Are Microsoft and Meta being better allies to chip companies than their own shareholders?
It’s time to turn to the statement of cash flows.
Microsoft and Meta, the two hyperscalers that reported earnings on Wednesday, are experiencing wildly divergent fortunes: the former is slumping severely after its cloud business failed to impress, while the latter is up after CEO Mark Zuckerberg doubled down on his commitment to the AI boom.
You can’t really complain too much about the returns you’ve gotten from these companies since the AI boom kicked off. Well, maybe Microsoft leaves something to be desired.
But to zoom out and understand how dramatically investments in this new technology are impacting these companies’ financials, we need to move away from some of the most high-profile quarterly numbers that determine whether analysts say, “Great quarter, folks!” or not.
Line items like net income and earnings per share can’t provide the full picture. For those, only depreciation (i.e. the capital stock you’ve “used up” over the course of the quarter to make all that money) is included.
It’s time to take a page out of the Wu-Tang Clan’s book: C.R.E.A.M. — cash rules everything around me — and turn to the statement of cash flows.
The cash flows that these companies generate have been under pressure amid their AI spending binges, even as earnings per share rapidly expand.
“Free cash flow was $6.5 billion, down 29% year-over-year reflecting higher capital expenditures to support our cloud and AI offerings,” according to Microsoft’s earnings call slides. Meta, for its part, saw this metric fall about 18% quarter on quarter.
Microsoft’s total AI revenues in Q4 were in the neighborhood of $3.25 billion, versus capex approaching $16 billion. Of course, the returns from investment presumably accrue over time, though this does hint at the thorny question of how much in ongoing capital outlays will be required to stay competitive in this space over time.
If you’re spending billions on AI and not making as much cash, that can lead to some belt-tightening in other areas. Of note: how much cash these companies are giving to shareholders in the form of buybacks.
Meta made a grand total of $0 in buybacks in the fourth quarter (or, per its press release, “nil”).
And despite its buybacks, Microsoft’s shares outstanding have increased since the end of July 2023, and share repurchases have shrunk over that time frame.
Implicit in all this AI spend is that down the road, shareholders get their due. That’s cold financial logic that’s presumably underpinning these outlays and the significant rallies in their share prices.
One could wonder how consistent this thesis is with a stylized version of the Jevons Paradox argument (in short: MOAR COMPUTE!), but let’s leave that aside for a moment.
Is there a step function for returns on investment in the offing? Perhaps artificial general intelligence, a vaguely and oft differently defined term, will provide that sort of jump in ROI that allows for something of a payback period in the cash flow statement and shareholder returns.
On the “Odd Lots” podcast, Zvi Mowshowitz said, “Generally, it is understood to mean you can do any task that can be done on a computer, that can be done cognitively only, as well as a human.”
Industry experts (particularly AI-adjacent tech executives) appear optimistic on how soon we’ll get a breakthrough there.
Failing that, more downward pressure on free cash flow and buybacks going forward is likely to lead to more pressure on management teams to justify their expenditures.
Especially when competitors are seemingly doing more (or as much) with less.