been holding a small QCOM position since around $175 and honestly wasn't expecting this pop. the AI chip stuff is cool on paper but I keep thinking about how many times a legacy chipmaker announces a data center pivot and it ends up being 2% of revenue three years later. not selling into the summit but also not adding here, want to see actual customer names before I believe the story


TehCcc
intel closed 91.00 on monday, up 0.89%, which passes for excitement in a stock that was above 110 two weeks ago. this whole community is arguing about memory pricing while the cheapest way to own the datacenter buildout just sits there doing nothing. started a small position at 92, two year hold not a two week one.
$Oracle(ORCL.US) near 124, a 52-week low, down about 33% on the month on AI-spend-return worries. and today's TSMC capex scare is the exact same anxiety, the market suddenly cares whether all this AI investment pays off. Oracle's cloud revenue was up 47% so the growth is real, but the multiple compression is brutal. i'm watching, not catching, because a stock making new lows on a sector-wide fear can keep making them. no knife-catching.
$Roundhill Memory ETF (DRAM.US)$ pulled back with the whole complex on the profit-taking, and holding the basket is exactly why i'm calm instead of panicking over which single name dumped hardest. up 27 one day, down 8 the next, the basket smooths that insanity. i'm not adding into the whipsaw, waiting for TSMC this afternoon and for the group to stop bleeding before i put more in.
cool CPI did exactly what $Invesco QQQ Trust (QQQ.US)$ holders needed, 3.5% vs 3.8% expected, biggest monthly drop since 2020, and July hike odds collapsed from 42% to 17%. that's the rate pressure lifting off the whole index in one print. i hold QQQ long term and this is the kind of macro relief that lets the AI names run. still watching PPI today though, one print doesn't end the debate.
holding $Invesco QQQ Trust (QQQ.US)$ into this week feels like walking through a minefield: CPI Tuesday, five big banks Tuesday, Warsh testifying, Hormuz shut. the index is at highs and priced for a benign print. i'm not selling my long-term core but i'm definitely not adding before that CPI number. respect the calendar, one hot print resets everything.
$Direxion Daily Semiconductor Bull 3X (SOXL.US)$ had a monster day with the SOXX up 5%, this is the leverage working for you for once. but that is exactly the trap, the up days feel amazing and lure you into holding through the chop that decays you to zero. i took my intraday scalp and closed it. never overnight, never in size. the bear-market week should be a fresh reminder why.
funny detail: $Invesco QQQ Trust(QQQ.US) was forced to buy SpaceX two days ago and it's already below its debut price at 149. so every index holder now owns a fresh all-time-low position they didn't choose. that's the hidden cost of passive during hype inclusions. i hold QQQ long term and just shrug at these, but it's a reminder index buying isn't always smart buying.
$AMD(AMD.US) reportedly up about 7% Monday on Samsung's guidance and the NVDA denial. AI accelerator demand not going away, AMD is the second beneficiary after Nvidia. i'm in a small position, mostly watching whether the Samsung print translates into AMD custom silicon orders. NVDA denial helps AMD too, a healthy competitor makes the whole AI chip market look real.
$Intel(INTC.US) reopens today after rolling over hard off the $139 high, sitting around $127. the comeback trade got way ahead of itself and i already trimmed into the strength weeks ago, glad i did. keeping a small starter, not adding until it proves it can base out here. turnaround stories up 200% are where i get cautious, not greedy.
if you want to gamble on the memory bounce, the 2x Micron ETF exists, but understand what today did to it. $Micron Tech(MU.US) fell 5.5% so the 2x lost roughly 11% in a single session. leverage on a name already this volatile is a coin flip with extra steps. i'd rather own the underlying and actually sleep at night.
SK Hynix listing on Nasdaq July 10, basically 58 percent of the HBM market going public right when memory is the hottest trade on the street. soon the three giants all trade on US exchanges, with $Micron Tech(MU.US) already leading the pack. adding SK Hynix to my watchlist, not chasing day one though
the part of the Micron story people are underrating imo is the supply math, not the deal headline. if AI data centres are eating roughly 70% of memory supply and 2026 DRAM pricing forecasts are sitting at +125%, the earnings beat almost becomes secondary to the guidance. that's what i'm watching Wednesday: the FY26 commentary, not the FQ3 number.



