Tech didn't win last decade as a sector.
A handful of names inside tech did.
$QQQ returned 20.3% a year from 2016- 2026. The S&P managed 14.7%.
That gap wasn't diversification. It was concentration.
The same pattern is forming in AI.
Opportunity isn't spread across 500 stocks. It’s concentrated in a few bottlenecks.
Currently in:
$IREN - power + AI compute
$MU - memory
Watching to build:
$CRDO - connectivity
$VRT - power / cooling
$INTC - CPU
Which bottleneck presents the greatest opportunity right now?
3
2
570
To be Successful in the Markets You Need to Make Few Moves Not Many.
In the past 18 months I have made very few major moves:
$IREN | $15 → $44 | +193%
$NBIS | $36 → $295 | +719%
$VRT | $175 → $300 | +71%
$MU | $340 → $930 | +174%
$IREN and $MU still held. $NBIS trimmed at $295 and exited at $280. $VRT exited at $300.
4
3
33
3,360
5,000 followers.
When I started this account I had no idea it would grow into this.
Thank you to everyone who follows, interacts and challenges my thinking.
The account has evolved. The thesis has deepened. The community has grown.
Honest analysis, transparency and my thoughts on the future. That’s what you’ll continue to get.
Next target: 10,000 by EOY. 🙏
21
35
1,573
Jacob Coxon Wrote a Viral Thread on X
He stated that AI could kill us all by the end of the decade.
Turns out he was financially motivated to post.
Is this the part we pretend to be Shocked?
JUST IN: Investigation reveals former Anthropic researcher Jacob Coxon reportedly worked with a “doomer PR firm” while launching his viral campaign warning of AI’s existential risks.
4
839
I publicly called $NBIS in the $40s and $IREN at $18.
When this post was written $IREN was $50-60s and $NBIS was $90s.
I specifically called $NBIS the better buy at that moment, not $IREN. Despite holding both as my two largest positions.
I held $NBIS from $30 to $300. Trimmed publicly at $295 and again between $258-280.
$NBIS was 30% of my portfolio through that entire run. $IREN held since $15 , called it a buy every time it hit the $30s.
Through that entire period I trimmed $NBIS twice only.
Track records matter. Mine has been public.
$IREN now represents over 60% of my current portfolio and is the better buy.
If you aren't following already, you should be.
15
3
148
20,883
Dylan calls it fake news. Two things he doesn't address:
1. He affirms $NBIS is a SemiAnalysis customer. A paid commercial engagement, commissioned TCO studies and joint marketing exists between them.
2. Kent Draper confirmed $IREN does not currently offer managed services. ClusterMAX 3.0 specifically evaluates managed GPU clusters.
How was $IREN rated on a criteria they don't yet offer?
Fake news.
We’ve done work with dozens of neoclouds and their equity and credit holders.
We do not pay to play.
Google Microsoft Amazon Oracle are all larger customers of SemiAnalysis than Nebius and Coreweave.
How about you actually look at the technical side of it
14
7
133
24,648
Why do I get comments about $NBIS when I post about $IREN ?
These comments are surface level and from accounts that recently invested in $NBIS .
Meanwhile I was buying $NBIS in the 30’s, being vocal on X in the $60’s and screaming it’s a buy between $90-110.
If you are going to come saying “ $NBIS is the better company” , be ready to explain why.
8
1
50
3,954
$NBIS commissioned SemiAnalysis to model total cost of ownership across GPU workloads in March 2026.
The study, paid for by $NBIS concluded Nebius has the lowest TCO in all scenarios.
Six months later SemiAnalysis places $NBIS in the Platinum tier and rates $IREN as Underperforming in ClusterMAX 3.0.
The same firm. A paid client relationship on one side.
Draw your own conclusions.
38
16
234
51,795
To be Successful in the Markets You Need to Make Few Moves Not Many.
In the past 18 months I have made very few major moves:
$IREN | $15 → $44 | +193%
$NBIS | $36 → $295 | +719%
$VRT | $175 → $300 | +71%
$MU | $340 → $930 | +174%
$IREN and $MU still held. $NBIS trimmed at $295 and exited at $280. $VRT exited at $300.
4
3
33
3,360