Micron, AMD, Intel Stocks On Pace For Abrupt Reversal: Top AI Expert – Advanced Micro Devices (NASDAQ:AMD
Four of the most heavily traded AI semiconductor stocks just hit “extreme exhaustion” readings on a composite technical model that has historically preceded sharp reversals.
Visser, head of AI Macro Nexus Research at 22V Research and former chief investment officer at Weiss Multi-Strategy Advisers, said in his latest weekly video he scaled fully out of his Micron position in response.
“Definitely out of all of Micron,” he said.
4 AI Stocks Most At Risk Of Rally Reversal
The exhaustion score is a 1-to-100 composite that combines several technical inputs – such as the RSI on both the 14-day and 5-day timeframes, Williams %R, the stock’s 20-day position relative to its range, 20-day and 50-day extension from average true range – into a single reading.
A score above 75 means the stock is statistically stretched across most of these measures simultaneously. The reading does not predict the timing of a reversal. It quantifies how far a name has traveled from its normal trading range, and historically how vulnerable that makes it to a sharp pullback when the next catalyst arrives.
The gap Visser’s exhaustion model is highlighting is between fundamentals and price velocity.
The earnings justify higher prices. They do not necessarily justify the pace at which prices have arrived at those levels.
Qualcomm Inc. — Exhaustion Score 91.9. The highest reading in the entire AI Macro Nexus universe. Qualcomm has rallied more than 50% in just over a month on the back of its expanding AI inference chip business, with management positioning Snapdragon platforms as the on-device AI standard. The recent move sits nearly five standard deviations above the stock’s own historical noise.
Advanced Micro Devices Inc. — Exhaustion Score 90 AMD is the most stretched of the four on a pure price basis, trading at the very top of its recent range as investors price the MI400 series ramp and a wave of hyperscaler order announcements. RSI readings on multiple timeframes confirm overbought conditions on every measure.
Micron Technology Inc. — Exhaustion Score 88.2. Micron has been the poster child of the AI memory supercycle, returning 154% year-to-date as HBM3E and HBM4 capacity sold out through 2026. The options market is pricing the highest near-term volatility of the four, a sign traders already expect a larger-than-normal move in either direction.
Other Signals Point To Downside Risks Too
“AI semiconductor momentum is still alive, but it is stretched,” Visser said. “The next phase is likely more selective and more volatile.”
The exhaustion scores alone are not the trigger. Paired with them are reversal day patterns on the daily charts that suggest distribution may already be underway in the names the model is flagging.
Visser identified reversal candles this week on the South Korean KOSPI 50, on SK Hynix, on Japan’s broader chip complex, and on Micron itself, which printed large wicks on both sides of the bar in seven of the last eight trading sessions.
Large wicks at exhaustion readings mean buyers and sellers are fighting at exact price levels, often the technical sign that distribution has begun.
Visser has described the current market environment as one of “bubbles, parabolas and speed crashes.”
The distinction matters operationally. A traditional bubble takes years to inflate and unwind.
The current parabolas are forming in weeks and unwinding in days.
Visser’s Position Change On Micron
Visser bought Micron during 2024 when, in his words, “nobody wanted to touch it.” His entry came in the low-$100s range.
He sold the first tranche in the high-$300s, scaled out further as the stock pushed through $500, and finished closing the position into the recent $700 area.
His average exit, by his own estimate, came in the high-$300s to low-$400s.
“This is not a negative view. This is not me wanting to short this stuff,” he said.
The framing matters. Visser is not abandoning the AI trade thematically.
Nvidia CEO Jensen Huang‘s comment this week that agentic AI requires roughly 1,000 times the compute of generative AI is, in Visser’s view, the kind of demand signal that makes any short call on the secular trend a mistake.
What he is doing is reducing exposure to the names where the technical setup says the next two-to-three months carry more downside than upside, while accumulating in assets where the equivalent setup looks washed out rather than stretched.
Image: Shutterstock
