Publikuj
MilkRoadAI
$3.2 trillion is now locked into AI commitments and that changes everything about this buildout (Save this).
Across Amazon, Google, Meta, Microsoft, and Oracle, that total splits between equipment purchases, current leases, future lease obligations, and pending stakes in AI companies.
Google's total jumped to $946 billion, up a stunning 116% from the prior quarter, making it the largest single commitment pile in this entire group.
Meta comes in at $731 billion, up 59% quarter over quarter, followed by Microsoft at $672 billion, up 50%, then Amazon at $453 billion, up 13%, and Oracle rounding things out at $349 billion, up 16%.
Most of that money isn't a one-time purchase you can cancel next quarter, it's future lease obligations locked in years ahead of time, and that's a fundamentally stickier commitment than ordinary capital spending.
Meta's future lease commitments alone jumped 90% quarter over quarter to $347 billion, and Microsoft's future leases grew 67% to $329 billion, while Oracle's future lease book more than doubled from where it started, sitting at $260 billion.
Google's equipment purchase commitments spiked 144% to $811 billion, by far the largest jump of any single category across any company, showing just how aggressively it's been signing purchase orders it can't easily unwind.
The reason this framing matters is that a normal capex cycle can be slowed or paused whenever a company decides demand has cooled off.
A multi-year lease or a signed equipment purchase order doesn't work that way, since these are contractual obligations the companies have to honor regardless of what happens to AI demand next year or the year after, which makes this buildout far stickier than a typical spending cycle.
That's the real reason hyperscalers are locking in power, compute capacity, and physical infrastructure years ahead of when they'll actually need it, since waiting to sign these contracts until demand shows up would mean losing access to capacity entirely in a market where everyone else is racing to secure the same scarce resources.
$3.2 trillion of AI commitments has to flow somewhere, and Milk Road subscribers are already up massively on several of the trades positioned around this buildout.
If you want to see exactly what we’re buying as that money flows into power, data centers, chips, memory, and networking, come join us for just $1 using the link below.
Zastrzeżenie: Treść na OKX Orbiter ma charakter wyłącznie informacyjny. Dowiedz się więcej
Odpowiedzi
Brak komentarzy. Bądź pierwszą osobą, która odpowie!
Dzisiejsze trendy rynkowe
1#NVIDIA kontaktuje się z CEO Arm, Hassem, wewnętrzne prognozy dotyczące nowego kandydata na CEO
2#Amerykański PPI za październik wzrósł o 6,5% rok do roku, co jest poniżej oczekiwań
3#Anthropic rozpoczyna roadshow IPO, wycena wzrasta do 1,8 biliona dolarów
