lower lows, concurrent with RSI divergence; then switched to convergence with lower highs and higher lows, forming consolidation.

The true confirmation signal was breaking the 50-week moving average. Afterwards, SMH experienced a drawdown of about 26%, with the key low appearing about 61 weeks after the initial signal.

The process in 2014 was even longer. In the week ending July 7, SMH closed at $25.47, RSI breaking above 80. There was a pullback, then a new high, but RSI divergence appeared. The first round saw a drawdown of about 15.7%, and the 50-week moving average gave support, allowing the trend to continue.

But divergence did not disappear. Price later formed an ascending wedge top, with two bearish engulfing patterns at the high. In June 2015, after breaking wedge support and then in July breaking the 50-week moving average, the subsequent decline from 2015 highs was about 27%. Eventually, price formed a double bottom near the 200-week average, with two key lows arriving about 60 and 84 weeks after the initial signal.

These two samples show that post-extreme overbought tops may drag out for a long time. The longer the duration, the easier it is for investors to overlook the initial risk signals.

2000 is the reference for tail risk

The 2000 sample is the most extreme version.

In the week ending February 25, 2000, SOX closed at 1032.02, weekly RSI breaking above 80. The index peaked three weeks later. Over the next six months, the index entered a distribution phase, with six swings ranging 31% to 55%, forming a triangle top.

Even as price repeatedly rebounded, RSI continued declining, showing clear exhaustion of momentum. Eventually, price broke the 50-week moving average, confirming a major top, then continued downwards.

The initial attempt to stabilize near the 200-week moving average also failed. The potential inverse head-and-shoulders did not materialize as the right shoulder failed to break the neckline and 50-week moving average. Price then broke below the 200-week average, extending the bear market. Even though there was a bear market rally of about 86%, it did not reverse the main trend.

The entire cycle did not bottom until the second half of 2002, with a peak-to-trough fall of about 85%.

2000 is not the baseline forecast for the current market, but it does provide a reference for tail risk: when extreme overbought stacks with topping patterns, the 50-week moving average fails, and even the 200-week average cannot hold, the technical outlook switches from correction to bear market.

Three main confirmation signals are most important now

Currently, SMH does not have a clear topping pattern or a definitive blow-off end. RSI alone and TD exhaustion signals cannot directly conclude that “the top is here.”

Investors need to focus on three types of confirmation signals.

First, whether price and RSI diverge. If SMH keeps hitting new highs but the 14-week RSI does not, historically this usually signals an acceleration of the topping process.

Second, whether high-level candlesticks weaken. Long upper shadows, long-legged doji, bearish engulfing, outside reversal weekly bars are all more important than simply “too much rise.”

Third, whether the 50-week moving average is breached. Historical samples show that pulling back to the 50-week average is not necessarily fatal; the real danger is failing to recover quickly after breaking below it. Deeper corrections often only find stability near the 200-week moving average.

Bulls can still hold the trend, but should not neglect protection

Before there is confirmation of a top, betting directly on a reversal is not advantageous. The trend could still drive SMH to $595, $619, $642, or even further extended resistance.

But continuing to hold and allowing risk to run unchecked are not the same. Approaches within the framework include three types: using trailing stops to preserve trend exposure while protecting realized gains; moderately reducing positions and allocating released funds for downside protection; selling out-of-the-money calls to finance protective put options.

The commonality of these actions is not being bearish on semiconductors, but acknowledging that the current position is no longer suitable for bearing risk without protection. Particularly, option strategies themselves entail expiry, zeroing, and assignment risks, and are not suitable for all investors.

The semiconductor rally has not yet signaled an end. But from historical experience, the most dangerous stage is often not when it can’t rise, but when it rises too smoothly, too quickly, too far, and the market starts treating extremes as normal.

~~~~~~~~~~~~~~~~~~~~~~~~

The above highlights are from Chase Trading Desk.

For more detailed analysis, including real-time commentary and frontline research, please join [Chase Trading Desk ▪ Annual Membership]

Risk Disclosure and DisclaimerThe market carries risk, and investment requires caution. This article does not constitute personalized investment advice and does not take into account individual users’ specific investment objectives, financial situation, or needs. Users should consider whether any opinions, views, or conclusions in this article fit their circumstances. Investment based on this is at your own risk. ```