Although I bought ORCL some time ago, I have recently done a lot of research (read many others' studies), especially on SEMI analysis.

I found that Neo Cloud is not really a good business. Or rather, it's not that good.
The investment logic of Neo Cloud essentially relies on investing in the phase mismatch of supply and demand.
Taking Oracle as an example, Oracle was one of the earliest to leverage debt to compete for data center capacity. So when supply and demand for computing power are mismatched, Oracle can meet excess demand with relatively high gross margins during a specific period.
But if we look at a longer time horizon, Oracle and companies like NBIS, which don't have their own chips, find it very difficult to outperform Google and Microsoft on the cost side. Not to mention Amazon.
Amazon has already entered TaaS (Token as a Service), while Oracle and others can only offer IaaS for now. The margin gap will be amplified.
Therefore, my strategy is to buy Oracle in the short term, capturing the short-term supply-demand mismatch, and in the long term, shift profits to Amazon.