Full Transparency Portfolio Returns: Firstly, I am a small account trying to grow to $1M. Total return this year: +98.80% 🚀 What a ride. We were knocking on the door of 180%% YTD before the recent correction in neocloud names cooled things off, but nearly doubling the account after being down -35% in Q1? I’ll take it. The volatility was wild, but the conviction held. Looking ahead to 2026, I am incredibly bullish. I believe the Semiconductor and Data Center supercycle is just getting started, and Bitcoin is primed to play some serious catch-up. The setup for next year looks massive. ⚡️ Majority of my holdings : $NBIS $RKLB $GOOG $AMZN $ONDS $TSLA $IREN $CIFR $BITF $TEM $PATH $PRME $NVDA $KRKNF $EOSE $QS

I bought more $CRWV $NBIS $MU $META today! Also entered small swing position in $IREN ! The prices are simply too attractive!

Memory names got smoked on the China CXMT story this week, and $MU slipped back under $900 after that $860-to-$931 bounce. SK Hynix is down ~26% in five days, Samsung ~22%, Micron ~11%, on fears denser Chinese DRAM hits late 2026 and floods the market by 2029. My take: treating HBM the same as commodity DRAM is the mistake. AI clusters still need high-bandwidth memory that multi-year contracts and tool constraints keep scarce, while CXMT's profit print looks mostly like standard server DRAM catching up. Elon even posted that China will make lots of RAM in 2028, and the tape priced that like every dollar of Hynix and Micron revenue is interchangeable with that stack. The only level I care about on $MU is whether the weekly channel rail near $914 holds on a close, not the hourly panic. Is CXMT a 2026 HBM problem or a 2029 commodity DRAM problem? NFA, just my read. DYOR.

Posting another loser, because the wins don't mean much without these: my $ASTS Jan 2028 $100 calls are down about 80%. I paid $31.71 a share ($3,171 per contract) for 3 of them ($9,512 all in), they were marked around $6.09 on my last screenshot, and that's roughly -$7.6K unrealized, still open. Here's the math that actually stings: breakeven at expiry is $131.71, which is 98% of the stock's 52-week high of $133.86. With $ASTS closing at $44.46 yesterday, it has to roughly triple (+196%) in the 462 days left just to get my money back. The selloff has real reasons: Piper Sandler cut it to Neutral with a $54 target on launch delays, and traders fear SpaceX's direct-to-cell push could crowd it out with the carriers. The bull case hasn't vanished either, with 13 satellites in orbit, over $1B of backlog, and Deutsche Bank still at Buy with a $93 target. The lesson for me: when the premium is almost a third of the strike, the breakeven quietly becomes "new all-time high", and that's a much bigger bet than it looks on the order ticket. Would you hold a call that needs a new ATH, or cut it and redeploy? NFA, just my trades. DYOR.

$GOOGL daily is squeezing inside a falling wedge with about 11 trendline touches, and price is just under the $351.69 trigger. Watch, not an entry: nothing has triggered yet. Levels: $351.69 trigger, $333.49 where the setup fails, $364 first resistance. Which way does it break? Not advice, just the chart. DYOR.

Memory just ate another roundhouse: $SKHY -26% in five days on China DRAM and 2029 oversupply chatter. AI demand's still screaming and DRAM is somehow 230% above 2025 levels. NFA, but this feels like a cycle scare inside a structural boom. Which story wins by Christmas?

Germany's lining up a $NVDA AI deal for factories, with Jensen in Berlin Monday. From Munich that feels less like hype and more like the EU finally buying GPUs instead of just regulating them. NFA — sovereign AI needs silicon, not slides. Does Merz write the check?

My $LITE buy order filled at $836 today, after the stock fell 27% from its $1,143 record high in just four sessions. I set that order on Monday because $836 was the first support on my chart, and today it sliced right through to tag $800 before bouncing to about $851. What I'm actually buying is plumbing: Lumentum makes the lasers and optics that move data between GPUs, and every bigger cluster needs more of them. The CEO told Bloomberg capacity is sold out through early 2029, some products can't cover 70% of next year's demand, and customers are already ordering into 2030. So the market is dumping a supplier that can't build fast enough, which looks to me like the $CRWV debt scare and Gartner's neocloud call splashing onto the whole AI chain. My math: $1,000 has capped it three times and sits about 20% above my fill, while a close under roughly $780 breaks the uptrend about 7% below, so I'm risking about $56 a share to make $164, roughly 3 to 1. If it closes under $780 I'll post that too, because I share the losers here as well. Are optics the AI bottleneck everyone keeps forgetting, or did I just catch a falling knife? NFA, just my trade. DYOR.

