$IREN Over two trading days, IREN fell from 34.80 to 26.90, a drop of 22.7%. It hit a 13-month intraday low tonight, while Bitcoin fell only about 2% over the same period. Every oversold indicator is flashing, but I looked through nearly five years of daily charts, and an oversold reading this extreme doesn’t guarantee a rebound.
First, let me own yesterday’s call. Last night I said I wouldn’t touch it unless it could reclaim 34.80, and I stood by that. But I used the July 29 low of 28.93 as a reference point below, and also mentioned that it had bounced back to 38 the very next day. Tonight, it plunged straight through that level at the open. My mistake was using the prettiest example as my benchmark.
Here’s how the backtest was done: since November 2021, IREN has had 23 instances of a two-day cumulative drop greater than 18%, with adjacent signals grouped together as one. Over the next 10 trading days, it rose only 11 times, and the median return was -3.2%. Put simply, the two weeks after a sharp drop are basically a coin toss. The July 29 instance, which bounced 30.5% the next day, was the biggest one-day rebound among those 23 instances—and I chose that one as my reference.
The source of the selling pressure over the past two days is clear: hotter-than-expected CPI lifted rate-hike expectations, raising financing costs for companies that borrow to build data centers. Funds are cutting their exposure to the entire AI-computing basket, selling everything indiscriminately, and the most actively traded names are getting hit first. NBIS fell 11% tonight, and CRWV fell 9%. IREN is now being priced for AI debt risk, not Bitcoin mining capacity. MARA, which is more purely focused on mining, fell only 7%—that’s the contrast.
Technically, the daily Bollinger lower band is at 28.90, and the price has already fallen below it. The KDJ J-value is -1.6; ATR(14) is 2.76, meaning a normal daily move is already around 10%. The 28.90–28.93 area above is where the lower band and the July low overlap; it becomes resistance first. Above that is yesterday’s close at 30.56. Below, I’m watching around 25.3, near the September 2025 price action.
My take: oversold readings are just noise for IREN. Unless rate-hike expectations cool, I’ll treat rallies in computing-power stocks as short covering. I’ll only consider testing a small position once the daily close gets back above 28.93. Until then, IREN won’t be in my account. Personal opinion
#美股 #AI算力 $BTC
First, let me own yesterday’s call. Last night I said I wouldn’t touch it unless it could reclaim 34.80, and I stood by that. But I used the July 29 low of 28.93 as a reference point below, and also mentioned that it had bounced back to 38 the very next day. Tonight, it plunged straight through that level at the open. My mistake was using the prettiest example as my benchmark.
Here’s how the backtest was done: since November 2021, IREN has had 23 instances of a two-day cumulative drop greater than 18%, with adjacent signals grouped together as one. Over the next 10 trading days, it rose only 11 times, and the median return was -3.2%. Put simply, the two weeks after a sharp drop are basically a coin toss. The July 29 instance, which bounced 30.5% the next day, was the biggest one-day rebound among those 23 instances—and I chose that one as my reference.
The source of the selling pressure over the past two days is clear: hotter-than-expected CPI lifted rate-hike expectations, raising financing costs for companies that borrow to build data centers. Funds are cutting their exposure to the entire AI-computing basket, selling everything indiscriminately, and the most actively traded names are getting hit first. NBIS fell 11% tonight, and CRWV fell 9%. IREN is now being priced for AI debt risk, not Bitcoin mining capacity. MARA, which is more purely focused on mining, fell only 7%—that’s the contrast.
Technically, the daily Bollinger lower band is at 28.90, and the price has already fallen below it. The KDJ J-value is -1.6; ATR(14) is 2.76, meaning a normal daily move is already around 10%. The 28.90–28.93 area above is where the lower band and the July low overlap; it becomes resistance first. Above that is yesterday’s close at 30.56. Below, I’m watching around 25.3, near the September 2025 price action.
My take: oversold readings are just noise for IREN. Unless rate-hike expectations cool, I’ll treat rallies in computing-power stocks as short covering. I’ll only consider testing a small position once the daily close gets back above 28.93. Until then, IREN won’t be in my account. Personal opinion
#美股 #AI算力 $BTC