Oracle, one of the initiators of the "AI bubble," has released its latest quarterly report. Naturally, the market responded negatively: shares dropped over 10% in after-hours trading, and the stock price is now significantly lower than it was before the announcement of the $300 billion contract with OpenAI.
Looking at the clearly deteriorating cash flow; looking at the sale of Ampere to support massive capital expenditures (which is really just a drop in the bucket); looking at the analysts who were actually quite friendly during the conference call; looking at the repeated explanations that "capital expenditure occurs when the data center is ready, and customers can typically reach full capacity and generate revenue within two or three days"; looking at the mounting RPO (Remaining Performance Obligations)...
Honestly, what else is there to say? You either believe it or you don't. I've said before that this is a business model worse than commercial real estate. However, if you believe the growth figures in the earnings report and perform a linear extrapolation, at this price level, you certainly can't call it a severe bubble anymore.
It all depends on how each individual views it.
Analyzing earnings reports and obsessing over every number is too exhausting; it's more interesting to research AI production.
By the way, the quality of the "nano banana pro" model has clearly declined: there are two visible errors—one with the timeline and one with text rendering. The latter never used to occur in previous 4K infographic modes.
Of course, reducing information load by spreading it across more slides helps significantly, but the issue of consistency remains impossible to solve perfectly.