My buy order in $LITE got executed at 836!

I bought four long-dated AI compute calls in the last few days, and the market stress-tested every one of them by Wednesday. $NBIS closed -10% at $186, $IREN -12% at $30.56, $CIFR -12% at $9.94 and $INDI -9% at $2.01, on hot CPI, $CRWV bond yields near 15% and Gartner saying most neoclouds won't survive to 2030. So here's the only number I care about on a long call, the price each stock has to beat by expiry just to get my money back. NBIS Dec 2027 $200 call bought at $61.50 needs $261.50 (+41%), IREN Aug 2027 $35 call at $8.80 needs $43.80 (+43%), INDI May 2027 $2.50 call at $0.57 needs $3.07 (+53%), and CIFR Mar 2027 $17 call at $1.95 needs $18.95 (+91%). That's about $23.8K of cost, and NBIS is 78% of it, so one ticker carries most of the risk. CIFR is the long shot: the shortest runway (155 days), the steepest hill, and the $850M convertible that sank it Tuesday means more shares splitting any upside. My reason for going long-dated hasn't changed: compute still looks supply-constrained to me, and 428 days gives NBIS time, starting with earnings on Oct 22. If the neocloud bears are right, these can go to zero, and that's the deal you sign with long calls. NFA, just my trades. DYOR.

$MU on the weekly: $914 is sitting right on the lower rail of the rising channel it's been climbing since last autumn, and that's the spot I'm watching today. The 5 EMA is still above the 20 EMA, so the trend structure hasn't broken even after this week's red candle took price down from around $1,000. The chart's levels: a bounce entry near $914, with the invalidation (stop) under the channel at $816, about 11% below. The upside map from there is $1,036, $1,109 and $1,179, and that last one sits just under the June peak near $1,200. Support bounces fail when the channel line breaks on a weekly close, so that's the only thing I care about here, not the daily noise. Not advice, just the chart. DYOR.

Posting the losers too. My $CRWV CoreWeave Dec 2027 $135 calls (7 contracts) are down about 75%: avg cost $32.28, now trading around $8.13. That's roughly -$16.9K unrealized on $22.6K of cost basis, and it's still open, so it's not a realized loss yet. I bought them as a long-dated bet on AI compute demand, but I paid a rich premium for a strike far above the stock, and that's what hurts when the shares fade. The honest lesson: long-dated OTM calls feel cheap in time but are expensive in implied volatility, and a 428-day runway doesn't protect you from a bad entry. Same month, my $NBIS calls made +$6.8K, so my Oct 1-14 options realized P&L is +$9,020.78. The winners paid for the mistakes, but I want better sizing so one position can't hurt like this. Not advice, just my trades. DYOR.

4 days back I wrote that $RKLB may head to 53 & now we are at 54.6 After hours. $RKLB is at $54.60 after hours, another 7.7% down after closing at $59.18 (−7.9%). That's about −15% since yesterday. The regular session held just above the rising trendline at $58.48, but the after-hours drop broke straight through it. There's very little price history underneath, so the next supports are only rough round numbers, around $52 and then $50. The RSI is 33, close to oversold, and price is well below all its moving averages (50-day at $70, 200-day at $81). The first resistance on any bounce is $58.50 to 61.

$MU closed at $915.66, just shy of 912 I wrote below. That's below the ~$1,026 buy zone from the Oct 10 post, so that setup failed. It's now testing the floor of the rising channel at about $924.50. If it gets back above that level, the channel targets are about $1,018 in the middle and $1,111 at the top. A close below about $891 breaks the channel entirely. $ARM $AMD $ASML $NVDA $TSMC $CBRS $NBIS $IREN $CRWV $RKLB $META

$ANET -4.6% to $195. Broke a 6-touch rising trendline at $199.60 and lost the 50-day. Next support $182. Networking names led today's AI selloff. #AI #Semis

$ARM -6.5% to $238.72, sitting right on 3x-tested support at $238.40 after breaking its trendline at $248.45. Close below $238 and the floor gives way. #AI #Semis

$MRVL -7.5% to $248.53. Lost its 4-touch rising trendline at $259.70. Next support $242.86. A reclaim of $260 negates the break; until then, rallies get sold. #AI #Semis