CFA. Private Equity professional. Investing in all asset classes including real estate,stocks,cryptocurrency,start-ups and private equity funds across the world
In a few months profits will temporarily flow out of TradFi (cool off phase after IPOs).
People will want to know where to put their money because everything already went up so much.
And there Bitcoin will be, ready to begin the next four year cycle.
The AI-Revolution is heavily compute constrained.
Hyperscalers are paying billions of dollars to accalerate the buildout.
These are the best positioned businesses across the AI-Infrastructure buildout stack.
Layer 1: Data centers
- $CRWV
- $IREN
- $ORCL
Top pick: $NBIS
Layer 2: Connectivity
- $AOI
- $ALAB
- $LITE
Top pick: $CRDO
Layer 3: Energy
- $OKLO
- $FLNC
- $EOSE
Top pick: $BE
Layer 4: Supply Bottlenecks
- $AXTI
- $IQE
- $COHR
Top pick: $WOLF
Layer 5: Thermal Management
- $NVT
- $MOD
- $PH
Top pick: $VRT
What are your top picks?
$NBIS market cap: $70B
$CRWV market cap: $68B
Meanwhile:
• Expected end-2026 revenue run rate
$CRWV: $18-19B
$NBIS: $7-9B
• Backlog
$CRWV: $99.4B
$NBIS: ~$45B
• Expected connected power by end of 2026
$CRWV: ~1.7 GW
$NBIS: 800MW-1GW
Both have deep partnerships with $NVDA.
$NVDA owns:
• 11.5% of $CRWV
• 8.3% of $NBIS
1Y return:
$NBIS: +660%
$CRWV: +5%
There is no reason why $NBIS should trade at a higher market cap than $CRWV today.
Buy the laggard.
$MSFT unveiled its next-generation quantum chip called Majorana 2 and says it is targeting useful quantum machines by 2029.
Microsoft says AI-designed materials helped unlock a major performance jump.
IBM $IBM is planning to invest over $10 billion in the next five years in an effort to build the first large-scale fault-tolerant quantum computer by 2029.
$QCOM CEO Cristiano Amon says agentic AI will require “gazillions” of tokens with token generation expected to rise from 32B every 10 seconds today to 1.3T by 2030.
That 40x increase shows why inference, edge compute and AI efficiency are becoming the next major bottlenecks.
$GOOGL didn’t raise $80 billion because they needed a lifeline.
They raised it because customer demand for AI compute is outpacing what they can physically supply. Power, land, data centers they can’t scale fast enough. Sundar Pichai literally said compute capacity is what keeps him up at night.
In April they revised capex guidance to $180-190B for the year. Now they’re going back to markets for more. Alphabet, Microsoft, Meta, and Amazon combined are expected to spend over $700B on capex this year alone.
And we’re still in the early innings. $1 trillion in AI capex by 2027 isn’t a prediction anymore it’s a roadmap.
$ARM Arm CEO says the company could reach its $15B AI chip revenue target earlier than expected, driven by surging AI demand and rapid growth in data centers. Meta is expected to be the first major customer for Arm's upcoming AI-focused CPU chips 🤖 🚀
$AMD is trading at almost double the $NVDA valuation despite $NVDA having faster expected growth by 2028.
Consensus growth estimates by 2028:
- $AMD: 43.7% CAGR
- $NVDA: 61.3% CAGR
Yet, $AMD is trading at 31x 2028 earnings while $NVDA is at 17.6x.
One of them is mispriced...