To The Moon

$Direxion Daily Semiconductors Bear 3x Shares(SOXS)$  Just for fun, I'm calling the high somewhere between 51.74 and 53.60. I held overnight, which isn't something I do often, with an average at 44.89.

$IREN Ltd(IREN)$  Been watching since after earnings. I jumped in at 27.53. It looks primed for a recovery at this level.

$Direxion Daily Semiconductors Bear 3x Shares(SOXS)$ If we could get a 20% move up to around 55, I'd definitely close out all my calls at that point and rotate into SOXL.

$IREN Ltd(IREN)$  Daily and 4hr RSI are both below 25. That's deeply oversold territory at this point, so I'm expecting a bounce. If it was worth liking above 30, it's hard not to like it even more down here. Nothing has changed in the business fundamentals, and AI isn't going anywhere.

A 60% pullback on a company growing revenue around 69% and building out production in Texas feels like an overreaction to me. $Applied Optoelectronics(AAOI)$  isn't risk-free, but the demand story didn't vanish in one week. Watching for stabilization.

YTD neocloud performance: $NEBIUS(NBIS)$  +83% is the only notable winner. The rest are all in the red: $TeraWulf Inc.(WULF)$  -8.32%, $CoreWeave, Inc.(CRWV)$  -22%, $APPLIED DIGITAL CORP(APLD)$  -36%, $IREN Ltd(IREN)$  -37%.

$Micron Technology(MU)$ $SK hynix(SKHY)$  KOSPI. Korea pumping. At least they came to their senses after the FUD around CXMT selling. Cheap prices today were a bonus. Now the next wall of worry is CPI.

$Applied Optoelectronics(AAOI)$ AMD has recovered its earlier losses and just turned green. It's still up 15% over the past month. The Nasdaq is heading higher too, so maybe the panic will settle down soon.

$SpaceX(SPCX)$ Bought 160c 10/16 contracts. Light position. I'm willing to let the premiums go to zero. Discipline over emotions. NFA.

$Invesco QQQ(QQQ)$ Price target $3000. CXMT claimed they would make DDR5 memory, but it isn't being sold in China or anywhere else. CXMT doesn't have EUV, so they can't make 4E2 memory. $SK hynix(SKHY)$  is 2-4 years ahead of CXMT in 4E2, and they have no production plans for it. 4E2 memory has issues with leakage and ECC failures. CXMT can't make HBM4 or HBM4E, and there's no plan for it either. $Micron Technology(MU)$ , $SK hynix(SKHY)$ , and Samsung generate most of their profits from HBM. They do make some other memory products, but the HBM revenue is what keeps their margins stable. CXMT can only se