RT Investing visuals
The AI data center bottlenecks map & key players:

Demand:
• $MSFT
• $GOOGL
• $META
• $AMZN
• Leading LLM providers

Layer 1 - Systems & colocation:
• $DELL
• $HPQ
• $SMCI
• $EQIX
• $DLR

Layer 2A - Memory:
• $MU
• $HYNSE
• $SMSD

Layer 2B - Networking & optics:
• $AVGO
• $COHR
• $MRVL

Layer 2C - Power & cooling:
• $VRT
• $ETN
• $BE

Layer 3 - In-rack connectivity:
• $AVGO
• $ALAB
• $CRDO

Layer 4 - Foundry & packaging:
• $INTC
• $ASX
• $TSM
• $AMCR

Layer 5 - Semiconductor equipment:
• $ASML
• $AMAT
• $LRCX
• $KLAC

Note that this isn't an exhaustive list, but rather an overview of the most important businesses per segment.

I hope you found this breakdown insightful!

I've covered the supply chain in more detail in the article linked below.

Investing visuals: I just published my latest research article: The AI data center bottlenecks map, covering:

• The AI supply chain
• The key players at critical bottlenecks
• Where to look as constraints move down the supply chain

I loved diving into this over the past few weeks with the goal

The hyperscalers aren’t cutting AI capex.

$660-700B committed for 2026. Contracts signed. Fabs running.

Here’s where that money flows:

Compute: $NVDA, $AMD, $AVGO
Memory: $MU, $SKX
Neoclouds: $IREN, $NBIS, $CIFR
Equipment: $AMAT, $LRCX, $KLAC
Power: $CEG, $VST, $ETN
Cooling: $VRT,
Networking: $ANET, $LITE, $AAOI

The Iran war slows sentiment, but it doesn’t slow AI buildouts.

Save this for later.

-BP

Please note, this is not financial advice.

🚨 Taiwan has ONLY ELEVEN days of natural gas reserves.

Taiwan runs 53% of its power grid on natural gas.

Qatar supplied 33% of that gas through Hormuz.

That route is now totally DISRUPTED!

Japan has 3 weeks. South Korea has a month.

Taiwan has eleven days, the LOWEST buffer in East Asia.

The government has secured supply through April.

May is being worked on. June is an open question.

And summer electricity demand in Taiwan is 40% higher than it is right now.

Here’s what this means for semiconductors:

Gas-fired plants generate more than 53% of Taiwan’s power. TSMC alone consumes roughly 8% of the entire national grid, rising toward 12.5% by year end. 

$TSM controls 70% of global chip foundry capacity and near-monopoly on the bleeding-edge nodes that power EVERY ai system on earth.

The hyperscalers committed $660-700B in AI capex for 2026.

Every dollar of that ASSUMES $TSM keeps the lights on.

Here’s where the real plays are:

1. $LNG (Cheniere Energy)

Taiwan is reorienting its entire energy supply chain toward the US. Cheniere is the direct beneficiary.

2. $AMAT $LRCX $KLAC

Every day this crisis continues, the case for accelerating US and Japanese fab buildouts gets stronger. Equipment orders are being front-loaded. Geographic diversification becoming existential.

3. $INTC + TSMC Arizona

Not better fabs. Not cheaper fabs. Fabs that aren’t on a grid with 11 days of gas reserves.

4. $NVDA

Any TSMC production constraint tightens GPU allocation. Scarcity pricing. That’s not bearish, that’s the opposite.

Watch for:

→ TSMC Q1 earnings, any energy commentary changes everything

→ Taiwan government statement on May/June LNG supply security

→ JKM Asia LNG benchmark price already doubled. Watch for further moves in April. 

→ Summer demand peak. Taiwan grid stress starts July

The financial shortage is here now.

The physical shortage risk builds toward summer.

The market hasn’t priced either one fully.

-BP

Please note, this is not financial advice.

RT KaizenInvestor
I have written a full article on the AI chip supply chain.

The supply chain is structured into 4 different phases with 13 layers:

1. Raw Materials: $SHECY, $SUOPY, GlobalWafers, $WAF.DE, $SHWDF, $AXTI, $IQE
2. Manufacturing Equipment: $ASML, $ASM.AS, $AMAT, $LRCX, $KLAC
3. EDA & Core Intellectual Property: $SNPS, $CDNS, $ARM, $RMBS
4. Chip Design: $NVDA, $AMD, $INTC, $QCOM
5. Foundries: $TSM, Samsung Semiconductor, $SMIC
6. Memory and HBM: SK Hynix, Samsung Electronics, $MU
7. Packaging and OSAT: ASE Technology, $AMKR, JCET Group
8. Server and Rack Integration: $SMCI, $DELL, $HPE, Foxconn
9. Networking Silicon: $AVGO, $MRVL, $CSCO, $ANET
10. Photonics and Optical Components: Ayar Labs, $ALAB, $CRDO, $COHR, $LITE
11. Power, Thermal management and Grid: $VRT, $MOD, $NVT, $SU.PA, $IREN, $CIFR
12. Hyperscalers: $AMZN, $GOOGL, $MSFT, $META
13. AI Storage, platforms and Data: VAST data, Weka, NetAPP, $PLTR, Blue Yonder, $KXSCF

The article covers it all.

KaizenInvestor: http://x.com/i/article/2029218092270628864

RT TacticzHazel
Which AI company would you buy right now?

Chip design:
- $AMD
- $NVDA
- $ARM
- $INTC

Manufacturing:
- $TSM
- $INTC
- $SAMSUNG

HBM/Memory:
- $MU
- SK HYNIX

Equipment:
- $ASML
- $KLAC
- $LRCX

Power:
- $VST
- $CEG

Networking:
- $ANET
- $MRVL