FromRedToRiches

Belgian private investor managing a €2M+ portfolio.
AI, semis & asymmetric bets.
Sharing positions, decisions, wins and mistakes. Nothing to sell.

[FromRedToRiches @FromRedToRiches]

1/6

This is what my €2.22M portfolio looks like today.

€1.6M initially invested.
€2.4M at the peak.

30 positions. Top 10 = 55.5%.

Still heavily concentrated in AI infrastructure.

Holdings, thesis, mistakes and what I’m changing next 🧵

[FromRedToRiches @FromRedToRiches]

Full blotter — Jul 27

OUT
$GOOGL 300 @ $327.66 · $NOW 800 @ $102.85 $ZS 432 @ $145.73 · $APP 145 @ $407.13 · $ISRG 100 @ $355.52

TRIMMED
$RBRK, half the line @ $75.35

IN
$AMD 140 @ $479.56 · $SNDK 51 @ $1,264.07 $MPWR 40 @ $1,300.87 · $GLW 240 @ $138.77

ADDED
MRVL, AMAT, KLAC, VRT, COHR, FORM, AMKR, TER, APH, GEV

Best: INOD +6.3%, RKLB +4.7%
Worst: AMKR −6.5%, Cerebras −5.3%, FORM −5.0%

(today's new positions excluded, their % is vs my entry, not a full session) How to read the numbers: the headline % is the change in my total portfolio value vs Friday's card, no deposits or withdrawals :€2,114,557 vs €2,138,037 = −1.1%. Line moves are close-to-close in EUR, so FX can shift them slightly vs the USD chart you'll pull up. 26 orders, €330,456 sold and put straight back to work. Portfolio €2,114,557.

[FromRedToRiches @FromRedToRiches]

Daily recap - Jul 27

Portfolio -1.1% vs SOX -2.2%. +1.1 pt.

Premarket had semis pointing to +2.7%. Instead the SOX printed 11,217 mid-session, a 5% slide from Friday's close, a swing that had me more stressed than I'd like to admit (my own gauge on the card says 9/10).

I used it: sold the last lines outside my AI basket : Alphabet, ServiceNow, Zscaler, AppLovin, Intuitive Surgical, €330,456, every one green at my exit, and put all of it back into the drop. 4 new positions, 10 adds.

Amkor reported tonight. Very strong numbers, guidance a touch soft. After-hours swung from −5% to +5%: no signal for the open. But the print backs my core view, this cycle's growth is real and far less cyclical than the market prices it. I still expect the SOX to retake its highs before the end of 2026. That's 27% from here.

Big week: the Fed decides Wednesday, and the hyperscalers report their capex. My whole basket trades on those numbers.

Full blotter in reply ↓

[FromRedToRiches @FromRedToRiches]

Ouch. That one really hurt.

Semis were indicated up 2.7% premarket. We opened up 0.6%. In between, The Information reported that a state-backed Shanghai company has started producing immersion DUV lithography machines, the tools ASML has effectively owned for twenty years.

What followed: SOX from 11,949 to 11,217 in about an hour. Minus 6.1% high to low. ASML went from +1.9% to -9.5% in Amsterdam, an 11% range on the most important company in the industry, on 2.7x its average volume. That's not noise, that's people repositioning.

Second brutal session in a row after Friday's -4.25%. At today's low we were down 9% in two sessions and more than 23% below the June high.

We've bounced back over 11,300 as I write this. I'm not calling a bottom, I have no idea.

My book is built on semis. So today hurt the P&L and it hurt the morale, and I'm not going to pretend otherwise. That's what running it this way costs, and I knew that going in.

Now here's what I keep coming back to. The same report says five machines this year. Around twenty in 2027. Still behind on performance and reliability, still needing more testing. That's the actual size of the thing that just put an 11% range on ASML.

It matters. It's a real structural shift and I'm not dismissing it. But it's a 2030 problem being priced as a 2026 one.

So my conviction hasn't moved. My opinion, not advice, and I'm the guy who's down today, so weigh it accordingly.

Full breakdown of my moves tonight.

[FromRedToRiches @FromRedToRiches]

J'étais dessus, mais j'avais mal joué l'entrée : Acheté 100 SAP à 204.7€ le 5 janvier, vendu a 152,96€ le 25 mai. -25%. j'ai laché tout mon logiciel ce jour là. Capgemini et Wolters Kluwer avec. Ce que j'ai bien fait c'est de sortir, le titre est passé sous 130 en juillet. Je n'ai pas encore repris de position et vous? Vous le prenez ce rebond ou vous attendez une confirmation?

