Great analysis, and I believe it adds color to management’s decision to more than double Purchase Commitments and Other Contractual Obligations from $332.4bn to $811.0bn in a single quarter. Anybody whose thesis depends on the durability of AI capex should take note.
Thank you, very interesting. I made a similar exercise taking SpaceX Colossus 1 as example. Using neocloud on-demand prices I got decent ROIC 25-45% depending on the capex assumptions. I agree with you that Alphabet is in a strong position even if token prices will decline. They have structural cost advantage (TPUs, cost of capital, scale) and funding capacity to outcompete pretty much everyone.
The $123 billion sitting in not-yet-in-service assets is the number worth tracking going forward, since that’s the capacity due to come online by mid-2027, and if even a fraction of that converts at similar returns, current Street estimates for operating income growth look conservative.
Great analysis, and I believe it adds color to management’s decision to more than double Purchase Commitments and Other Contractual Obligations from $332.4bn to $811.0bn in a single quarter. Anybody whose thesis depends on the durability of AI capex should take note.
Thanks Melvin!
Thank you, very interesting. I made a similar exercise taking SpaceX Colossus 1 as example. Using neocloud on-demand prices I got decent ROIC 25-45% depending on the capex assumptions. I agree with you that Alphabet is in a strong position even if token prices will decline. They have structural cost advantage (TPUs, cost of capital, scale) and funding capacity to outcompete pretty much everyone.
Thanks! Glad you enjoyed the piece!
The $123 billion sitting in not-yet-in-service assets is the number worth tracking going forward, since that’s the capacity due to come online by mid-2027, and if even a fraction of that converts at similar returns, current Street estimates for operating income growth look conservative.