So Oracle is going to build a Cloud that leverages machine learning as the competitive differentiation against AWS? Another company already has a Cloud like that, is considered one of the top Cloud providers, is growing much faster than Oracle and is typically less expensive than the competition (and certainly Oracle).
The only thing real avenue LE that to compete is to try to use fear, uncertainty and doubt to get existing customers (i.e. of database and ebusiness apps) not to move to the competition. While the database is the top tier option for large scale OLTP use, it is overpriced and alternatives are increasingly "good enough". Apps, other than NetSuite, are having a hard time competing. Three bulk of the rest of Oracle's software is mostly bloated, legacy stuff.
The one thing LE has always done "well" is market and use his hubris to display confidence in the face of superior competition. He is playing the same game again. The problem this time is that he can't buy out his competition to find a way into the market as all the other major players are significantly larger than Oracle. He misscalled the move to Cloud and now is stuck having to try to find a way to stay afloat until the next technology shift comes along.