[FromRedToRiches @FromRedToRiches]

The threat is real, no argument there. Attackers already have frontier AI.
But read the list. Almost everyone on it sells something the standard will end up blessing. And half the software names there are companies AI could disrupt. A seat at the table where the rules get written is cheap insurance. I'm not neutral either. I own 1000 shares of $NVDA, about 8% of my portfolio. Nvidia pushing open weights isn't charity. A wide ecosystem of models buys more GPUs than three closed labs with pricing power. That trade works for me.
What I'd want independent is the audit. Engineers with nothing to sell. Who funds them, I don't know (Mb Linux?). That's the weak part of my own argument.

[Jensen Huang @JensenHuang]

Attackers have frontier AI. Defenders need a frontier AI ecosystem—the best open and closed models, force-multiplied by a global community.

During the Hugging Face incident, closed AI blocked essential forensics. An open-weight frontier model helped contain the intrusion.

That’s why we created the Open Secure AI Alliance.

[FromRedToRiches @FromRedToRiches]

Same shock this morning. For me the real question is not how they close the gap without EUV. it's how they got this far without it. Older ASML DUV machines bought before the restrictions, and brute force. Apple or not, the new US rules on CXMT protect $MU, $SKHY and Samsung where it matters, in the US market. All three are building fabs there, and those investments are note innocent timing. They are an answer to $CXMT. The lead is still on their side. I own 200 shares of Micron, one of my biggest positions, plus SK Hynix where i'm still down 13%.

[FromRedToRiches @FromRedToRiches]

The bubble crowd will be right one day. Being five years early costs you about as much as being wrong.
I bought into Friday's Sox -4.4% $INTC, $AMKR and i added to $AMAT and $ONTO. It was not comfortable. Everything i own was red and i had no idea if monday would be worse. What i keep coming back to is the equipment order books. Signed quarters ahead. Two red days don't touch them !

[FromRedToRiches @FromRedToRiches]

Worth noting where that backlog comes from: it's mostly Cloud and mostly signed before the capex it wiill require gets spent.
Alphabet is 4% of my book and one of the few AI names i hold without a semis correlation. The number i'll be watching is how much of this converts in 2027, not the headline.

[FromRedToRiches @FromRedToRiches]

On May 18 last year I posted +200 replies in one day.

I have one mode: all in. I'd come back to X after a long break and decided the way to get noticed was to reply to everyone, everywhere. I thought I was doing the right thing. Showing up, being present, talking to people.

My reach collapsed instead. I stopped posting for thirteen months.

Came back this month with better replies and the same instinct: too much, too fast. Reach dropped again, from one day to the next. And nothing tells you what changed, or when, or for how long.

So I'm doing the boring thing now. One portfolio update a day, real numbers, real decisions, a few conversations worth having.

If this ever happened to you: did your reach come back on its own ?

[FromRedToRiches @FromRedToRiches]

@GregInvestFr Oui ca devient clairement interessant a surveiller, au moins pour profiter d’un rebond court terme.

[FromRedToRiches @FromRedToRiches]

@ApprendEnsemble Pour moi c'était $Intel hier soir, dans la chute.
Je pense que le SOX va rebondir rapidement et qu'Intel revenir tester ses plus hauts.

[FromRedToRiches @FromRedToRiches]

Full blotter — Jul 24, 11 orders

SELLS
- 206 $SNPS @ $373.65 → exited
- 350 $AMZN @ $232.45 avg (2 fills) → exited
- 118 $CEG @ $277.60 → 3.4%

BUYS
- 365 $INTC @ $92.48 → 1.4%
- 970 $OUST @ $35.65 → 1.4%
- 500 $AMKR @ $64.60 → 1.3%
- 315 $FORM @ $107.21 → 1.3%
- 100 $ONTO @ $269.48 → 4.5%
- 35 $APP @ $389.40 → 2.3%
- 30 $AMAT @ $530.75 → 5.1%

Best: $NOW +5.9%, $ZS +2.1%
Worst: SK Hynix −9.8%, $CRDO −9.7%, $COHR −9.7%

All moves computed in EUR on prior-day close, FX included.
Portfolio: €2,138,037 (−€77,773).
Book stays ~76% semis + AI infra.
No cash rebuilt today : this was a rotation, not a de-risking.

[FromRedToRiches @FromRedToRiches]

Daily recap — Jul 24

Portfolio −3.5% vs SOX −4.4%.
+0.9 pt vs my benchmark on a brutal tape.

