Mag 4 Capex & Cloud Recap: $725B CapEx, Who’s Going to Secure the Bag?
Markets rally prompted by good earnings. Big Tech took turns proving the bull case, recovering March's tariff-driven selloff.
How's everything going so far?
-
$Alphabet(GOOG)$ surged +10% in a single session after Cloud revenue grew +63.4%, killing the "Google is losing the AI race" narrative.
-
$Apple(AAPL)$ +2.56% post-earnings on a record March quarter.
-
$Amazon.com(AMZN)$ posted $23.9B in operating income, a 14% beat.
-
$Meta Platforms, Inc.(META)$ delivered +28.7% ad revenue growth but lost 9% due to capex concerns.
-
$Microsoft(MSFT)$ is worse, still the worst performer among mag 7. Its capex slows down.
Another company affected by the mag earnings is$$NVD$$ -5%, falling back below the $5T market cap.
A new high-efficiency model release that the market read as "better models = less compute demand." But $725B in committed hyperscaler CapEx is already locked, B300 servers pricing near $1M, and supply tightness hasn't changed.
Let's take a look at the most important parts: capex and cloud.
💰 CapEx Summary
Cloud Revenue Comparison: Constrained by supply — not by demand.
Google Cloud's acceleration was the biggest surprise of the night: +63.4% from +48% last quarter, Cloud op margin cracking 30% for the first time.
Management's exact words: "If we had more compute, cloud revenue would have been higher."
——————-
Jefferies analyst Brent Thill: "We're seeing bottlenecks across the board" — memory, fiber, power, cooling water, undeveloped land.
Every layer in the AI infrastructure stack is supply-constrained and repricing. $725B of committed spend means the picks-and-shovels trade just got a hard floor under it.
$725B in committed CapEx. Who actually captures it?
How do you view the sky-high capex?
Leave your comments to join our Mag7 series to win at least 5 tiger coins~
Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.
On capex, I don’t see a bubble — I see barriers forming. Despite concerns around $Meta Platforms, Inc.(META)$ and $Microsoft(MSFT)$ , the key takeaway is unchanged: demand exceeds supply, and constraints are real, not cyclical excess. To me, this looks like early-stage infrastructure buildout rather than late-cycle overinvestment.
For the $725B spend, I still favor the infrastructure layer like NVIDIA. Efficiency gains won’t kill demand — they expand it. This looks like a narrative-driven pullback, not a structural shift.
@TigerStars @Tiger_comments @TigerClub
The Silicon Kings : $NVIDIA(NVDA)$ $Taiwan Semiconductor Manufacturing(TSM)$ $Broadcom(AVGO)$ are the 3 winners. Every dollar Big Tech spends goes to these 3 companies.
Nvidia is no longer just selling chips. It is selling AI factories. At March 2026 GTC, CEO Jensen Huang unveiled the Vera Rubin platform. This isn't just a GPU. It is a vertically integrated system of 7 new chips designed to act as a single supercomputer.
Broadcom: It is the King of connectivity & custom silicon.
As AI clusters scale to millions of chips, the bottleneck is communication. Broadcom Tomahawk 6 switch holds an estimated 80% market share in high speed Ethernet.
Broadcom co develops custom AI chips XPU for Meta & Google's TPU.
TSMC is the Foundry to the world. It is mass producing 2nm chips in early 2026.
These are the Big 3 companies getting Big Tech money.
@Tiger_comments @Tiger_SG @TigerStars
Amazon (AMZN): AWS grew ~28% (fastest in 15 quarters) and delivered a strong operating income beat. Solid returns on ~$200B capex.
Meta (META): Strong ad revenue (+28-33%), but shares fell on higher capex guidance ($125-145B) as AI monetization remains more indirect.
Microsoft (MSFT): Azure growing well (~40%), yet the stock has lagged peers amid massive ~$190B spending and questions on return timing.
Apple continues with low capex intensity and strong services margins.Bottom Line: Among the big AI spenders, Alphabet currently shows the best bang for the buck
do i use ai... as a voice activated calculator thats it.
now im just a hood queen budget $10pw trader because never keep your eggs in one basket.
customers vote with their feet. working class chatter want to pop the ai bubble allready and get back to basics.
im the mean time $Southern Copper Corp(SCCO)$ shares did realy well with all the push ai made for copper. moral of the story us hood level traders need to invest in the industries that feed material to the big cat industries. big cats invest in big tech big tech buys raw materials. often a smaller easier market to get imto when your starting out are the raw material supliers.
fair trade winds fill your sails my fellow hood traders
I would be taking profits at every juncture then buying back in in hopes the bubble continues.
But I will gradually start picking up HALO shares to start hedging for the eventual bubble pop.
I don't see a need to be the one holding the bag when shit happens...
Thank you @koolgal for invitation 😘
thank you @Tiger_comments for the event. @TigerClub @TigerStars @CaptainTiger @MillionaireTiger
Join in the fun @Terra_Incognita @Emotional Investor @GoodLife99 @Barcode @vodkalime @bigfatdog123dog @ahyi @DCamel
1. NVDA, the compute toll booth. If GPU demand stays supply-constrained, pricing power remains exceptional.
2. MU / HBM memory suppliers, because memory is now mission-critical, not a commodity add-on.
3. VRT and cooling/power infrastructure names, the overlooked backbone of AI scaling.
4. Fibre/networking plays such as ANET, where bandwidth becomes as valuable as compute.
5. Select utilities, land and data centre REITs, where scarcity can drive repricing.
On sky-high capex: bullish near term, but execution risk rises sharply.
If AI monetisation lags infrastructure build, markets may start questioning ROI. Until then, the shovel sellers are still in the strongest position.
Management's exact words: "If we had more compute, cloud revenue would have been higher."