Their combined free cash flow turned negative in the September quarter, down from $51.4 billion a year earlier. Their long-term debt has tripled in a year, and Microsoft has just sold $25 billion more.
The number that stuck with me is $329.3 billion of long-term debt against $107.7 billion a year ago. The build is no longer funded by the cash machine, it is funded by the bond market, which makes the marginal investor in AI capex a lender rather than a shareholder. Your Microsoft 2066 watch is the right instrument for this: the 1.05 point spread held at issue, and if it widens, the equity story has to carry weight it has not carried in years. One question: with AWS at 39% and guiding above 40% through 2028, is there a growth rate at which the negative free cash flow stops mattering, or does the debt number keep the question open no matter what the top line does?
Growth alone won't settle it. What matters is whether each new dollar of capacity earns more than it costs once depreciation and interest are counted. Amazon's record 13.9% margin is the line to watch as depreciation from this build rises. If AWS keeps growing near 40% and that margin holds, the negative cash flow looks like an investment phase. If the margin slips, the debt keeps the question open, however fast revenue grows.
That's the right reframe, and it moves my question from the revenue line to the margin line. If AWS keeps growing near 40 percent with the 13.9 percent margin intact, the negative free cash flow is just an investment phase and the debt is noise. What I'll be watching is the depreciation line creeping up as the new capacity comes on, and whether that 1.05 point spread on the Microsoft 2066 starts to move. If the margin slips and the spread widens together, the debt question stays open no matter what the top line does.
Molto bello, ottimo articolo chiaro. Sembra che abbiano cominciato un ciclo di investimenti che durerà anni, non pochi trimestri o solo uno o due anni. Le spese non sono basse.
Yes, and the filings put dates on it. Most of Alphabet's $924bn of purchase commitments run through 2031, and Microsoft's unstarted leases begin as late as fiscal 2034. Lenders are pricing it that way too: Microsoft just sold bonds that don't mature until 2066. The risk with a long cycle funded by debt is that the payments keep coming even if demand cools.
Io sto sviluppando una mia idea del settore e di come si potrà evolvere i prossimi anni, alla luce di queste informazioni e di nuove; sto creando un articolo o una serie che cercheranno di spiegare meglio questo ciclo di spese IA e di cosa potrebbero comportare in futuro per queste aziende, tracciando parallelismi con aziende del passato e i grandi cambiamenti infrastrutturali e tecnologici. Sarà un bel pezzo impegnativo, perciò ci vorrà tempo per cercare le fonti, organizzare le idee e scriverlo. Grazie delle tue info preziose che aggiungerò alle altre informazioni. ☺️
Looking forward to it. A good parallel is the fiber buildout of the late 1990s. It was largely funded with debt, and the cables stayed useful even after many of the companies that built them went under. This time, the builders have some of the strongest balance sheets in the market, so the risk is more about returns than survival.
The number that stuck with me is $329.3 billion of long-term debt against $107.7 billion a year ago. The build is no longer funded by the cash machine, it is funded by the bond market, which makes the marginal investor in AI capex a lender rather than a shareholder. Your Microsoft 2066 watch is the right instrument for this: the 1.05 point spread held at issue, and if it widens, the equity story has to carry weight it has not carried in years. One question: with AWS at 39% and guiding above 40% through 2028, is there a growth rate at which the negative free cash flow stops mattering, or does the debt number keep the question open no matter what the top line does?
Growth alone won't settle it. What matters is whether each new dollar of capacity earns more than it costs once depreciation and interest are counted. Amazon's record 13.9% margin is the line to watch as depreciation from this build rises. If AWS keeps growing near 40% and that margin holds, the negative cash flow looks like an investment phase. If the margin slips, the debt keeps the question open, however fast revenue grows.
That's the right reframe, and it moves my question from the revenue line to the margin line. If AWS keeps growing near 40 percent with the 13.9 percent margin intact, the negative free cash flow is just an investment phase and the debt is noise. What I'll be watching is the depreciation line creeping up as the new capacity comes on, and whether that 1.05 point spread on the Microsoft 2066 starts to move. If the margin slips and the spread widens together, the debt question stays open no matter what the top line does.
Molto bello, ottimo articolo chiaro. Sembra che abbiano cominciato un ciclo di investimenti che durerà anni, non pochi trimestri o solo uno o due anni. Le spese non sono basse.
Yes, and the filings put dates on it. Most of Alphabet's $924bn of purchase commitments run through 2031, and Microsoft's unstarted leases begin as late as fiscal 2034. Lenders are pricing it that way too: Microsoft just sold bonds that don't mature until 2066. The risk with a long cycle funded by debt is that the payments keep coming even if demand cools.
Io sto sviluppando una mia idea del settore e di come si potrà evolvere i prossimi anni, alla luce di queste informazioni e di nuove; sto creando un articolo o una serie che cercheranno di spiegare meglio questo ciclo di spese IA e di cosa potrebbero comportare in futuro per queste aziende, tracciando parallelismi con aziende del passato e i grandi cambiamenti infrastrutturali e tecnologici. Sarà un bel pezzo impegnativo, perciò ci vorrà tempo per cercare le fonti, organizzare le idee e scriverlo. Grazie delle tue info preziose che aggiungerò alle altre informazioni. ☺️
Looking forward to it. A good parallel is the fiber buildout of the late 1990s. It was largely funded with debt, and the cables stayed useful even after many of the companies that built them went under. This time, the builders have some of the strongest balance sheets in the market, so the risk is more about returns than survival.
Exactly that! I will deep dive into that and crunching the numbers to show a perspective on that!