SOX just re-entered bear-market territory.
I don't think we're in one.. and I'm positioned for a violent rebound.

VIX closed at 18.5, DOWN 0.9%. DAX +1.4%, CAC +0.9%, Dow +0.5%, S&P 500 +0.1%.
Semis got hit. Nothing else did.
That's a rotation, not a systemic de-rating.

I'm wrong the day the VIX breaks 25 and the S&P starts following the SOX down.

So I sold what held up to fund immediate buys on what got hit hardest.
Out: $SNPS, $AMZN · Trimmed: $CEG
In: $INTC, $OUST, $AMKR, $FORM
Added: $ONTO, $APP, $AMAT

11 orders — full blotter in reply ↓

[FromRedToRiches @FromRedToRiches]

The signature list is the real story here. The signers sell what surrounds the models: chips, cloud, infrastructure. OpenAI and Anthropic sell the models, and they didn't sign. Both sides are rational.
Open weights push value toward hardware and applications. Good for the SOXX, good for me. I own 1,000 shares of $NVDA, about 8% of my portfolio.
If you spend billions training frontier models, I get why you want them closed. But as a shareholder of the picks and shovels, I'll take this outcome.
Welcome to X.

[FromRedToRiches @FromRedToRiches]

Sold out: $AMZN
Full exit today. 350 shares at $232.45 avg, about $81K, 3.3% of the portfolio.
Earnings are next Thursday. I expect EPS and revenue in line. But Google raised its capex guidance this week and the stock dropped 5%. I see the same setup for Amazon.
I still think Amazon is undervalued. This is a timing call, not a valuation call. I'll probably buy back after the print.
The capital went to $OUST, $FORM and more $APP. Details in tonight's recap.

[FromRedToRiches @FromRedToRiches]

Sure, but look at what came before. Even after 7 red days, $CRWD is still up more than 60% this year. Keeps falling? I might take a starter position. But my first picks in this sector are the outsiders. $ZS is down about a third this year and got hit again this week. $RBRK sold off too. Beaten down names, none of the premium. That's where my next euros go if this weakness continues.

[FromRedToRiches @FromRedToRiches]

@elonmusk I try it now

[FromRedToRiches @FromRedToRiches]

Daily recap — Jul 23
Portfolio: +€589 (+0.03%)
vs SOX -0.5%, Nasdaq -2.2%
Best: $GEV +4.9%, $CRDO +3.9%, $MU +3.6%
Worst: $GOOGL -6.9%, $NOW -6.3%, $AMZN -4.3%

Trades today, 6 orders:
• Trimmed 40 $AMAT @ $565. The position had grown too big for a tactical bet, freed some cash. Still 4.5% of the portfolio.
• Trimmed 20 $MU @ $1,002. Same logic, but into strength. Still my #2 at 7.9%.
• Sold 78 $VST @ $167. Profit taking, up 11% on my avg. Still 4.2%.
• Added 415 $NOW @ $94.71. Strong quarter, badly paid. Revenue up 24%, AI ACV past $1B, guidance raised. The market only saw the GAAP miss and a soft backlog. I took the other side. It closed lower, we'll see who's right. Now 2.9%.
• Added 36 $APP @ $399. Too cheap versus fundamentals at this price. Playing the rebound, not a marriage. Now 1.7%.

Nasdaq down more than 2%, my book closed flat. Semis and power carried what the mega-caps bled. Same rotation as Monday.

One more thing: $INTC just reported, up 12% after hours. I don't own it, but a bid like that lifts the whole semi complex. Tomorrow could be interesting.

[FromRedToRiches @FromRedToRiches]

Le levier des particuliers coréens est un vrai signal de court terme, d'accord. Et j'ai encore allégé Micron à 1 002 $ aujourd'hui. Mais je garde le cœur de position. À 5 ans de bénéfices forward pour racheter toute sa capitalisation (4 pour SK Hynix, 37 pour AMD, 86 pour ARM), la mémoire n'est pas pricée comme une bulle, elle est encore pricée comme le cycle d'antan. Les contrats HBM pluriannuels ont relevé le plancher. La rechute décrite peut arriver, mais à ces multiples elle ressemblerait plus à un cadeau qu'à un piège.

[FromRedToRiches @FromRedToRiches]

@TradexWhisperer The market still prices memory like a commodity cycle, treating peak earnings as fake. HBM multi-year contracts moved the floor. Cycle risk isn't dead, it's mispriced. When that clicks, these multiples rerate fast :) Long $MU (7.9%) and $SKHY (3.8%) here.