Most people diversify and stay average.
I run a concentrated AI-focused portfolio (March 2026) π
Top winners:
π£ππ§π₯ +π±π¬π¬%
π₯πππ +ππ³π¬%
ππ’ππ‘ +ππ²π΅%
Hereβs exactly how Iβm positioned π
Really interesting take on MU - the supply shortage angle makes sense given what Mehrotra said. But Iβm curious, with fabs taking 3β5 years to come online, at what point does all that new capacity start flooding the market and turn this from a tailwind into a headwind for the stock?
@SWEETSANDY007@dannycheng2022 Bro, that AMD entry is crazy, congrats on the 222% gains. What else is in your portfolio right now? Are all your plays hitting like this, or do you have a couple bad ones balancing it out?
$NOW CEO Bill McDermott just laid out the 2030 plan positioning them as control tower for AI governance, routing & oversight:
β’ $30B+ 2030 subs revenue (~18% CAGR)
β’ Rule of 60 target & highest renewal rate in SaaS
β’ $600B TAM (+100% YoY) supported by 6 $1B+ business lines
BREAKING: Initial details are emerging as the US and Iran conduct their first direct meeting since 1979.
Details include:
1. The Strat of Hormuz remains a point of "serious disagreement"
2. US military says 2 US warships have transited the Strait of Hormuz today
3. "Mood swings" are being seen on both sides in talks, per Pakistani mediators
4. The US will be assisting Iran in removing mines from the Strait of Hormuz
5. Talks are expected to continue tonight and may extend into tomorrow
We expect to receive much more detail in the coming hours.
Been watching $AMZN closely β Anthropic choosing Trainium for their most capable (and most restricted) model feels bigger than a chip story.
Itβs validation.
AWS is quietly becoming the place where frontier models get trained β especially the ones you canβt just ship publicly.
Thatβs a different game than trying to beat Nvidia on raw benchmarks.
Trainium doesnβt need to win β just needs to be good enough for the models that actually matter.
Been watching $USO β WSJ leaking a target list publicly isnβt an attack signal, itβs a negotiation tactic.
The TACO trade has paid out every single time. But it only has to be wrong once to wipe out every gain from playing it.
That asymmetry is the only thing worth pricing in right now.
Been watching oil - markets arenβt rallying on a ceasefire, theyβre reacting to both sides still talking.
A 10-point proposal while rejecting terms isnβt a rejectionβ¦ itβs an opening bid.
If Trumpβs already calling it βsignificant,β is the war premium coming out before any deal is close?
Robinhood going from meme stock villain to government-backed brokerage for Americaβs kids in 3 years is one of the biggest brand rehabilitations in fintech history.
BNY handles custody, $HOOD gets the relationship β and every one of those kids becomes a potential lifetime customer.
Which fintech actually loses the most if Robinhood locks in the next generation of retail investors?
$1.5B buyback from a company still growing ARR at 20%+ β thatβs not capital return, thatβs management saying the stock is on sale.
Most companies buy back shares when growth slows. CrowdStrike is doing it while accelerating.
So what does management know that the market is still missing?
$50M+ sounds like a step up β but itβs really the signal that ONDS is moving from small wins to scalable defense programs.
Early contracts validate the tech, but they donβt guarantee repeat or expansion.
How many of these initial wins actually turn into long-term program revenue?