millionaire

What Tuesday tells us about SpaceX

SpaceX Q3 is the first real earnings test for proxy holders, and the bar is simple: show that Starlink is funding the launch business rather than the other way around. The 18.4 billion spend only works if subscriber growth and the new airwaves push into mobile keep compounding. Buying Grain Management spectrum is a direct move on the carriers, and if Starlink Mobile works indoors it reprices the whole wireless map. The capex is enormous and a private mark is not a public multiple, so risks are real. But if Tuesday shows cash generation inflecting, the space premium starts to look earned rather than hoped for. I am watching margins and the Starlink mix, not the rocket count.

Tesla holds the Semi rollout today at the new Sparks plant, which is built for up to 50,000 trucks a year. That number is the part worth focusing on. Semi has been in low volume production since late April, so today is about showing the line can actually scale, not about the truck itself. Fleet buyers care about total cost per mile and service network density, neither of which a launch event solves. I would treat this as a milestone for the manufacturing story and wait for real fleet order numbers before rerating anything.

sold my mrvl at 198 in july because it looked extended to me. closed 245 yesterday and then popped to 256 after hours on nvda's print. average marvell holder, except i am not one anymore 🤡

still +188% ytd and here i am upset about one 5% day. touch grass moment for me honestly 🫠

bought $Micron Tech(MU.US) at 1180 in june because the thesis was so obvious. thesis still obvious, i am still down 30% 🤡

soxl printed today, 3x semis and the chip index up 5% so you can imagine. rode it up and took some off into the close, not greedy. leverage cuts both ways and i've been burned before. lock in the gains on green days like this

intel 18A yield reportedly went from 65 to 85% and there is chatter about Nvidia and OpenAI foundry orders. if even half of that is real the turnaround thesis has legs. but i have been burned by intel promises before, so i am waiting for the 7/23 print before i believe anything 🫠

$Unity Software(U.US)11.SG$ pushing near its all-time high around S$40 while global tech had a rough session. the SG bank trade has quietly been one of the best places to hide this year, records while the AI names whipsaw 30% peak to trough. i hold UOB for the yield and the stability, it's the least dramatic position in my book and that's the point. not adding at the ATH, just letting the dividend compound.

the 2x SK Hynix ETF is the perfect illustration of why leverage on a whipsaw kills you. up huge Tuesday on the +27%, dumped hard Wednesday on the give-back. you can be right on direction over a week and still get chopped to pieces daily. i keep memory exposure in the plain $Roundhill Memory ETF (DRAM.US)$ basket, never the 2x. the decay plus this volatility is a blender.

the 2x SK Hynix ETF just did a moonshot with the underlying +27%, and this is exactly the tool that also nuked people on Monday's 15% crash. leverage amplifying a V-shaped whipsaw is a heart attack machine. i hold plain memory exposure, no 2x, and i sleep. if you rode the 2x through both days respect, but that's gambling not investing lah.

$Roundhill Memory ETF (DRAM.US)$ down with the whole complex on the SK Hynix profit scare, but this is exactly why i hold the basket instead of picking SanDisk or Micron individually, no single -15% Korean session guts me. the undersupply-through-2030 story didn't change today, only the margin timing did. adding slowly on weakness, not chasing, and definitely waiting for CPI tonight first.

$Roundhill Memory ETF (DRAM.US)$ is the lazy way i'm playing "memory undersupplied through 2030." SK Hynix debut, Micron near 980, SanDisk with a $3,000 target, all in one basket without picking the winner. the whole complex got validated Friday. i'm adding on any CPI-week dip, this is a multi-year hold not a trade for me.

SK Hynix when-issued trading starts tonight under SKHYV, regular trading Monday. i'm not touching the debut, IPO first-day moves are a casino. but the fact that the second-biggest US IPO in history is a memory company, landing right as memory bounced out of a bear market, is a pretty loud signal. holding my $Roundhill Memory ETF (DRAM.US)$ basket and watching how it opens.

the chip divergence is brutal for $Direxion Semicon Bull 3X(SOXL.US) holders because the index nets out NVDA up and Intel down into chop, and chop is where the 3x daily reset decay eats you alive. i only ever scalp this intraday. if you're holding it overnight through a sector this split, you're paying the decay tax every single day. respect the instrument.

today was exactly why i never hold $Direxion Daily Semiconductor Bull 3X (SOXL.US)$ overnight. SMH fell over 3% so the 3x got hit around 9-10%. i was flat into the close, thank god. if you're bag-holding this through a chip bear market the daily reset decay will bury you even if semis eventually recover. this is a scalp tool, not an investment.

the semi bounce Monday was real and $Direxion Daily Semiconductor Bull 3X (SOXL.US)$ ripped hard with it. i scalped a piece intraday and closed it, i do not sleep with leverage overnight. if the rally holds today i might do it again. if it fades the leverage cuts the other way fast. Samsung's $58B print is a tailwind but 3x is 3x.

Sembcorp Industries quietly had one of the most active trading days on SGX friday while everyone watched US chips. FY25 net profit around S$1B, analysts target near S$6.89 vs ~S$6.19 now, and earnings land July 30. i hold a starter for the yield and the energy transition angle. adding a bit before the print, this one flies under the radar.

reminder from today: $Direxion Daily Semiconductor Bull 3X (SOXL.US)$ got torched when the SOX fell 5%, that's the 3x working against you. i only ever scalp this intraday, never hold overnight. the leverage plus daily reset decay will bleed you dry in choppy tape. respect the tool or it eats you alive.

nasdaq 100 chopping around as tech splits, some names red some green. not making any big bets at these levels, just dollar cost averaging the index and ignoring the daily noise. boring but it works for me

Numbers are insane no doubt, but the stock already ripped 17% and it's sitting at $1,231. Buying memory the day after a blowout print is exactly how people end up holding the cycle top. I want to see it hold $1,180 before I touch it.

SNDK got dumped -13% then another -5% into the Micron print and now memory rips. classic shake out the weak hands right before the good news sia. NAND supercycle story is intact, im looking at this dip around 1,985 as a gift not a warning.

SanDisk ripped to a fresh all-time high close around $2,273 today riding the Micron wave, and the analyst tape is stacking up: BofA top-5 AI pick, Barclays upgrade to Overweight, Mizuho lifting FY27 revenue to roughly $45B. i added a small starter here, NAND pricing forecasts at +234% next year is hard to ignore.