“Clown-1-1” (CC BY-SA 2.0) by americanbulldogbully007
Oracle announced Q3 earnings yesterday, and talked up their CAPEX spending. This post pulls together for posterity some of my Twitter reactions and provides another exciting installment of CAPEX Clown Watch.
Oracle spent $1.85 billion on CAPEX over the previous twelve months (by obscuring the quarterly number, they must hope you’ll compare their annual number to others’ quarterly numbers), up almost 20% over the previous year. That annual spend is about what Amazon spent in a week on CAPEX in the last quarter of 2020.
Founder, Chairman and CTO Larry “Darth” Ellison: “We are opening new regions as fast as we can to support our rapidly growing multi-billion dollar infrastructure business.”
CEO Safra Katz: “CapEx for Q4 is expected to be $1 billion.”
Zoom Boom and Gloom
The Zoom episode is a stark illustration that BS is a very poor substitute for CAPEX. After winning some business from Zoom almost a year ago, Oracle marketed the deal very aggressively. In time-honored Oracle fashion they exaggerated their role but, in particular, used it to hype the capacity of their cloud.
How It Started
Oracle press release (April 2020):
“We recently experienced the most significant growth our business has ever seen, requiring massive increases in our service capacity. We explored multiple platforms, and Oracle Cloud Infrastructure was instrumental in helping us quickly scale our capacity and meet the needs of our new users,” said Zoom CEO Eric S. Yuan.
An important clarification (May 2020):
“Reached for comment about Oracle’s deal, an Amazon spokesperson referred to a recent video call with Zoom CEO Eric Yuan’s where he said that the company “primarily” relies on Amazon.”
How It’s Going
Larry Ellison (February 2021):
But as time went on, Oracle ultimately didn’t anticipate just how significant the “incredible upward spike” and “hyper-growth” would be, and it wasn’t able to completely meet that additional capacity, he said.
“We got part of it right, but it was so extreme,” Ellison said at the event, that Oracle was unable to meet the capacity demands.
“That was just a complete change in demand in our business,” Ellison said. “We never thought it would go as high as it really did.”
“Hyper-growth” in Oracle terms is evidently just shy of 20% a year (which is, admittedly, almost an order of magnitude greater than their top line growth, so these things are all relative…).
Priorities, Priorities…
The guidance for a billion dollars of CAPEX in Q4 is a big step up for Oracle, whose spend has typically been less than half that in any quarter. But their growing CAPEX enthusiasm pales in comparison to new financial engineering commitments. The company raised its dividend by 33% and committed an “additional $20 billion for the repurchase of Oracle shares.” Needless to say, that is money that could be spent on CAPEX (and ensuring there is a dividend beyond the short run…).
Meanwhile, in the Cloud
Lets see how Oracle’s CAPEX stacks up with the hyperclouds.
Once again “That red line you may mistake for the x-axis is Oracle’s CAPEX spending.”
If Oracle really wants to get back in the game, they need to spend $100 billion this quarter, not $1 billion. Microsoft spent more on CAPEX last year than Oracle has in its entire multi-decade history. And it is so much worse to be the x-axis on this chart:
It is hard to make the case you will catch up in the infrastructure game when you’re 1.) ~$100 billion behind 2.) smaller in revenue (Oracle is less than $40 billion in revenues and growing at low single digits) and 3.) spend about a third as much per dollar of revenue on CAPEX: